NHS CFHEP 001 Extension: Final report

Transcription

The Impact of eHealth on the Quality and
Safety of Healthcare
An updated systematic overview & synthesis of the
literature
Final report for the NHS Connecting for Health
Evaluation Programme (NHS CFHEP 001)
Aziz Sheikh, Susannah McLean, Kathrin Cresswell, Claudia Pagliari,
Yannis Pappas, Josip Car, Ashly Black, Akiko Hemmi, Ulugbek
Nurmatov, Mome Mukherjee, Brian McKinstry, Rob Procter and Azeem
Majeed
March 2011
Table of contents
List of authors
Acknowledgements
Abbreviations
Executive summary
SECTION 1: BACKGROUND
1. Introduction
SECTION 2: METHODS AND FORMATIVE WORK
2. Methods
3. The National Programme for Information Technology, NHS Connecting for Health
and the Information Revolution
4. Exploring, describing and integrating the fields of quality, safety and eHealth
SECTION 3: FINDINGS
Descriptive overview of results
5. Health information exchange and interoperability
6. Electronic health record
7. Computer-assisted history taking systems
8. Computerised decision support systems
9. Case study: Computerised decision support in mammography screening
10. ePrescribing systems
11. Case study: ePrescribing for oral anticoagulation therapy in primary care with
warfarin
12. Case study: eHealth-based approaches to reducing repeat drug exposure in
patients with known drug allergies
13. Telehealthcare
14. Case study: Telehealthcare for asthma
15. Case study: Internet-based interventions for smoking cessation
16. Human factors
17. Importance of organisational issues in the implementation and adoption of
eHealth innovations
18. Case study: Lessons in relation to the design, implementation and adoption of the
NHS Care Record Service in secondary care
SECTION 4: CONCLUSIONS, DISCUSSION
PRIORITIES
19. Conclusions, discussion and recommendations
20. Future research priorities
AND
FUTURE
RESEARCH
SECTION 5: APPENDICES AND GLOSSARY
Appendix 1: Search strategy
Appendix 2: Quality assessment form
Appendix 3: PRISMA flow diagrams
Appendix 4: Included systematic reviews
Appendix 5: Data extraction and critical appraisal scores of included reviews
Appendix 6: Programmes for eHealth Masterclasses
Appendix 7: Academic outputs
Glossary
2
List of authors and contact details
Ashly D. Black, Research Assistant, Health Unit, Department of Primary Care
and Social Medicine, Imperial College London
Dr Josip Car, Director of eHealth Unit, eHealth Unit, Department of Primary
Care and Social Medicine, Imperial College London
Kathrin Cresswell, Research Associate, eHealth Research Group. Centre for
Population Health Sciences, The University of Edinburgh
Dr Akiko Hemmi, Research Fellow, eHealth Research Group, Centre for
Population Health Sciences, The University of Edinburgh
Professor Azeem Majeed, Professor of Primary Care, Department of Primary
Care and Social Medicine, Imperial College London
Dr Susannah McLean, Clinical Research Fellow, eHealth Research Group,
Centre for Population Health Sciences, The University of Edinburgh
Dr Brian McKinstry, Reader, eHealth Research Group, Centre for Population
Health Sciences, The University of Edinburgh
Mome Mukherjee, Senior Research Analyst, eHealth Research Group, Centre
for Population Health Sciences, The University of Edinburgh
Dr Ulugbek Nurmatov, Clinical Research Fellow, eHealth Research Group,
Dr Claudia Pagliari, Senior Lecturer in Primary Care & Chair of eHealth Group
Dr Yannis Pappas, Deputy Director eHealth Unit, Department of Primary Care
and Social Medicine, Imperial College London
Professor Rob Procter, Professor of e-Social Science, Manchester eResearch
Centre, The University of Manchester
Professor Aziz Sheikh, Professor of Primary Care Research & Development,
Correspondence to: A Sheikh, Centre for Population Health Sciences, The
University of Edinburgh, Teviot Place, Edinburgh EH8 9AG, Email:
aziz.sheikh@ed.ac.uk
3
Acknowledgements
Identifying, retrieving, collating, synthesising and interpreting the vast body of
evidence that we have drawn on in compiling this review has been a
challenging task and would not have been possible without the help of a
number of people who we here take great pleasure in acknowledging.
Throughout the process of undertaking this work we have had helpful support
from colleagues at the NHS Connecting for Health Evaluation Programme,
namely Professor Richard Lilford, Jo Foster, Nathalie Maillard and Lee Priest.
Richard, Jo and Lee have kindly also represented the funders on our External
Steering Group which, under the able chairmanship of Professor Denis Protti,
and with the helpful support of Professor David Bates, Professor Jeremy
Wyatt, Dr Maureen Baker, Mr Antony Chutter and Ms Alison Turner provided
thoughtful and constructive advice.
Marshall Dozier, The University of Edinburgh’s librarian kindly helped with
developing the search strategy and we were aided in the selection of
references by Dr Tomislav Bokun, Matko Marlais and Chuin Ying Ung. Dr
Chantelle Anandan contributed to an earlier iteration of this report.
We have had the welcome opportunity to present our findings to colleagues at
a number of meetings both in the UK and internationally and we are grateful
for the feedback that has been received, which has either explicitly or
implicitly informed our thinking. In addition, Dr Chris Burton, Dr Sangeeta
Dhami, Dr Bernard Fernando, Professor Dave Fitzmaurice, Professor MariePierre Gagnon, Dr Hilary Pinnock, Professor Michael Sharpe and Dr Paul
Taylor kindly provided critical feedback on earlier versions of draft chapters,
as also did members of the External Steering Group. Our thanks also to the
three external reviewers who provided constructive suggestions on an earlier
version of this report.
Our thanks are also due to the many colleagues–both from the UK and
abroad–whom we have had the pleasure to work with on the various
4
Cochrane systematic reviews, in particular Dr Marta Civljak, Dr Lindsay Stead,
Dr Joe Liu and David Cahndler who worked with us on exemplar systematic reviews
incoroporated into this report. Our thanks also to the many colleagues who have
supported us in putting on the series of eHealth Masterclasses.
Anna Wierzoch provided administrative support and also proof-read the final
manuscript.
To all of these individuals, who so generously gave us their time, we wish to
record our sincerest thanks.
Finally, we are grateful to publishers to allow reproduction of figures, boxes,
tables and text referred to in individual chapters. Our thanks are in this latter
respect particularly due to The Cochrane Collaboration.
5
Abbreviations
ADE
ADR
AHR
AHRQ
AI
AMR
AQLQ
ATC
BA
BCSH
BMI
BNF
BP
BS7799
BWH
C
C&B
CAD
CAHTS
CASP
CBMR
CBPR
CBPRS
CCT
CI
CT
CCR
CDR
CDSS
CFHEP
CHI
CHIN
CIS
CMR
COPD
COTS
CP
CPOE
CPR
CPT
CRDB
CRS
CS
CSA
CSC
CTS
DCR
Adverse drug events
Adverse drug reaction
Automated Health Records
Agency for Healthcare Research and Quality
Artificial intelligence
Automated Medical Records
Asthma Quality of Life Questionnaire
Anatomical Therapeutic Chemical Classification Index
Before after (pre-post) study
British Committee for Standards in Haematology
Body Mass Index
British National Formulary
Blood Pressure
British Standard for Information Security Management
Brigham and Women’s Hospital
Case study
Choose and Book
Computer-aided detection or diagnosis
Computer-assisted history taking systems
Critical Appraisal Skills Programme
Computer-Based Medical Record
Computer-Based Patient Record
Computer-Based Patient Record System
Controlled Clinical Trial
Confidence Interval
Controlled Trial
Continuity of care record
Clinical Data Repository
Computerised Decision Support System
See NHS CFHEP
Community Health Index number
Community Health Information Network
Clinical information system
Computerised medical record
Chronic Obstructive Pulmonary Disease
Commercial off-the-shelf
Cerebral Palsy
Computerised Provider Order Entry
Computerised patient record
Current Procedural Terminology
Care Record Development Board
see NHS CRS
Cross-sectional study
Clinical Spine Application
Computer Sciences Corporation
Computer Telephony System
Detailed Care Record
6
DH
DIN
DIRC
dm+d
DMR
DSS
ECR
ECRI
ECS
EHCR
EHR
EM
eMAR
EMIS
EMR
EPHR
EPR
EPS
ERP
ETD
ETP
EV
FAA
FDA
GEP-HI
GINA
GMS
GP
GPRD
GPSoC
HbA1c
HCI
HF
HFE
HIE
HIEI
HIMSS
HIS
HL7
HRO
HT
HTA
ICBI
ICD
ICHPPC
ICNP
ICPC
ICR
ICU
IEHR
Department of Health
Doctor Independent Network
Dependable Interdisciplinary Research Collaboration
Dictionary of Medicines and Devices
Digital Medical Record
Decision support system
Electronic Care Record
Emergency Care Research Institute
Emergency Care Summary
Electronic Health Care Records
Electronic Health Record
Experimental methods
Electronic medication administration record
Egton Medical Information Systems
Electronic medical record
Electronic Personal Health Records
Electronic Patient record
Electronic Prescription Service
Enterprise resource planning
Education Training and Development
Electronic Transmission of Prescriptions
economic evaluations
Federal Aviation Authority
Food and Drug Administration
Guidelines for Best Evaluation Practices in Health Informatics
Global Initiative for Asthma
General Medical Services
General Practitioner
General Practice Research Database
GP Systems of Choice
Glycated Haemoglobin
Human-computer interaction
Human factors
Human factors engineering
Health information exchange
Health information exchange and interoperability
Healthcare Information and Management Systems Society
Hospital information system
Health Level Seven
High Reliability Organizations
hypothesis testing
Health Technology Assessment
Internet and Computer Based Interventions
International Classification of Diseases
International Classification of Health Problems in Primary Care
International Classification of Nursing Practice
International Classification of Primary Care
Integrated Care Record
Intensive care unit
Interoperable electronic health record
7
IM&T
Information management and technology
InPS
In Practice System
INR
International Normalised Ratio
IQAP
Information Quality Assurance Programme
IP
Internet protocol
ISAAC
International Study of Asthma and Allergies in Children
ISB
Information Standards Board
ISO
International Organization for Standardization
ISIP
Integrated Service Improvement Programme
IT
Information technology
JACHO
Joint Commission on Accreditation of Healthcare Organizations
KLAS
Keystone Library Automation System
LAN
Local Area Network
LHR
Longitudinal health record
LIS
Laboratory information system
LOINC
Logical observation identifiers names and codes
LPfIT
London Programme for Information Technology
LSPs
Local Service Providers
MA
Meta-analysis
ME
Medical Error
MeSH
Medical Subject Headings
MHRA
Medicines and Healthcare Products Regulatory Agency
MRC
Medical Research Council
MRHA
Medicines and Healthcare Products Regulatory Agency
N
narrative
NANDA
North American Nursing Diagnoses
N3
National Network for the NHS
N3SP
N3 Service Provider
NASP
National Application Service Provider
NHLBI
National Heart Lung and Blood Institute
NHS
National Health Service
NHS BSP
NHS Breast Screening Programme
NHS CFH
NHS Connecting for Health
NHS CFHEP NHS Connecting for Health Evaluation Programme
NHS CRS
NHS Care Records System
NIC
Iowa Nursing Intervention Classification
NICE
National Institute for Health and Clinical Excellence
NIH
National Institute for Health
NIHR
National Institute of Health Research
NISP
National Infrastructure Service Provider
NKS
National Knowledge Service
NLOP
National Programme for IT Local Ownership Programme
NOC
Iowa Nursing Outcome Classification
NPfIT
National Programme for Information Technology
NPSA
National Patient Safety Agency
NPT
Near patient test
NS
Not Specified
NSF
National Service Framework
OAT
Oral anticoagulation therapy
OBS
Output-based specification
8
OCCO
OITIM
OSCHR
OSI
OXMIS
P
PA
PACS
P&P
PACQLQ
PAQLQ
PAS
PDA
PBR
PC
PCEHR
PCR
PCT
PDA
PDF
PDS
PEHR
PHQ
PHR
PICS
PMCS
PMR
PMRI
PMS
POC
PRIMIS
PRO
PSIS
QI
QMAS
QOF
QoS
R
RCT
RDBT
RFID
RHIOs
RID
RTC
Rx
SBCCE
SC
SCI
SCR
SDS
Office of the Chief Clinical Officer
Organisation Information Technology/System Innovation Model
Office for Strategic Coordination of Health Research
Open System Iinterconnection
Oxford Medical Information Systems
Prospective
Predictive analysis
Picture Archiving and Communications System
Pen and paper
Paediatric Asthma Carer’s Quality of Life Questionnaire
Paediatric Asthma Quality of Life Questionnaire
Patient administration systems
Personal Digital Assistant
Payment by Results
Personal computer
Person-Controlled Electronic Health Record
Patient care record
Primary Care Trust
Personal digital assistant
Portable Document Format
Personal Demographic Service
Personal electronic health record
Patient Health Questionnaire
Personal health record
Prescribing Information and Communication System
Predictive modelling of cost savings
Patient (carried) medical records
Patient medical record information
Personal medical services
Point-of-care
Primary Care Information Services
Patient reported outcomes
Personal Spine Information Service
Qualitative interview
Quality Management and Analysis Subsystem
Quality and Outcomes Framework
Quality of Service
retrospective
Randomised controlled trial
NHS CFH Requirements, Design, Build and Test
Radiofrequency identification
Regional Health Information Organizations
Regional Implementation Director
Roadmap for Transformational Change
ePrescribing
Santa Barbara County Care Exchange
Secondary Care
Scottish Care Information
Summary Care Record
Spine Directory Service
9
SE
Software engineering
SHAs
Strategic Health Authorities
SI
Service implementation
SIGN
Scottish Intercollegiate Guidelines Network
SMD
Standard mean difference
SMS
Short message service
SNOMEDCT Systematized Nomenclature of Medicine Terminology-Clinical
Terms
SOAP
Subjective information, objective information, assessments and
plan
SR
Systematic review
STARE-HI Statement on Reporting of Evaluation Studies in Health
Informatics
SUS
Secondary Uses Service
TGC
Tight Glycemic control
TMS
Transaction Messaging Service
TRaMS
Training Messaging Service
TS
Time series
UK
United Kingdom
UKCRN
United Kingdom Comprehensive Research Network
US
United States of America
VATAM
Validation of Telematics Applications in Medicine
VME
Vancomycin resistant Enterococcus faecium
VPN
Virtual private network
WAN
Wide area network
WHO HEN World Health Organization’s Health Evidence Network
10
Executive summary
INTRODUCTION
•
Increasing life expectancy, improved survival in people with acute and
long-term conditions and a greater array of available treatment options
are combining to place an increasing burden on healthcare systems
internationally.
•
This picture is particularly true of the economically developed world
where high salaries for healthcare professionals and ever-increasing
public expectations contribute to the challenges facing governments
trying to contain spending on healthcare provision and planning.
Similar pressures are also emerging in economic transition and
developing countries.
•
There is now a substantial body of research, both domestic and
international, identifying considerable shortcomings in the current
provision of healthcare.
•
Key issues emerging from this literature are substantial variations in
the quality of healthcare, the considerable risks of iatrogenic harm and
inefficiencies in healthcare provision; this latter issue has become a
particularly acute cause for concern in the light of the economic
downturn and the resulting pressures on healthcare budgets.
•
These failings contribute in a major way to the high rates of potentially
avoidable morbidity, mortality and healthcare expenditure.
•
There have been substantial developments in information technology
(IT), hardware and software capabilities over recent decades and there
is
now
considerable
potential
to
apply
these
technological
developments in relation to aspects of healthcare provision (the
application of IT in this way will henceforth in this report be referred to
by the term “eHealth”).
•
Of particular international interest is the development and deployment
of an array of eHealth applications, with a view to improving the quality,
safety and efficiency of healthcare delivery.
11
•
Whilst these eHealth applications have considerable potential to aid
professionals in delivering healthcare and patients in self-management,
it is not widely appreciated that use of these new technologies may
also introduce significant new risks to patients.
•
Also of concern is that even when potentially useful interventions are
developed and deployed, they frequently fail to live up to their potential
when deployed in the “real world”; a major factor contributing to this
paradox is the failure of these technological fixes to integrate effectively
with existing work patterns or to be adequately conceptualised as a
change management reform which can then be used to reshape
delivery of care.
•
Given that England’s National Health Service (NHS) is engaged in one
of the largest eHealth-based modernisation programmes in the world, it
is appropriate and timely to critically review the international eHealth
literature with a view to identifying lessons that can usefully be learned
with respect to the development, design, deployment, integration and
evaluation of eHealth applications.
AIMS AND OBJECTIVES
•
We were originally commissioned by the Patient Safety Research
Programme and then by the NHS Connecting for Health Evaluation
Programme (NHS CFHEP) to produce a systematic overview of the
literature examining the effectiveness of IT (eHealth) applications to
improve the quality and safety of healthcare.
•
We focused in our initial phase of work on evidence relating primarily
to:
o the storage and retrieval of medical information
o tools to support healthcare professionals in making clinical
decisions
o ways of promoting the effective development, deployment and
use of eHealth applications in routine healthcare settings.
•
In the second phase of work, this was expanded to:
o include evidence on delivering healthcare from a distance
12
o undertaking de novo systematic reviews in key eHealth areas
o co-ordinating a series of “eHealth Masterclasses” aimed at
policymakers, academics and practitioners
o developing a database of the critically appraised and critiqued
eHealth literature
o updating and integrating this with the previous body of work on
storing and managing healthcare information, supporting clinical
decision making and promoting the successful adoption of
eHealth applications.
•
This report focuses on providing a synopsis of the main findings from
the synthesis of this body of work: findings from the new Cochrane
systematic reviews that we have initiated and/or contributed to are
integrated, where appropriate, into chapters and a summary of these
outputs is detailed along with the programmes for the Masterclasses in
the report’s appendices; the final database of eHealth systematic
reviews will, with the agreement of the funders, be submitted later this
year. An application to support the publication of this database through
NHS Evidence will be submitted in April 2011.
METHODS AND FORMATIVE WORK
Methods
•
We conducted a systematic search and critique of the empirical
literature on eHealth applications and their impact on the quality and
safety of healthcare delivery and synthesised this with relevant
theoretical, technical, developmental and policy-relevant literature with
a view to producing an authoritative and accessible overview of the
field.
•
Whilst we drew on established Cochrane review principles to
systematically search for, critique and synthesise the literature, this
approach needed to be adapted in several respects in order to produce
a meaningful overview of the literature (see below).
•
Searching the literature was complicated by the lack of internationally
(or indeed in some cases nationally) agreed terminology relating to
13
eHealth applications, the lack of agreed definitions of quality and
safety, and consequently poor indexing of these constructs in
databases of published literature.
•
In order to undertake a thorough review of the literature, we therefore
needed to undertake initial developmental work to formulate a
comprehensive search strategy.
•
Using the set of comprehensive Medical Subject Headings (MeSH) and
free text search terms developed, we systematically searched major
medical databases over a 13-year period (1997–2010) to identify
systematic
reviews,
technical
reports
and
health
technology
assessments, and randomised controlled trials investigating the
effectiveness of eHealth applications.
The specific databases
searched were:
o MEDLINE
o EMBASE
o The Cochrane Database of Systematic Reviews
o Database of Abstracts of Reviews of Effects
o The Cochrane Central Register of Controlled Trials
o The Cochrane Methodology Register
o Health Technology Assessment Database and NHS Economic
Evaluation.
•
We, in addition, searched key national and international databases to
identify unpublished work and research in progress.
•
The systematic reviews have then been subjected to critical review
using the Critical Appraisal Skills Programme (CASP) approach,
adapted for use with eHealth applications.
•
The de novo systematic reviews that have contributed to this review
have been conducted using established Cochrane review methods.
•
To provide a broader appreciation of the context of this work and
furthermore to aid conceptual development and interpretation of
findings, we supplemented this systematic search for empirical
evidence with a more emergent approach to identify relevant
background and theoretical literature in relation to the essential
concepts underpinning this overview—namely: eHealth; quality; safety;
14
and efficiency. This involved drawing on our personal databases of
relevant papers, identifying seminal papers and reports as well as
searching the grey literature.
•
The overall body of literature identified was too diverse to make any
quantitative synthesis of the literature meaningful, nor indeed was it
possible. Rather, we chose to qualitatively synthesise the literature
drawing on the relevant preliminary conceptual work to guide this
narrative synthesis.
•
Our overall assessment of the volume and strength of evidence in
relation to key findings are summarised in this Executive summary
using a modified version of the World Health Organization’s Health
Evidence Network (WHO HEN) system for public health evidence,
which grades evidence into three main categories:
o strong (consistent, good quality, plentiful or generalisable)
o moderate (consistent and good quality)
o limited to none (inconsistent or poor quality).
NHS Connecting for Health and the National Programme for Information
Technology
•
The NPfIT remains one of the most comprehensive, ambitious and
expensive eHealth-based overhauls of healthcare delivery ever
undertaken.
•
This Programme has its origins in the 1998 Department of Health
strategy Information for Health, which committed the NHS to lifelong
electronic health records for everyone with round-the-clock, on-line
access to patient records and information about best clinical practice
for all NHS clinicians. The current Programme, launched in 2002, was
to be a 10-year initiative aiming initially to create the infrastructure,
tools and environment through which it is possible to deliver:
o a longitudinal electronic patient record (from “cradle to grave”)
accessible to multiple users throughout the NHS; this (i.e. NHS
Care Records Service or NHS CRS) together with the dedicated
NHS broadband (National Network for the NHS or N3) and the
15
national database on which these records will be held (the
Spine); represented the backbone of the Programme
o a service through which prescriptions can be transferred
electronically from the general practitioner and other prescribers
to pharmacists (Electronic Prescriptions Service) and eventual
integration of this with NHS CRS (Electronic Transmission of
Prescriptions or ETP)
o an electronic appointment booking service enabling general
practitioners to electronically book hospital appointments
(Choose and Book).
•
The Programme was however subsequently expanded to include
amongst other things:
o a Picture Archiving and Communication System (PACS)
o GP2GP, which is a system that enables transfer of patient
records between GP practices
o Quality
Management
Analysis
System
(QMAS),
which
automates assessment of GP practice performance against
criteria included in the new GP contract
o ePrescribing.
•
Whilst these represented the headline deliverables of the Programme,
our scoping of the field has identified a number of other related eHealth
projects or applications which are also closely inter-connected with the
delivery of the Programme.
•
Originally managed directly by the Department of Health, oversight of
NPfIT was transferred in 2005 to a newly created arm’s length body,
namely NHS CFH.
•
Foremost amongst the roles of NHS CFH was responsibility for
nationally procuring systems and services that will be needed to ensure
delivery of NPfIT.
•
Given the extremely high level of public expenditure, the Programme
attracted considerable public, professional, legal, financial, political and
international scrutiny.
16
•
It is thus probably no exaggeration to say that in addition to it being the
most comprehensive, ambitious and expensive eHealth reform
programme in the world, it is also the most influential in that its success
or failure is likely to have major domestic and international ripples for
many years to come.
•
NHS CFH has more recently been subsumed into the Department of
Health’s Informatics Directorate.
•
Following the election of the new coalition government in 2010, the
scope and ambition of NPfIT/NHS CFH has been reduced with a rapid
move
to
less
centralised
and
more
local
procurement
and
implementation of IT into NHS healthcare settings.
Exploring, describing and integrating the fields of quality, safety and
eHealth
•
eHealth remains a relatively new and rapidly evolving field and so
many of the concepts, terms and applications are still in a state of flux.
•
There is furthermore no agreed definition of eHealth, with some
researchers using this to relate primarily to the area of consumer
informatics, whereas others use it more generically to refer to any of
the ways in which IT can be employed to aid the deliver of healthcare.
For the purposes of this review, we considered it important to use an
inclusive definition and chose to use Eysenbach’s definition, as
adapted by Pagliari, for the basis for our work:
‘eHealth is an emerging field of medical informatics, referring to the
organisation and delivery of health services and information using the
Internet and related technologies.
In a broader sense, the term
characterizes not only a technical development, but also a new way of
working, an attitude, and a commitment for networked, global thinking, to
improve healthcare locally, regionally, and worldwide by using information
and communication technology.’
•
Whilst the number of eHealth applications is potentially vast, these can
nonetheless be divided into three broad domains relating to key
activities they support:
o storing, managing and sharing data
17
o informing and supporting clinical decision-making
o delivering expert professional and or consumer care remotely.
•
The effective commissioning, development, deployment and routine
use of eHealth applications is a cross-cutting area that impacts on each
of these three domains.
Quality
•
There are no internationally agreed definitions of healthcare quality.
•
Most frameworks of quality currently in use do, however, incorporate
the following key dimensions of care:
o effectiveness of treatments
o appropriateness of means of delivery
o acceptability
o efficiency
o equity.
Safety
•
Whilst there are no internationally agreed definitions of patient safety,
adaptations of the National Patient Safety Agency’s definition of Patient
Safety Incidents are increasingly being used. This, in its original form,
states that:
‘A patient safety incident is any unintended or unexpected incident which
could have harmed or did lead to harm for one or more patients being
cared for by the NHS.’
•
There are a number of patient safety taxonomies currently in existence,
however, our scoping of this literature found that the Joint Commission
on Accreditation of Healthcare Organizations (JCAHO) Patient Safety
Event Taxonomy is the most comprehensive and clinically relevant in
that it incorporates five key primary areas:
o impact of medical error
o type of processes that failed
o domain, i.e. the setting in which an incident occurred
o cause or factors leading to the safety incident
18
o prevention and mitigation factors to reduce risk of recurrence
and or improve outcomes in the case of a further incident.
Integrating eHealth, quality and safety
•
Integrating
the
fields
of
eHealth,
quality
and
safety
clearly
demonstrates the numerous ways in which technology has the
potential to improve the efficiency of many facets of healthcare
delivery. This can occur through, for example, helping clinicians readily
to access comprehensive information on their patients, aiding
monitoring of their conditions and the treatments being issued,
reducing inappropriate variability in healthcare delivery, and proactively
identifying and alerting clinicians to threats to patient safety.
•
This integration of these domains, however, also highlighted the many
ways in which introduction of new eHealth applications could
inadvertently increase risks.
MAIN FINDINGS
•
Our initial searches retrieved a total of 67 systematic reviews. Updating
and expanding this review during the second phase of work has
identified an additional 95 systematic reviews.
We thus in total
identified 162 systematic reviews published during the period 1/1/199731/10/2010.
•
We have initiated 20 eHealth-related systematic reviews during the
course of the second phase of work, the majority of which have or are
now nearly completed. The key findings from this body of work have
also been incorporated into the review.
•
The volume of primary and secondary literature is large, rapidly
expanding, poorly collated and of very variable quality; as a result the
literature surrounding eHealth poses major challenges to synthesis and
interpretation.
•
In synthesising the available evidence, we used the following generic
approach in relation to different eHealth applications and their related
considerations:
19
o clarifying definitions, description and scope for deployment
o drawing on conceptual maps to reflect on the potential benefits
and risks of each application
o identifying the empirically demonstrated benefits and risks, using
exemplar subject areas and or detailed case studies on issues
that are of direct or potential future relevance to the UK
o based on a synthesis of the above, highlighting the policy,
clinical and research implications for the individual areas of
interest with a view to realising the potential that eHealth has to
offer.
•
Our findings’ chapters are grouped together in relation to the four main
foci of this report, namely the domains of:
o managing, storing and transmitting data
o supporting clinical decision-making
o supporting the remote delivery of care
o the cross-cutting issue of the socio-techno-cultural dimensions
of developing and implementing eHealth applications.
Data storage, management and retrieval
•
Effective and efficient sharing of clinical information is essential to the
future development of modern healthcare systems, which are
increasingly characterised by the involvement of many specialist
healthcare providers, often working from different sites, contributing to
the care of individual patients.
•
The ideal in this respect is for professionals and patients themselves to
have the ability simultaneously to access and seamlessly transfer,
contribute to and integrate clinical data from disparate sources.
•
The potential gains in relation to improving the quality, safety and
overall efficiency of healthcare delivery are potentially enormous, as
suggested by a recent US-based economic analysis.
•
Most UK healthcare settings are however currently characterised by
relatively low levels of health information exchange and interoperability
20
(HIEI) capability, this being particularly true of the hospital sector,
where paper-based records are still the main means of recording and
communicating clinical information.
•
NPfIT has however substantially increased the potential for HIEI, for
example, through N3 and the Spine; deployment of the NHS Care
Record Service (NHS CRS) aims to result in the creation of summary
and detailed electronic health records that have the potential to be
shared, to varying degrees, across healthcare settings and between
providers.
•
NHS CFH has insisted that new eHealth applications must be Health
Level Seven (HL7) compliant—a voluntary but nonetheless widely used
standard for interoperability—thereby assuring a degree of ability to
exchange information between systems. However, while this is
undoubtedly welcome, none of the headline NPfIT applications will
achieve the ideal levels of HIEI, with the result that patient safety may
needlessly be compromised.
•
The current empirical evidence-base that supports the role of HIEI in
improving organisational efficiency, practitioner performance or indeed
any clinical patient outcomes.is, however, weak.
•
Although improving HIEI to the optimal level to allow seamless transfer
and access to data in all settings, will probably result in societal costsavings in the longer run, this will inevitably require considerable
upfront investment in hardware and software
•
An important paradox to further developments in this area is that whilst
increasing levels of HIEI are clearly desirable for many reasons,
greater availability of data also inevitably increases the risk of threats to
breaches of patient confidentiality and data security.
•
Key outstanding issues that face healthcare systems in realising the
potential for seamless exchange of information include the need to
develop and deploy standard coding structures across all care settings
(e.g. using Systematized Nomenclature of Medicine-Clinical Terms
(SNOMED-CT)), facilitate integration of the increasing amounts of
patient-generated data (e.g, through HealthSpace, home sensors
21
and/or telemetry devices), and improve secure audited access to
electronic records to minimise the risks of breaching confidentiality.
•
Making progress in optimising the ability to share and integrate data
seems particularly relevant given the expressed government objective
of creating an “information revolution” within the NHS.
Electronic health record
•
The electronic health record (EHR) represents the backbone of all
major international eHealth developments currently taking place
internationally, including NHS CFH’s NPfIT.
•
The ultimate goal is to have available comprehensive longitudinal
health information for all members of the population, with the potential
for accessing and contributing to these records by multiple users
working across a range of healthcare settings; no country has however
yet to achieve this comprehensively and if successful the NHS CRS will
be one of the first in the world to come close to this aspiration.
•
Electronic health records range from simple storage devices to those
with varying degrees of added functionality, including the ability to
electronically prescribe (ePrescribing) and access to computerised
decision support systems (CDSSs), which are active knowledge
systems, which use individual patient data to generate case-specific
advice.
•
The main potential advantages of EHRs relate to improved legibility
and, particularly if structured templates are used, comprehensiveness
of recording information, access by multiple users that is not
geographically-bound (if interoperable), the ability to incorporate
professional support tools, and time and cost savings.
•
There are, however, important potential risks associated with the EHR,
these in the main relating to disruptions to work flow, professional
frustration and dissatisfaction, data security considerations; there is in
addition, the concern that clinically important information may be
overlooked, particularly in contexts where there is parallel recording of
data using both electronic and paper records.
22
•
The empirically demonstrated benefits relating to introduction of EHRs
are currently limited to improved legibility, time savings for some
professionals (nurses), and the facilitation of higher order functions
such as audit, secondary analysis of routine data and performance
management.
•
Time taken for doctors to enter and retrieve data has in contrast been
found to increase; studies have furthermore found that the time
disadvantage for clinicians to record and retrieve information did not
attenuate with increased familiarity and experience with using EHRs.
•
There is moderate evidence that these can help improve patient
outcomes, particularly in relation to provision of preventive care.
•
Standardised and widely accepted measures of data quality in EHRs
are lacking and their development should be a priority.
•
Important potential national future development for EHRs include better
human-machine interfaces which allow, for example, the use of voice
recognition to speed up data input and the ability to incorporate multimedia files such as heart sounds, retinal screens and audio- or videorecordings of consultations.
•
Most computer-assisted history taking systems (CAHTS) are designed
for use by healthcare professionals, although some elicit information
directly from the patient, as in the case of pre-consultation interviews.
•
Computer-assisted history taking systems can be used in a variety of
clinical settings and have, when eliciting data directly from patients,
proven particularly useful in identifying potentially sensitive information
such as alcohol consumption, sexual health and psychiatric illnesses
(e.g. suicidal thoughts).
•
Computer-based questionnaires are particularly useful for gathering
important background data prior to the consultation, which can then
allow more time for focusing on key aspects of the health problems in
the actual consultation.
These systems may also save money by
reducing administrative costs.
23
•
Speech software and speech completed response CAHTS allow
adaptability for those with particular needs such as non-English
speaking patients, patients with hearing impediments and those who
are illiterate.
•
These can also prove helpful in the context of delivering high volume
care in specific contexts such as the screening of patients in the
context of pandemic illness (e.g. H1N1-pandemic influenza).
•
There is moderate evidence that data collected electronically tend to be
more accurate and contain fewer errors than data captured manually
with traditional pen and paper techniques; such data are also more
legible.
•
The current generation of computers is however not adept at detecting
non-verbal behaviour; these systems should therefore be seen as not a
substitute, but rather an adjunct to the clinical history.
•
There have as yet been no comparative studies that have formally
assessed
the
effectiveness
and
cost-effectiveness
of
different
CAHTSs.
•
It is important for the DH’s Informatics Directorate to carefully consider
the considerable potential efficiency gains to be made from
incorporating CAHTS functionality—particularly if this involves direct
entry of data by patients—into future iterations of the informatics
strategy. HealthSpace could facilitate this as could a number of other
modalities such as touch-screen or voice-recognition equipped
computers available in waiting rooms. This will, however, need to be
introduced within a clear evaluative context.
Supporting professional decision-making
Computerised decision support systems
•
There are strong theoretical reasons for believing that improved access
to relevant clinical information for healthcare professionals, at the point
of care, can translate into improvements in healthcare quality, patient
safety and organisational efficiency.
24
•
Defined as software applications that use individual patient data,
computerised decision support systems (CDSSs) utilise a repository of
clinical information (knowledge-base) and an inference mechanism
(logic) to generate patient-specific output.
These applications are
highly variable in sophistication, output and the extent to which they
can integrate with other clinical information systems.
•
Computerised decision support systems have the potential to improve
clinical decision-making by providing practitioners with real time
patient-specific
and
evidence-based
support
and
by
providing
individually tailored feedback.
•
Although numerous evaluations of CDSSs have taken place, very little
consistent and generalisable evidence exists on their ability to improve
patient outcomes; the available evidence suggests however that
practitioner performance can be improved, particularly in relation to
relatively simple tasks such as aspects of preventive care (e.g.
vaccination reminders) and monitoring of those with long-term
conditions
(e.g.
blood
pressure
monitoring
in
people
with
hypertension).
•
The use of computerised reminders for provision of preventive care has
been empirically demonstrated to be of benefit. However, trials have
tended not to focus on patient outcomes; this may be because for
most preventive care interventions, the numbers needed to be studied
and/or and the time needed to demonstrate an effect on patient
outcomes are prohibitively large/long.
•
Commercial CDSSs are largely unregulated in the US and UK due to
exclusion from the Federal Drug Administration and Medicine and
Healthcare Products Regulatory Agency respectively.
•
Without formal quality and safety assurances in relation to CDSS
applications, the potential risks to patient safety need to be seriously
considered as they may in certain situations inadvertently introduce
new errors.
•
As CDSSs work best when interfacing with or integrated within existing
clinical information systems, and the evidence of benefit is clearest and
25
risk of harm is least for provision of preventive healthcare, the DH
should consider introducing a range of computerised health promotion
tools into primary care and with the roll-out of the EHR into secondary
care.
•
Research should focus on understanding the contexts in which these
decision support tools are most likely to prove effective and this should
be a priority consideration for the DH as it introduces new eHealth
applications with built-in decision support functionality.
ePrescribing
•
There is considerable variation in the quality of prescribing. Medicines
management errors are common, costly and an important source of
iatrogenic harm.
•
ePrescribing can be defined as the use of computing devices to enter,
modify, review and output or communicate prescriptions. ePrescribing
systems are highly variable in genealogy, functionality, configurability
and the extent to which they integrate with other systems.
•
ePrescribing has the potential to greatly improve the quality and safety
of prescribing, through facilitating cost-conscious evidence-based
prescribing and in particular reducing errors associated with knowledge
gaps and routine tasks such as repeat prescribing.
•
There is moderate evidence that practitioner performance is improved
through better access to these guidelines.
Patient outcomes are,
however, less well studied and, when assessed, most studies have not
been able to demonstrate a clinical benefit. This is likely to be due to
the prohibitive numbers needed to demonstrate a clear impact on
infrequent, but serious outcomes.
•
Evidence of benefit from ePrescribing systems has in the main been
demonstrated from evaluations of “home-grown” systems in a few
centres of excellence. Most systems in use are however commercially
procured and these systems typically lack the sophistication, clinical
relevance and sense of ownership associated with the tailored homegrown systems.
26
•
Poorly designed ePrescribing systems and a failure to appreciate the
socio-techno-cultural issues associated with their introduction can
introduce unexpected new risks to patient safety.
•
In the UK, the Medicines and Healthcare Products Regulatory Agency
does not consider ePrescribing systems to be a medical device and
does not therefore require these systems to be quality assured. This is
an important policy failing that needs to be addressed.
•
Further research into the design features, knowledge-bases and
underlying algorithms, clinical relevance of output, interoperability of
ePrescribing systems and socio-technical factors that enhance use is
needed in order to replicate the benefits of ePrescribing that have been
demonstrated in US centres of excellence. This research needs to
focus on commercial systems.
•
There is merit in considering the introduction of ePrescribing systems
into NHS hospitals, but in an evaluative context.
Remote delivery of care
Telehealthcare
•
Current systems of healthcare are inadequate to deal with rapidly
increasing populations, many of whom have one or more long-term
conditions.
•
Telehealthcare can be defined as “personalised healthcare delivered
over a distance.” It thus involves transfer of data from the patient to a
professional using IT, which is then followed by personalised feedback
from the professional to the patient.
•
It has the potential to improve access to care, promote preventive care,
and help manage the health and social care needs of, amongst others,
those with long-term conditions.
•
Whilst potentially attractive, these new models of care can however
prove very costly and may also cause a considerable change to
consultation dynamics and workflow processes; these should therefore
be introduced only after a careful assessment of effectiveness, costeffectiveness and safety considerations.
27
•
Most patients have in general a positive view of telehealthcare, so long
as it is offered in addition to, rather than as a substitution for, the faceto-face consultation. Qualitative studies have highlighted that the needs
of the carer should also be taken in account when designing
telehealthcare systems.
•
Overall, there is for people with long-term conditions such as asthma,
chronic obstructive pulmonary disease and diabetes now moderate
evidence that telehealthcare can reduce hospital admissions in those
with more severe disease without increasing mortality; the evidence in
relation to improving quality of life in those with
milder disease is
however less encouraging.
•
There is a need for more cost-effectiveness studies which take into
account the full range of perspectives, i.e. that of the patient,
healthcare services and society.
•
Important potential pitfalls of telehealthcare that need to be considered
include user interface issues such as older people not understanding
how to operate systems, technical breakdown, and safety concerns
such as data loss and breaches of confidentiality.
•
The use of telehealthcare often results in a reshaping of the doctorpatient relationship and attention should be paid to the opportunities for
humanising this interaction. Workflows also need to be considered to
minimise the risk of unintended disruptions to normal working routines.
Socio-technical dimensions of designing, developing, and deploying
eHealth applications
Human factors
•
The aim of human factors (also known as ergonomics) is to bring to
bear an understanding of how people’s conduct is shaped by their
technological, organisational and social environments to processes of
design and innovation so that new technologies are compatible with
their users’ goals, capabilities and needs.
•
When new technologies are introduced into the workplace, success will
typlically depend on addressing a diverse range of human factors
28
issues,
including:
the
nature
of
existing
work
practices
and
performance requirements (including efficiency, dependability and
safety); user skills; possibilities for the re-design of work (including user
perspectives and preferences); and training requirements (including
formal, informal and on-the-job).
•
The healthcare sector has been slow to incorporate human factors into
the design and evaluation of eHealth applications, despite the
mounting complexity of care delivery systems and evidence of resulting
risk to patients from poorly designed systems.
•
A well-designed user interface is, for example, as important as
functionality and reliability in applications such as ePrescribing.
Confusion and frustration with an application interface are enough to
impede users’ acceptance, with an adverse subsequent knock-on
effect on implementation. User-centred design and evaluation for
usability are critical for acceptance.
•
There is a lack of consensus in the literature regarding how best to
incorporate human factors into the design, development and evaluation
of eHealth applications..The DH needs to establish best practice
guidelines for human factors in eHealth applications, including
standards for usability evaluation, and ensure that a human factors
audit and usability findings are available to inform assessment of new
eHealth applications prior to approval for use in the NHS.
•
The EHR, ePrescribing and CDSS systems are examples of complex
applications for which it is clearly essential that user needs have been
adequately considered in all stages of design, development and
deployment.
•
Critical feedback from users of new eHealth applications should not
only be facilitated, but must also be actively encouraged so as to
ensure that new applications are fit-for-purpose and minimise risks to
patient safety.
•
Embedding human factors principles and thinking is resource intensive;
the DH needs to ensure that adequate time, resources and
29
prioritisation are given to this so as to maximise the chances of
success of its various eHealth initiatives.
Effective implementation and adoption of eHealth applications
•
Many technological innovations fail to realise their potential and this
unfortunately has also been true with respect to the history of eHealth
applications.
•
Major factors contributing to these failures—which may in some cases
be spectacular—include the lack of appreciation and attention paid to
human factors’ issues during product development and socio-technical
factors that subsequently enable innovations to integrate with and
embed themselves into healthcare organisations.
•
In planning IT deployments, health systems need to be clear as to
whether the aim is to replicate existing practices or to use the
intervention as a basis for organisational reform.
•
There is a burgeoning change management literature, dating back to
the influential Diffusions of Innovation theory and stretching to the more
recent Diffusion of Innovations in Health Service Organisations work.
Much of this is however descriptive and so the predictive ability of
these models of change management is as yet unknown.
•
Using an adaptation of the above and other theories to render them
more relevant to the planned dissemination of eHealth applications, we
have developed and updated our Infusion of eHealth Innovations in
Health Services Organisations Model to
undertake a detailed case
study seeking to learn lessons from NHS CFH’s approach to promoting
the NHS CRS in secondary care.
CONCLUSIONS AND FUTURE RESEARCH PRIORITIES
•
In undertaking the first phase of work, we made four main
methodological contributions to this nascent field of enquiry, namely:
o development of a comprehensive search strategy for identifying
high quality primary and secondary literature investigating the
impact of eHealth on the quality and safety of healthcare
30
o development of integrated conceptual maps of eHealth, quality
and safety, which have, as demonstrated in this project, the
ability to draw attention to the major potential benefits
associated with use of different eHealth applications
o adapting a tool for critically appraising systematic reviews of
eHealth applications
o development of a framework with which to consider the planned
implementation of eHealth innovations into complex health
service organisations.
•
The second phase of work has allowed us to initiate and undertake:
o a suite of Cochrane systematic reviews on various aspects of
eHealth, in particular focusing on telehealthcare
o hold a series of national Masterclases, which have enabled
academics,
policymakers,
clinicians
and
industry
representatives to come together and discuss and debate how
best to maximise the health gains associated with eHealth.
•
The formative work for this project and the review of technical reports
and a variety of review documents clearly demonstrate that eHealth
applications have great potential to improve the quality of healthcare
delivery; even more importantly, perhaps, there is considerable
potential to improve the safety profile of medicine through elimination of
both latent and active errors, and through promoting real-time systems
checks.
•
The major finding from reviewing the empirical evidence—which is of
variable quality—however, is that there is still as yet only limited
evidence demonstrating that these technologies actually improve
patient outcomes.
•
The reasons for this are multi-faceted; these include:
o an assumption that the benefits associated with these
applications are obvious, thus resulting in many (and indeed
perhaps the majority of) interventions being implemented
without due consideration to evaluation
31
o using proxy outcome measures as opposed to those that are
actually clinically important
o poorly theorised interventions and studies
o overestimating likely effect sizes
o methodological
limitations,
particularly
the
failure
to
appropriately use cluster designs in studies that are at risk of
contamination, poor outcome definition and measurement, and
approaching these interventions as the equivalent of simple
interventions, which they are typically clearly not.
o naïve assumptions that these technologies will be equally
effective in all contexts and settings
o inappropriately short time-frames to study likely health gains
o the failure to involve end-users at a sufficiently early stage in the
design and deployment process such that they can actually
influence factors that are likely to increase acceptability of the
interventions to clinicians
o a failure to pay adequate attention to the socio-technical factors
that are likely to be important in relation to the diffusion,
implementation and adoption of these applications.
•
Despite these substantial gaps in the evidence-base—many of which
can be filled—we are, on the basis of the theoretical work and empirical
evidence reviewed, cautiously optimistic that a number of the eHealth
applications that have been introduced into the NHS are likely to result
in significant medium- to long-term benefits to organisation efficiency,
professional practice and in time patient care.
•
We would
in
particular encourage
the
DH to
prioritise
the
implementation of ePrescribing into hospitals, ideally with CDSS
functionality in an integrated way within the NHS CRS within secondary
and tertiary care settings, and also to improve ePrescribing
functionality in primary care.
•
There is also a need to ensure that adequate expertise exists to ensure
that the considerable body of data now being generated–whether
32
through routine healthcare data sources and/or the increasing array of
telehealthcare initiatives–are appropriately interrogated and exploited.
•
Realising the benefits is however likely to be crucially dependant on
menaingfully facilitating end-user input throughout the commissioning,
design, development and implementation process as this will maximise
the chances that aspects of these technologies that are actually
clinically useful are developed; this will also promote a sense of
ownership and buy-in into the technological innovation and any
accompanying service redesign.
•
Appreciating the structural, organisational and, to a lesser extent,
professional challenges that need to be overcome in the deployment
should not, however, be
transformative
overlooked, particularly when complex
technologies such as the NHS CRS are introduced.
The need for training in the use of new technologies and on-the-job
support also needs much greater emphasis than tends to be the case.
•
End-user consultation and feedback needs to be viewed as an ongoing process and should therefore continue after deployment to
ensure that problems are identified early, as are possible solutions,
which can be incorporated into system upgrades.
•
We have throughout this report identified a number of areas in relation
to specific technologies where further research is warranted. There are
however a number of broader research considerations that need to be
prioritised, which we emphasise here, namely:
o the need for further conceptual development and then
international consensus building on use of standardised
terminologies to facilitate future primary and secondary work
o the development of a methodological toolkit to facilitate
evaluation of eHealth applications throughout all aspects of the
development and deployment lifecycle of these technologies.
o the need to encourage researchers to, where appropriate,
combine rigorous quantitative and simultaneously conducted
qualitative assessments when evaluating the effectiveness of
new eHealth applications to allow a detailed appreciation of
33
relevant contextual factors that might help better understand the
reasons for the success or failure of the intervention to emerge
and also, where found to be successful, to allow an assessment
of the likely generalisability of the intervention to be made
o for consensus building on the contexts in which proxy measures
are appropriate surrogates for clinically important end-points.
•
Given the considerable gulf between the potential advantages
associated with eHealth applications and the actual empiricallydemonstrable benefits we make the following recommendations:
o our major policy recommendation is
the need for more
considered and cautious claims of the likely benefits and
possible risks associated with the introduction of new technology
into healthcare settings, and the need to be cognisant of the fact
that technical, human and organisational factors are all likely to
be important success factors, particularly in the context of
implementing disruptive innovations
o
our major
research
recommendation
is
to
ensure
that
independent evaluation into the risks and benefits of eHealth
continues: given the scale of the investment taking place, it is
vital that every opportunity is taken to ensure that this public
money produces the desired outcomes.
•
In conclusion, reviewing this large body of evidence has demonstrated
that eHealth is an area of considerable policy interest internationally
with considerable potential to impact on care delivery.
But despite
emerging evidence of positive impacts on aspects of professional
performance, there is as yet little clear evidence that these
developments are translating into improvements in patient outcomes.
Demonstrable benefits from many of the IT-related initiatives ushered
in through the NPfIT are, we suspect, most likely to be seen from the
reuse of digital data for the purposes of, amongst other things, policy
planning, local audit and governance activities, patient empowerment
and research. Recent IT developments are likely therefore, over time,
to help consolidate the UK’s acknowledged expertise and leadership in
34
the secondary uses of data, which aligns closely with the DH’s new
policy priority of initiating an “information revolution”.
SECTION 1
BACKGROUND
35
Chapter 1
Introduction
Summary
•
Increasing life expectancy, improved survival of people with acute and
long-term conditions and a greater array of available treatment options are
placing an increasing burden on health systems internationally.
•
There is now a substantial body of research, both domestic and
international, identifying considerable shortfalls in the current provision of
healthcare.
•
Key issues emerging from this literature are substantial variations in the
quality of healthcare and the considerable risks of iatrogenic harm.
•
These failings make a significant contribution to the high rates of
potentially avoidable morbidity and mortality, and rising healthcare
expenditure.
•
There have been substantial developments in information technology
hardware and software capabilities over recent decades and there is now
considerable potential to apply these technological developments in
relation to aspects of healthcare provision.
•
Of particular international interest is the deployment of eHealth
applications—that is the use of information technology in healthcare
contexts—with a view to improving the quality, safety and efficiency of
healthcare.
•
Whilst these eHealth technologies have considerable potential to aid
professionals in delivering healthcare, the use of these new technologies
may also introduce significant new unanticipated risks to patients.
•
Also of concern is that even when high quality interventions are
developed, they frequently fail to live up to their potential when deployed in
the “real world”; the reasons underpinning these failures remain poorly
understood, but include amongst other things, professional resistance to
change, poorly designed applications and the resulting disruption to
clinical practices and care provision.
36
Summary contd…
•
Given that the NHS has embarked on one of the largest eHealth-based
modernisation programmes in the world, it is appropriate and timely to
critically review the international eHealth literature with a view to identifying
lessons that can be learned with respect to the future development,
design, deployment and evaluation of eHealth applications.
1.1 Background to this study and policy context
The Institute of Medicine has defined quality as the ‘…degree to which health
services for individuals and populations increase the likelihood of desired
health outcomes and are consistent with current professional knowledge.’1
The quality of services provided by the National Health Service (NHS) in
England varies widely, and there is often a large gap between the optimal
standard of services, when judged based on professionally agreed criteria,
and the actual quality of practice.2 3
This quality gap can have serious health consequences and major
implications for the intimately inter-related notion of patient safety.4 These
deficiencies can manifest, for example, as misdiagnosis,5 medication errors,6 7
increased rates of complications in patients with long-term diseases,8 hospital
admissions for adverse drug reactions,9
10
and harm in already hospitalised
patients.11
This gap between optimal and actual levels of healthcare also has large
financial costs for the healthcare system, national governments, and society,
as well as affecting patients' quality of life and life expectancy.10 12 13 Problems
with the quality of care can have a considerable impact on healthcare costs;
for example, through increased hospital admission rates for unscheduled care
and/or prolonged hospital stay.10 Problems with the safety of healthcare can
similarly have major cost implications; for example, adverse drug reactions in
the UK are estimated to result in approximately 250,000 hospital admissions
37
per year9 10 at a total cost of about £500 million (€700 million; $1,008 million).9
10
Mounting evidence of the disease burden associated with variations in
standards of care, especially with regard to medical errors, has informed a
number of recent national policy and strategic reports on patient safety.14-23
These documents have highlighted the pressing need for healthcare policy to
focus more clearly on developing systems—and strategies for their
implementation—that facilitate the delivery of safe and effective high quality
healthcare. This response being catalysed by the identification of a series of
failings, both with respect to policy and approaches to strategically implement
these healthcare reforms.24
There are also other relevant factors, including: the fact that economicallydeveloped (and indeed also transition and developing) countries are facing
increases in the proportions and number of older people in their populations;
increasing public expectations about rapid access to high quality patientcentred health services; and an expectation from governments, professionals
and the public alike that people should be able to live with the minimum
possible impact from disease and disability on their quality of life and day-today activities.15
Capitalising on the information technology (IT) revolution is increasingly seen
as pivotal to redesigning healthcare systems so that they are able to deliver
safe, effective and convenient healthcare. Such objectives lay at the heart of
the NHS’s Information Strategy and led to the creation of structures such as
the National Patient Safety Agency (NPSA),25 the National Reporting and
Learning System (NRLS),26 NHS Connecting for Health (NHS CFH) and its
National Programme for Information Technology (NPfIT) and the more recent
creation of the Department of Health’s (DH) Informatics Directorate (see
Chapter 3 for a detailed discussion of NHS CFH, NPfIT and the DH
Informatics Directorate).
38
1.2 Agenda for system redesign: quality and safety in healthcare
The Department of Health (DH) estimates that 10 per cent of patients
admitted to NHS hospitals are unintentionally harmed, this high rate being
similar to that in other developed countries. If lessons had been learnt from
previous incidents, around 50 per cent of these patient safety incidents could
have been avoided.17 In its report An Organisation with a Memory, an expert
committee, chaired by the then Chief Medical Officer Sir Liam Donaldson,
identified four conditions that the NHS must meet to learn effectively from
failures and offer the best possible protection to patients in the future.14 These
relate to the development of:
•
a more open culture, in which clinical errors or failures of service
delivery can be reported and discussed by NHS staff
•
unified mechanisms for reporting and analysing patient safety incidents
in healthcare delivery
•
mechanisms to ensure that, where lessons are identified, the
necessary changes to improve the delivery of healthcare are put into
practice
•
a much wider appreciation of the value of system-based approaches in
preventing, analysing and learning from errors.
These requirements called for complex and comprehensive solutions as no
single “magic bullet” will be sufficient.15 As an example of the role of IT in
these endeavours, evidence-based technological solutions may provide
significant support for the development of standardised incident reporting and
analysis mechanisms, as well as providing a platform for automated
interventions to ensure that lessons learned from past mistakes are
incorporated into daily practice.26 Such interventions also demonstrate the
system-based approach called for in a number of policy documents.
1.3 Importance of technology in designing safe healthcare systems
Enthusiasm about the potential that IT offers for improving health services has
resulted in unprecedented investments into IT by the NHS (and other health
systems internationally).
Technology-based health services and clinical
39
systems come in many different forms, have myriad aims, and can be
implemented in numerous ways. Information technologies (sometimes also
referred to as information and communication technology (ICT)), health
informatics or eHealth applications, the latter being our preferred term in this
report) may be used to record, collate and share information on patients (e.g.
electronic health records (EHRs)). eHealth applications may also be used to
support evidence-based practice both in general terms (e.g. guideline-linked
reminders)27 or
more specifically
individual patients (e.g.
through advice on the management of
advice on avoiding drug-drug interactions).28
29
eHealth applications can also serve to facilitate care from a distance30 (e.g.
telehealthcare),
patient
self-care
(e.g.
through
consumer
informatics
applications), epidemiological and quasi-experimental research (e.g. using
databases that collate data derived from electronic health records)31 and
healthcare management activities such as quality improvement initiatives.32 33
As discussed in subsequent chapters, many of these technology-based
approaches
are
vigorously
being
pursued
both
in
the
NHS
and
internationally.34 35
eHealth solutions range from computer-assisted history taking systems
(CAHTS),36-42
computerised
ePrescribing,44-52
decision
computerised
support
medication
systems
administration
(CDSSs),43
records,53
automated pharmacy systems and bar-coding to point-of-care personal digital
assistants, robot-assisted surgery, telehealthcare and “smart” homes.54-60
Because of the often complex nature of these systems, there are important
considerations relating to appropriately evaluating their impact and we reflect
on these issues in more detail in Chapters 2 and 20. The potential offered by
these eHealth applications is considered in Chapter 4 and the empirical
evidence in relation to a number of key exemplary eHealth applications is
summarised in Chapters 6–18.
The desire to realise the potential of technology to improve the quality and
safety of healthcare often leads to the introduction of new IT solutions, on a
large or even national scale, with typically only limited evidence in support of
the systems’ overall effectiveness and safety. Of particular relevance here is
40
that the hardware-software development mismatch often creates unrealistic
expectations about the likely impact of eHealth systems. Processing power
doubles approximately every 18 months and storage capacity and network
bandwidth also increase rapidly within relatively short time periods.
In
contrast, productivity in software development has been increasing much
more slowly. The Chaos studies performed by the Standish Group have, for
example, shown that the majority of software projects (across all industries)
are delivered late and over budget.61 Specifically, software applications often
become a bottleneck for healthcare organisations trying rapidly to improve
their processes, this despite substantial investments in IT.62 To meet the
expectations of the market, systems’ developers have thus far been more
concerned with meeting design briefs and satisfaction than with assessing the
efficacy and effectiveness of systems in terms of producing actual health
gains. Consequently, many eHealth innovations remain either unstudied or
under-studied. Indeed, when evaluated, the evidence on the question of their
effectiveness is often inconsistent, frequently reflecting, as we will see, a lack
of attention to issues pertaining to context, which is so crucial to success.63-68
Researchers have paid even less attention to patient safety (i.e. the potential
for harmful effects), as demonstrated by recent work in the UK evaluating the
safety profile of general practice (GP) computing software in providing
decision support for prescribing.45 69 Other areas of relevance in the context of
both clinical effectiveness and safety are user acceptance, interoperability,
workflow integration, compatibility with legacy applications, and system
maturity.70 71
Emerging eHealth systems have the potential to reduce errors and enhance
patient safety by, amongst other things, improving the legibility of clinical
communications, enabling shared access to health records, improving access
to care, reducing reliance on human memory and prompting evidence-based
prescribing.33
72-75
For example, automated monitoring and routine feedback
have been shown to reduce the rate at which hospitalised adults with renal
insufficiency received excessive doses of medication.76
41
However, such systems also have the potential to introduce risks and
compromise quality as demonstrated in Chapter 4, where we map the key
health risks associated with eHealth solutions.
For example, a study
examining the types of errors facilitated by a leading ePrescribing system
found that it regularly increased the risk of a variety of errors, with users
describing up to 22 different types of failures.77 Similarly, high rates of adverse
drug events have been found in a highly computerised hospital.78 Although
improved case finding (in part resulting from IT use) may partially explain
these high rates, the overall rate of adverse drug events (related to problems
in drug selection, dosage, and monitoring) per 1,000 patient-days was 5 to 19
times higher than that reported in previous studies. These examples illustrate
the point that, as new IT solutions are conceived and implemented, there is
also a need to develop mechanisms proactively to study the potential for
these systems to inadvertently cause harm.79-84
As with many innovative ventures, not all ideas ultimately prove beneficial.
Even when safe, some 50 to 70 per cent of eHealth projects fail.85 These
failures do not in the majority of cases represent technological problems, but
rather human and organisational factors.
Problems may arise due to
developers misjudging the functioning of health systems or not envisaging the
full range of consequences. Even when the ideas prove effective in controlled
research settings, they often challenge prevailing thinking and behaviour in a
way that renders them difficult to implement.86 87 There is ample evidence to
show that the past failures of technological innovations with respect to
improving health outcomes (or to drive improvements in other industries)88-90
have not necessarily been due to their clinical ineffectiveness, but rather to
social, technological and cultural issues relating to their implementation and
adoption.91
In particular, the human elements affecting the success of
technological implementation and adoption should not be underestimated,
whether these relate to organisational issues, e.g. managing complexity92—
complex systems involve many gaps between people, stages, and
processes;93 and or availability, accessibility and user attitudes towards, or
usability of the technology.70
87 94-100
Furthermore, integration of IT systems
into clinical settings fundamentally may change not only how clinicians view
42
their daily work practices, but also the very process of medical reasoning
itself, introducing new elements into care pathways, often with unpredictable
effects.101 102
Any significant adverse events due to a failure of new eHealth solutions have
the potential to lead to disastrous effects in the way in which these
programmes are perceived by professionals and the public alike.
It is
therefore vital to ensure that implementation and adoption considerations are
given at least equal importance as questions of IT effectiveness and safety.
Successfully integrating new IT solutions into the healthcare workflow is
crucially dependant on engagement of clinicians and patients right from the
start of the development and ongoing evaluation of these new applications. In
addition to providing important insights into user requirements, such
engagement leads to a sense of ownership and buy-in that is critical to the
success of the often radical programmes to transform health services that are
currently underway in different parts of the world.87
103
These socio-techno-
cultural considerations are reviewed in Chapters 16-18.
In summary, key challenges currently facing the DH’s/NHS CFH’s ambitious
NPfIT and the NHS quality and patient safety agendas include ensuring that
new eHealth solutions developed for and introduced into the NHS represent a
genuine clinical advantage, that their deployment is, wherever possible,
supported by a secure evidence-base, and that evaluative mechanisms are in
place to assess their effects with regard to both quality and safety of health
service delivery. As a recent Institute of Medicine report highlighted, patient
safety must be indistinguishable from the delivery of quality care23 and must
score top on our agendas.104
1.4 Aims of this review
This report, which builds on our scoping work,33 completed and ongoing
systematic reviews,105-110 both quantitative and qualitative studies on eHealth
approaches to enhancing quality and minimising threats to patient safety,
aims to provide a systematic overview of the impact of eHealth on the safety
and quality of healthcare. In so doing, we seek to identify, critically appraise
43
and then synthesise evidence to develop a theoretically informed framework
for understanding the potential and empirically demonstrated benefits and
risks associated with eHealth and to describe and understand approaches
that can be used to guide the successful implementation and adoption of
effective interventions. We also aim to identify areas of eHealth in need of
further study, highlighting possible approaches that might successfully be
employed to ensure a well-rounded, context-bound appreciation of the overall
impact of IT interventions in healthcare.
1.5 Structure of this report
After this introductory chapter, Section 2 begins with a detailed description of
our methods (Chapter 2) and this is followed by the findings of our attempts at
understanding the history, development and policy foci of the DH/NHS CFH
and NPfIT (Chapter 3) and describing and mapping the fields of eHealth,
quality and safety (Chapter 4).
In Section 3 we present our main findings, beginning with the essential
underpinning
issues
of
health
information
exchange
capability
and
interoperability (Chapter 5), and then moving on to reviews of key eHealth
applications, namely EHRs (Chapter 6), CAHTS (Chapter 7), CDSSs (Chapter
8), ePrescribing (Chapter 10) and telehealthcare (Chapter 13). We then turn
to the cross-cutting issues of how to design safe, effective and usable
systems (Chapter 16) and how then to ensure their effective implementation
and adoption into routine models of care (Chapter 17).
In each of these
chapters we begin by considering key definitional issues and the relevant
theoretical literature with a view to having a firm foundation from which to
interpret the empirical evidence; the chapters conclude with a reflection of key
policy, practice and research implications arising from this work. We then
reflect on this evidence in considerably more detail in the context of applying it
to five exemplary case studies looking at important clinical and policy
questions in the remaining chapters in this section (Chapters 9, 11, 12, 14, 15
and 18).
44
Our conclusions and policy recommendations from this review are presented
in Chapter 19. Section 4 concludes with our reflections on the quality of the
evidence reviewed and overarching suggestions for future research which
should, if pursued, help to improve both the quality and quantity of empirical
evidence for eHealth applications and their successful deployment (Chapter
20).
The final section contains a number of appendices relating to the methods
employed (Appendices 1-3), detailed findings (Appendices 4 and 5) and
outputs from this work (Appendices 6 and 7). This section also includes a
glossary of key terms.
1.6 Conclusions
As a result of the work carried out for this report, we have, we believe, made a
number of key conceptual and methodological advances in relation to
techniques for identifying, appraising, synthesising and interpreting the
evidence. Furthermore, our work has laid the foundation for the creation of a
comprehensive, critically appraised, and carefully interpreted database of
systematic reviews, which should, once fully developed, be of international
interest to all those with an interest in eHealth and its impact on healthcare
delivery. We hope that the theoretical and empirical evidence presented in
this report, when viewed in the context of ongoing policy developments,
provides a reliable and accessible overview of a vast corpus of hitherto
disparate knowledge and through so doing helps to inform on-going
deliberations on how to ensure that eHealth realises its enormous potential to
impact positively on both the quality and safety of healthcare delivery.
References
1. Institute of Medicine. Crossing the Quality Chasm: A New Health System for the
21st Century. Washington, DC: National Academy Press, 2001.
2. Klein R. Britain's National Health Service revisited. N Engl J Med
2004;350:937-942.
3. Martin RM, Sterne JAC, Gunnell D, Ebrahim S, vey Smith G, Frankel S. NHS
waiting lists and evidence of national or local failure: analysis of health service
data. BMJ 2003; 326: 188.
45
4. Institute of Medicine. To Err Is Human: Building a Safer Health System. .
Washington, DC: National Academy Press, 2000.
5. Chadwick D, Smith D. The misdiagnosis of epilepsy. BMJ 2002; 324: 495-96.
6. Wong IC, Ghaleb MA, Franklin BD, Barber N. Incidence and nature of dosing
errors in paediatric medications: a systematic review. Drug Saf 2004; 27:66170.
7. Snijders C, van Lingen RA, Molendijk A, Fetter WPF. Incidents and errors in
neonatal intensive care: a review of the literature. Arch Dis Child Fetal Neonatal
Ed 2007; 92: F391-98.
8. Alexander KP, Chen AY, Roe MT, Newby LK, Gibson CM, len-LaPointe NM et
al. Excess dosing of antiplatelet and antithrombin agents in the treatment of
non-ST-segment elevation acute coronary syndromes. JAMA 2005; 294: 310816.
9. Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ et al.
Adverse drug reactions as cause of admission to hospital: prospective analysis
of 18 820 patients. BMJ 2004; 329: 15-19.
10. Davies EC, Green CF, Mottram DR, Rowe PH, Pirmohamed M. Emergency readmissions to hospital due to adverse drug reactions within 1 year of the index
admission. Br J Clin Pharmacol 2010; 70: 749-55.
11. Davies EC, Green CF, Taylor S, Williamson PR, Mottram DR, Pirmohamed M.
Adverse drug reactions in hospital in-patients: a prospective analysis of 3695
patient-episodes. PLoS One 2009; 4: e4439.
12. Institute of Medicine. Preventing Medication Errors. 2007.
13. Institute of Medicine. Patient Safety: Achieving a New Standard for Care.
Washington, DC: National Academy Press, 2003.
14. Department of Health. An Organisation with a Memory: Report of an expert
group of learning from adverse events in the NHS. London: HMSO, 2000.
15. Department of Health. Equity and Excellence: Liberating the NHS. London:
HMSO, 2010.
16. Department of Health, Chief Pharmaceutical Officer. Building a Safer NHS for
Patients: Improving Medication Safety. London: HMSO, 2004.
17. National Audit Office. Department of Health. A Safer Place for Patients:
Learning to Improve Patient Safety. London: Stationery Office, 2005.
18. National Audit Office. Safety, Quality, Efficacy: Regulating Medicines in the UK.
London: The Stationery Office, 2003.
19. National Audit Office. A Comparison of International Practices in the
Management and Control of Hospital-acquired Infections. HC 876. London,
National Audit Office, 2003.
20. Department of Health. Safety First: A Report for Patients, Clinicians and
Healthcare Managers. London: HMSO, 2006.
21. NPSA. Safety in Doses. London: NRLS, 2007.
22. Department of Health. Saving Lives. London: HMSO, 1999.
23. WHO. World Alliance for Patient Safety: Forward Programme 2008-2009.
WHO, 2008.
24. Berwick DM. Not again! BMJ 2001;322:247-248.
25. NHS NPSA. National Patient Safety Agency. NHS National Patient Safety
Agency Available from: http://www.npsa.nhs.uk/ (last accessed 1/1/11)
26. NHS NPSA. Reporting Incidents. NHS NPSA Available from:
http://www.npsa.nhs.uk/patientsafety/reporting (last accessed 1/1/11).
27. Shiffman RN, Liaw Y, Brandt CA, Corb GJ. Computer-based guideline
implementation systems: a systematic review of functionality and effectiveness.
J Am Med Inform Assoc 1999; 6: 104-14.
28. Holdsworth MT, Fichtl RE, Raisch DW, Hewryk A, Behta M, Mendez-Rico E et
al. Impact of computerized prescriber order entry on the incidence of adverse
drug events in pediatric inpatients. Pediatrics 2007;120: 1058-66.
46
29. Kuperman GJ, Bobb AM, Payne T, Avery AJ, Gandhi TK, Burns G et al.
Medication-related clinical decision support in computerized provider order
entry systems: a review. J Am Med Inform Assoc 2007. 14; 29-40.
30. Hersh WR, Hickam DH, Severance SM, Dana TL, Pyle KK, Helfand M.
Diagnosis, access and outcomes: update of a systematic review of
telemedicine services. J Telemed Telecare 2006; 12 :S3-31.
31. Pringle M. Using computers in general practice research. Practitioner 1986;
230: 635-39.
32. Pagliari C, Sloan D, Gregor P, Sullivan F, Detmer D, Kahan JP et al. What is
eHealth (4): a scoping exercise to map the field. J Med Internet Res 2005; 7:
e9.
33. Pagliari C, Sloan D, Gregor P, Sullivan F, Kahan JP, Detmer D et al. Literature
Review and Conceptual Map of the Field of eHealth. Final report to the SDO
programme. 2004.
34. Department of Health. Health Informatics Review. London: HMSO, 2010.
35. Department of Health. Liberating the NHS: An Information Revolution. London:
HMSO, 2010.
36. Lilford RJ, Bingham P, Fawdry R, Setchell M, Chard T. The development of online history-taking systems in antenatal care. Methods Inf Med 1983; 22: 18997.
37. Lilford RJ, Glyn-Evans D, Chard T. The use of a patient-interactive
microcomputer system to obtain histories in an infertility and gynecologic
endocrinology clinic. Am J Obstet Gynecol 1983; 146: 374-79.
38. Lilford RJ, Bingham P, Bourne GL, Chard T. Computerized histories facilitate
patient care in a termination of pregnancy clinic: the use of a small computer to
obtain and reproduce patient information. Br J Obstet Gynaecol 1985; 92: 33340.
39. Lilford RJ, Chard T, Bingham P, Carrigan E. Use of a microcomputer network
for history taking in a prenatal clinic. Am J Perinatol 1985; 2: 143-47.
40. Lilford RJ. Comparisons between written and computerised patient histories. Br
Med J 1987; 295: 503.
41. Lilford RJ, Guthrie K, Kelly M. History-taking by computer. Baillieres Clin Obstet
Gynaecol 1990; 4: 723-42.
42. Pringle M. Using computers to take patient histories. BMJ 1988; 297: 697-98.
43. Bates DW, Kuperman GJ, Wang S, Gandhi T, Kittler A, Volk L et al. Ten
commandments for effective clinical decision support: making the practice of
evidence-based medicine a reality. J Am Med Inform Assoc 2003; 10: 523-30.
44. Avery AJ, Sheikh A, Hurwitz B, Smeaton L, Chen YF, Howard R et al. Safer
medicines management in primary care. Br J Gen Pract 2002; 52: S17-22.
45. Avery AJ, Savelyich BS, Sheikh A, Cantrill J, Morris CJ, Fernando B et al.
Identifying and establishing consensus on the most important safety features of
GP computer systems: e-Delphi study. Inform Prim Care 2005; 13: 3-12.
46. Barber N, Franklin B, Cornford T, Klecun E, Savage I. Safer, Faster, Better?
Evaluating
Electronic
Prescribing.
2006.
Available
from:
http://www.who.int/patientsafety/information_centre/reports/PS019_Barber_Fin
al_report.pdf (last accessed 1/1/11).
47. Barber N. Designing information technology to support prescribing decision
making. Qual Saf Health Care 2004; 13: 450-54.
48. Barber N, Cornford T, Klecun E. Qualitative evaluation of an electronic
prescribing and administration system. Qual Saf Health Care 2007; 16: 271-78.
49. Bates DW, Evans RS, Murff H, Stetson PD, Pizziferri L, Hripcsak G. Policy and
the future of adverse event detection using information technology. J Am Med
Inform Assoc 2003; 10: 226-28.
47
50. Bates DW, Evans RS, Murff H, Stetson PD, Pizziferri L, Hripcsak G. Detecting
adverse events using information technology. J Am Med Inform Assoc 2003;
10: 115-28.
51. Bates DW. Computerized physician order entry and medication errors: finding a
balance. J Biomed Inform 2005; 38: 259-61.
52. Bates DW. Preventing medication errors: a summary. Am J Health Syst Pharm
2007; 64: S3-S9.
53. Franklin BD, O'Grady K, Donyai P, Jacklin A, Barber N. The impact of a closedloop electronic prescribing and administration system on prescribing errors,
administration errors and staff time: a before-and-after study. Qual Saf Health
Care 2007; 16: 279-84.
54. Bates DW, Cohen M, Leape LL, Overhage JM, Shabot MM, Sheridan T.
Reducing the frequency of errors in medicine using information technology. J
Am Med Inform Assoc 2001; 8: 299-308.
55. Schiff GD, Kim S, Krosnjar N, Wisniewski MF, Bult J, Fogelfeld L et al. Missed
hypothyroidism diagnosis uncovered by linking laboratory and pharmacy data.
Arch Intern Med 2005; 65: 574-77.
56. Bates DW, Gawande AA. Improving safety with information technology. N Engl
J Med 2003; 348: 2526-34.
57. Patterson ES, Cook RI, Render ML. Improving patient safety by identifying side
effects from introducing bar coding in medication administration. J Am Med
Inform Assoc 2002; 9: 540-53.
58. Kaushal R, Barker KN, Bates DW. How can information technology improve
patient safety and reduce medication errors in children's health care? Arch
Pediatr Adolesc Med 2001; 155: 1002-07.
59. Garg AX, Adhikari NKJ, McDonald H, Rosas-Arellano MP, Devereaux PJ,
Beyene J et al. Effects of computerized clinical decision support systems on
practitioner performance and patient outcomes: a systematic review. JAMA
2005; 293: 1223-38.
60. Durieux P. Electronic medical alerts -- so simple, so complex. N Engl J Med
2005; 352: 1034-036.
61. The
Standish
Group.
CHAOS:
A
recipe
for
success.
http://www.standishgroup.com/ (last accessed 1/1/11).
62. Stead WW, Kelly BJ, Kolodner RM. Achievable steps toward building a national
health information infrastructure in the United States. J Am Med Inform Assoc
2005; 12: 113-20.
63. Kaushal R, Bates DW. Information technology and medication safety: what is
the benefit? Qual Saf Health Care 2002; 11: 261-65.
64. Wears RL, Berg M. Computer technology and clinical work: still waiting for
Godot. JAMA 2005; 293: 1261-63.
65. Koppel R, Metlay JP, Cohen A, Abaluck B, Localio AR, Kimmel SE et al. Role
of computerized physician order entry systems in facilitating medication errors.
JAMA 2005; 293: 1197-1203.
66. Kuperman GJ, Gibson RF. Computer physician order entry: benefits, costs, and
issues. Ann Intern Med 2003; 139: 31-39.
67. Adhikari N. Medical informatics in the intensive care unit: overview of
technology assessment. J Crit Care 2003; 18: 41-47.
68. Berg M. Health Information Management: Integrating Information Technology in
Health Care Work. London, England: Routledge; 2004.
69. Morris CJ, Savelyich BSP, Avery AJ, Cantrill JA, Sheikh A. Patient safety
features of clinical computer systems: questionnaire survey of GP views. Qual
Saf Health Care 2005; 14: 164-68.
70. Robertson A, Cresswell K, Takian A, Petrakaki D, Crowe S, Cornford T, et al.
Implementation and adoption of nationwide electronic health records in
48
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
secondary care in England: qualitative analysis of interim results from a
prospective national evaluation. BMJ 2010; 341: c4564.
Fox J, Thomson R. Clinical decision support systems: a discussion of quality,
safety and legal liability issues. Proc AMIA Symp 2002; 265-69.
Protti D. Comparison of information technology in general practice in 10
countries. Healthc Q 2007;10: 107-16.
Sullivan F, Wyatt JC. How decision support tools help define clinical problems.
BMJ 2005; 331: 831-33.
Sullivan F, Wyatt JC. How informatics tools help deal with patients' problems.
BMJ 2005; 331: 955-57.
Sullivan F, Wyatt JC. How computers help make efficient use of consultations.
BMJ 2005; 331: 1010-12.
Nash IS, Rojas M, Hebert P, Marrone SR, Colgan C, Fisher LA et al. Reducing
excessive medication administration in hospitalized adults with renal
dysfunction. Am J Med Qual 2005; 20: 64-69.
Koppel R, Metlay JP, Cohen A, Abaluck B, Localio AR, Kimmel SE et al. Role
JAMA 2005; 293: 1197-1203.
Nebeker JR, Hoffman JM, Weir CR, Bennett CL, Hurdle JF. High rates of
adverse drug events in a highly computerized hospital. Arch Intern Med 2005;
165: 1111-16.
Strom BL, Schinnar R, Aberra F, Bilker W, Hennessy S, Leonard CE, et al.
Unintended effects of a computerized physician order entry nearly hard-stop
alert to prevent a drug interaction: a randomized controlled trial. Arch Intern
Med 2010; 170: 1578-83.
Battles JB, Lilford RJ. Organizing patient safety research to identify risks and
hazards. Qual Saf Health Care 2003; 12: 2ii-7.
McNutt RA, Abrams R, Aron DC, for the Patient Safety Committee. Patient
safety efforts should focus on medical errors. JAMA 2002; 287: 1997-2001.
Overhage JM, Tierney WM, Zhou XH, McDonald CJ. A randomized trial of
"corollary orders" to prevent errors of omission. J Am Med Inform Assoc 1997;
4: 364-75.
Gillam M, Rothenhaus T, Smith V, Kanhouwa M. Information technology
principles for management, reporting, and research. Acad Emerg Med 2004;
11: 1155-61.
Bates DW, Teich JM, Lee J, Seger D, Kuperman GJ, Ma'Luf N et al. The impact
of computerized physician order entry on medication error prevention. J Am
Med Inform Assoc 1999; 6: 313-21.
Lorenzi NM, Smith JB, Conner SR, Campion TR. The success factor profile for
clinical computer innovation. Medinfo 2004; 11: 1077-80.
Stryer D, Clancy C. Patients' safety. BMJ 2005; 330: 553-54.
Greenhalgh T, Robert G, Bate P, Kyriakidou O, Macfarlane F, Peacock R. How
to Spread Good Ideas: A systematic review of the literature on diffusion,
dissemination and sustainability of innovations in health service delivery and
organisation. Report for the National Co-ordinating Centre for NHS Service
Delivery and Organisation R & D. 1-424. London, The National Co-ordinating
Centre for NHS Service Delivery and Organisation R & D (NCCSDO), 2004.
Neumann PG. Computer-Related Risks. New York, NY: ACM Press; 1995.
Eason K. Local sociotechnical system development in the NHS National
Programme for Information Technology. J Inform Technol 2007; 22: 257–64.
Rochlin GI. Trapped in the Net: The Unanticipated Consequences of
Computerization. Princeton, NJ: Princeton University Press; 1997.
Tenner E. Why Things Bite Back: Technology and the Revenge of Unintended
Consequences. New York, NY: Vintage Books; 1997.
49
92. Plsek PE, Wilson T. Complexity science: Complexity, leadership, and
management in healthcare organisations. BMJ 2001; 323: 746-49.
93. Cook RI, Render M, Woods DD. Gaps in the continuity of care and progress on
patient safety. BMJ 2000; 320: 791-94.
94. Lorenzi NM, Riley RT, Blyth AJC, Southon G, Dixon BJ. Antecedents of the
people and organizational aspects of medical informatics: Review of the
Literature. J Am Med Inform Assoc 1997; 4: 79-93.
95. Silver MP, Geis MS, Bateman KA. Improving health care systems performance:
a human factors approach. Am J Med Qual 2004; 19: 93-102.
96. Stryer D, Clancy C. Patients' safety. BMJ 2005; 330: 553-54.
97. Ball MJ, Douglas JV. Redefining and improving patient safety. Methods Inf Med
2002; 41: 271-76.
98. Short D, Frischer M, Bashford J. Barriers to the adoption of computerised
decision support systems in general practice consultations: a qualitative study
of GPs' perspectives. Int J Med Inf 2004; 73: 357-62.
99. Kaplan B, Farzanfar R, Friedman RH. Personal relationships with an intelligent
interactive telephone health behavior advisor system: a multimethod study
using surveys and ethnographic interviews. Int J Med Inf 2003; 71: 33-41.
100. Catwell L, Sheikh A. Information technology (IT) system users must be allowed
to decide on the future direction of major national IT initiatives. But the task of
redistributing power equally amongst stakeholders will not be an easy one.
Inform Prim Care 2009; 17: 1-4.
101. Patel VL, Kaufman DR. Science and practice: a case for medical informatics as
a local science of design. J Am Med Inform Assoc: 5: 489-92.
102. Hunt DL, Haynes RB, Hanna SE, Smith K. Effects of computer-based clinical
decision support systems on physician performance and patient outcomes: a
systematic review. JAMA 1998; 280: 1339-46.
103. Morrison Z, Robertson A, Cresswell K, Crowe S, Sheikh A. Comparing and
contrasting three approaches to nationwide implementations of electronic
health records: England, the USA and Australia. J Hlthcare Eng (in press).
104. Wilson T, Pringle M, Sheikh A. Promoting patient safety in primary care. BMJ
2001; 323: 583-84.
105. Wilson T, Sheikh A. Enhancing public safety in primary care. BMJ 2002; 324:
584-87.
106. Car J, Sheikh A. Email consultations in health care: 2--acceptability and safe
application. BMJ 2004; 329: 439-42.
107. Car J, Sheikh A. Email consultations in health care: 1--scope and effectiveness.
BMJ 2004; 329: 435-38.
108. McLean S, Chandler D, Nurmatov U, Liu J, Pagliari C, Car J, Sheikh A.
Telehealthcare for asthma. Cochrane Database Syst Rev 2010; CD007717.
109. Civljak M, Sheikh A, Stead LF, Car J. Internet-based interventions for smoking
cessation. Cochrane Database Syst Rev 2010; CD007078.
110. Gagnon MP, Légaré F, Labrecque M, Frémont P, Pluye P, Gagnon J, et al.
Interventions for promoting information and communication technologies
adoption in healthcare professionals. Cochrane Database Syst Rev 2009;
CD006093.
50
SECTION 2
BACKGROUND &
FORMATIVE WORK
51
Chapter 2
Methods
Summary
•
Drawing on our own systematic reviews and those conducted by others,
we conducted a systematic search and critique of the empirical literature
on eHealth applications and their impact on the quality and safety of
healthcare delivery; this body of work was then synthesised with relevant
theoretical, technical, developmental and policy relevant literature with a
view to producing an authoritative and accessible overview of the field.
•
Whilst
we
drew
on
established
Cochrane
review
principles
to
systematically search for, critique and synthesise the literature, this
approach needed to be adapted in several respects in order to produce a
meaningful “overview” of the literature.
•
Searching the literature was complicated by the lack of internationally (or
indeed in some cases nationally) agreed terminology relating to eHealth
applications, the lack of agreed definitions of quality and safety, and
consequently poor indexing of these constructs in databases of published
work.
•
In order to undertake a comprehensive review of the literature we
therefore needed to undertake initial developmental work to formulate
highly sensitive search strategies.
•
Using this comprehensive set of search terms, we systematically
searched: MEDLINE, EMBASE, The Cochrane Database of Systematic
Reviews, Database of Abstracts of Reviews of Effects, The Cochrane
Central Register of Controlled Trials, The Cochrane Methodology
Register, Health Technology Assessment Database, NHS Economic
Evaluation Database over a 13-year period (beginning with 1997 to
October 2010) to identify systematic reviews, technical reports and health
technology assessments, and randomised controlled trials investigating
the impact of eHealth applications on the quality and safety of healthcare.
We also searched databases of unpublished work to identify relevant ongoing and unpublished work.
52
Summary continued…
•
To provide a broader appreciation of the context to this work and
furthermore to aid development and interpretation of findings, we built our
interpretation of evidence around relevant theoretical literature. We used
a more emergent approach to identify relevant background literature in
relation to the core concepts underpinning this overview—namely eHealth,
quality, safety and the National Programme for Information Technology.
This involved drawing on our personal databases of relevant papers,
identifying seminal papers and reports, and searching the grey literature.
•
We integrated the findings from this broader theoretical and empirical
evidence-base in relation to each of the domains of enquiry.
•
The overall body of literature identified was too diverse to make any
meaningful quantitative synthesis of the literature desirable, nor was it
possible.
Rather, we chose to synthesise the literature qualitatively
drawing on the relevant preliminary conceptual work to guide this narrative
synthesis.
•
Our assessment of individual reviews is summarised in Appendix 6 using
the modified Critical Appraisal Skills Programme criteria; the overall
assessment of the volume and strength of evidence in relation to key
findings is summarised in the Executive Summary using a modified
version of the World Health Organization’s Health Evidence Network
system for public health evidence, which grades evidence into three main
categories:
o strong (consistent, good quality, plentiful or generalisable)
o moderate (consistent and good quality)
o limited to none (inconsistent or poor quality).
53
2.1 Outline of methods
Research in the field of eHealth is exponentially expanding in breadth and
increasing in volume thus rendering it impossible for any individual to read,
critically appraise and synthesise the state of current knowledge, let alone
remain up-to-date in this dynamic area. Systematic reviews have become
essential sources of information for policy-makers and practitioners who need
to remain abreast of important new developments in their fields of interest.
These also serve another important function, namely helping to identify areas
where there are important gaps in the literature and where further primary
studies are required.
Our aim in conducting this systematic overview of the literature was to survey
the relevant theoretical, technical and empirical literature to produce a
comprehensive summary of the evidence for eHealth applications to improve
the quality and safety of healthcare delivery, and identify key issues in relation
to the design, development, implementation and adoption of eHealth
applications into routine healthcare settings. More specifically, in keeping with
our brief we were initially interested in two broad areas of eHealth
applications, these relating to technologies concerned with improving the
storage and management of patient data and tools that support professional
decision-making; the scope of enquiry was extended in the second phase of
our work to include tools to support the delivery of healthcare from a distance
(see Chapter 4).
Given the extensive body of eHealth literature, we considered various
approaches to searching, appraising, interpreting and synthesising the
evidence. Considering the aim of this report and in order to make our findings
accessible to policy-makers, managers and end-users,1 We throughout
attempted to generate new insights and pragmatic recommendations suitable
for effective policy-making in healthcare.
Based on our recent experiences of conducting systematic reviews in the
subject area, we considered it important to take a broad approach to
identifying potentially relevant work in this systematic overview.2 This
54
approach has the advantage of allowing the strength (i.e. consistency and
generalisability) of research findings to be assessed across a wide range of
different settings, study populations, and behaviours, thereby reducing the
risks of drawing erroneous conclusions.
Established systematic review
methods were used to identify and critically appraise studies and then
abstract and synthesise findings.
We explicitly sought to factor into our
assessment important contextual factors. Detailed methods for the de novo
systematic reviews we conducted in the second phase of work are detailed in
the relevant publications (see Appendix 7).
2.1.1 Literature reviews of other sources of evidence
Standard systematic review methods (e.g. Cochrane), which focus on
identifying comparable studies addressing specific research topics and
combining their results to establish general effect sizes, were too narrow for
the purposes of meeting the objective of research brief.3-5 While evidence
collated from such reviews appropriately carried greater weight in relation to
assessing questions of effectiveness, our interpretation of evidence was also
informed by other research architectures or designs, primarily randomised
controlled trials (RCTs) and other experimental designs at relatively low risk of
bias,6 but also case series, instructive case reports, and conceptual papers to
inform safety considerations.
Systematic reviews of RCTs are, on account of their ability to control for
known and unknown confounders, the “gold standard” evidence source in
relation to studying the effectiveness of interventions.
Whilst RCTs and
systematic summaries of these are ideally suited for studying drug treatments,
they are typically far less appropriate for studying safety issues, a major
problem being the prohibitively large RCTs that are typically needed to study
adverse events. A meaningful study of safety considerations must therefore
embrace a broader literature than that needed to establish efficacy or
effectiveness. Randomised controlled trials are furthermore unable to shed
detailed insights on whether systems will be used or indeed how they will be
used—factors which greatly influence the effectiveness (as opposed to
efficacy) of interventions when implemented in routine practice.
As the
55
commissioning brief advised a focus on both quality and safety issues of IT
deployment in healthcare services, we therefore needed to develop a novel
combination of approaches to knowledge acquisition and evidence appraisal
to ensure that we produced a valid summary of the literature.
We began evidence synthesis by drawing primarily on high quality evidence
from systematic reviews. Where these did not provide adequate answers to
the questions of interest, we employed a “snowball” approach to identifying
additional potentially relevant literature, starting our additional searches for
and examination of evidence from RCTs and, as necessary and feasible, then
also considering other study designs (e.g. controlled before-after designs,
time-motion, descriptive and qualitative studies), as well as relevant
theoretical and technical papers.6 7 Since the field is rapidly developing, a key
aim of these additional searches was to ensure that any important primary
evidence that has emerged since the included reviews were conducted was
not overlooked.
Many of the included studies went beyond the usual measures of system
performance or doctors’ behaviour by focusing on “fit” of the system with other
aspects of professional and organisational life.5 We carefully considered
issues relating to the measurement of error.8 Patient safety research
initiatives can be considered in three different stages: (1) identification of risks
and hazards; (2) design, implementation, and evaluation of patient safety
practices; and (3) vigilance to ensure that a patient safety environment and
culture is maintained. Clearly, different research methods and approaches
are needed at each of the different stages of the continuum.
No single
method can be universally applied to identify risks and hazards in patient
safety. Rather, multiple approaches using combinations of these methods
increased identification of risks and hazards in terms of potential injury or
harm to patients.9
2.2 Planning phase
In terms of developing an overall search strategy, we considered several
options, such as employing individual strategies for each of the different
56
eHealth areas of interest (discussed below). However, instead we decided to
develop a single all-encompassing search strategy in order to improve overall
efficiency by avoiding overlapping multiple retrieval of publications in these
closely related areas.
We began by building on the findings of our eHealth scoping report to the
NHS Service Delivery & Organisation10 and our existing multi-disciplinary
programme of work on the impact of eHealth on the quality and safety of
health services.
Our search strategy for bibliographic databases was
developed in accordance with the recommendations for search strategies in
the Cochrane Handbook for Systematic Reviews of Interventions11 and
encompassed broad, comprehensive Medical Subject Headings (MeSH) and
free text terms in order to achieve high sensitivity in our search for possibly
eligible studies.
Initially, we also planned to employ computerised knowledge mapping
methods and analysis of non-standard data sources using free text terms.
However, due to the volume of literature from established databases, this step
was, in agreement with our External Steering Group, not performed due to
project time constraints (the Delphi method for identifying research priorities
was also omitted for the same reason).
2.2.1 Rationale for an overview of reviews
The success of the evidence-based medicine movement has resulted in a
proliferation of systematic reviews of the literature, this in turn necessitating
the need for overviews of systematic reviews to allow policy-makers, clinicians
and increasingly patients to obtain an accurate overview of the literature in a
broad field of medicine.
These overviews, also sometimes known as
“umbrella” reviews seek to collate and synthesise evidence from multiple
systematic reviews on the effectiveness of different approaches to addressing
a specific or set of related problems or, in contrast, the effectiveness of
related interventions for a range of conditions, into one accessible document.
57
Assessing and comparing the effectiveness of different approaches to
dealing with a particular problem is clearly of considerable potential
importance to healthcare professionals and patients.
Of note is that the
same or similar interventions may sometimes be used to treat different
conditions and whilst an overview of this evidence is likely to be of only
limited interest to clinicians or consumers, such an overview can, however,
be of considerable interest to policy-makers, basic scientists and technical
developers as well as academics working in related areas.
Such overviews of the high quality evidence are now routinely being
undertaken by a range of commercial (e.g. Clinical Evidence), professional
(e.g. the British Thoracic Society and Scottish Intercollegiate Guideline
Network Asthma Guideline), and regulatory bodies (e.g. the National Institute
for Health and Clinical Excellence (NICE).12 We sought to draw upon and
build on the approaches that have been developed by such organisations.
2.2.2 Detailing the scope and foci of the review and mapping the available
evidence
Getting the question(s) “right” is critical to the success of any systematic
review process.13 In the present case, whilst the review question was clearly
formulated by the commissioner of our work, this was very broad.
The
mapping exercise described below was employed to guide and refine the
focus of the review and assess the volume of potentially relevant literature.
In the context of reviews of the effectiveness of interventions, there is general
agreement that a well-formulated question involves the following key
components: (1) participants who are the focus of the interventions (in our
case both clinicians as users of eHealth applications and patients as those for
whom these applications are being used by clinicians); (2) the interventions;
and (3) the outcomes. As our review was also concerned with the factors that
influence the design, implementation and use of technological interventions,
our question included components related to (4) the context in which the
intervention was developed and implemented (see Chapter 1).
58
Shortly after the commissioning of this report, a systematic review was
published investigating the impact of IT on the quality, efficiency and costs of
healthcare.
This was commissioned by the US Agency for Healthcare
Research and Quality (AHRQ) and undertaken by Southern California
Evidence-based Practice Center, Santa Monica, CA and the RAND
Corporation.13 Their searches initially covered the time period (January 1995
to January 2004) and involved systematically searching MEDLINE, the
Cochrane Central Register of Controlled Trials, the Cochrane Database of
Abstracts of Reviews of Effects, and the Periodical Abstracts Database.
These searches yielded a total of 257 eligible studies, which were then
synthesised qualitatively (see Box 2.1 for key findings). A recent update of this
review has confirmed the overall impressions, with little discernible evidence
of progress.14
Box 2.1: Key findings from the Agency for Healthcare Research and Quality’s
report
•
IT has been shown to improve quality by increasing adherence to guidelines,
enhancing disease surveillance, and decreasing medication errors.
•
Much of the evidence on quality improvement relates to primary and secondary
preventive care.
•
The major demonstrated efficiency benefit has been decreased utilisation of
healthcare.
•
Evidence of the impact of these technologies on the professional time is
inconclusive.
•
There are only limited data on the cost-effectiveness of eHealth interventions.
•
Most of the high quality evidence on eHealth applications comes from four
benchmark research institutions.
•
Little evidence is available on the effectiveness of multi-functional commercially
developed systems.
•
There is scant evidence available on interoperability and consumer health
informatics.
•
A major shortcoming of the evidence is its limited generalisability.
Adapted from reference 13, permission requested from the Annals of Internal Medicine
59
This AHRQ review necessitated a need to reconsider the scope of our
proposed review as there was little point in simply duplicating this work. This
issue was raised with our External Steering Group (comprising of
representatives of the funders and independent expert advisors) and based
on the advice received, we refocused our efforts to expand and update this
review, but also to frame the findings within a relevant theoretical framework
to aid interpretation and assessment of the generalisability of findings. As the
AHRQ report did not cover the dimension of patient safety or implementation
issues, we retained our focus on these issues.
The field of eHealth spans a vast and complex range of applications,
technologies and issues for implementation. An important component of our
formative work has been the development and refinement of conceptual
frameworks, as an aid to managing and meaningfully interpreting the relevant
literature. A key objective of this exercise has been to triangulate the health
informatics topic areas with the particular needs and concerns of the
Department of Health’s Informatics Directorate/NHS Connecting for Health
(NHS CFH).
As noted in the commissioning brief, patient safety
considerations are often overlooked in the development and implementation
of IT in healthcare.
This aspect of our work built on the eHealth model
developed by our team (Pagliari C, unpublished work) and involved identifying
and synthesising existing frameworks for understanding dimensions of quality
and patient safety taxonomies, including that developed by the World Health
Organization (WHO),15 to develop a framework to understand the interplay
between eHealth applications of particular relevance to the DH/NHS CFH,
quality and patient safety. This represented a helpful conceptual framework
within which to organise thinking when considering development, deployment
and commissioning of IT in healthcare applications.
Based on this initial
mapping of these fields, we were able to refine the scope of the review. For
more details of the considerations that have informed this planning phase,
please see Chapters 3 and 4.
Whilst mapping the fields of eHealth, quality and safety, it became clear that
these areas are still undergoing significant conceptual and technical
60
development. For instance, there are major problems with the indexing of this
literature in the major biomedical databases. Identifying the relevant literature
therefore necessitated a considerable amount of developmental work to
identify relevant MeSH and free text terms. This paid dividends though, in that
it enabled us to identify a considerable body of relevant literature that was
overlooked by the AHRQ review.
Cognisant of the key areas of interest to the DH/NHS CFH’s National
Programme for Information Technology (NPfIT), we focused in the initial
phase of work on issues that are of particular interest to the Programme.
These related to the storing, managing and sharing of data, tools to support
professional decision-making and delivering care from a distance, and the
cross-cutting issue of how to develop and implement technological
interventions so as to maximise the chances that they will be successfully
used.
Our interest was not only on what works, but also on why things work or
conversely, do not work.16
We thus also explored the attributes of
effectiveness of interventions and critical features for success of an
intervention. In addition, we sought to understand how this varies according
to context.
2.3 Search methods for identification of evidence
The search strategy and subsequent decisions about inclusion and exclusion
of studies needs to be tailored according to the review questions, so it was
important that the questions were clearly articulated.
Despite having well
defined areas (see Section 2.4.3), searching the literature proved highly
complex and laborious. This resulted from the inherent immaturity of these
disciplines relative to others, the resultant flux that these disciplines are
currently in and the poorly developed indexing of core concepts and
technologies.
Inevitably, however, the comprehensiveness, sensitivity and
specificity of any search strategy for complex overviews of this kind have to
be influenced by pragmatic considerations such as the time and resources
available for the project.
61
While overviews of reviews use different methods from systematic reviews in
the sense that they summarise existing reviews rather than find and collate
original studies, our search strategy was developed in the same fashion as a
search strategy for a systematic review. Our focus was on identifying SRs,
whether exclusively of RCTs or also including studies employing other
designs.
2.3.1 Time period
Although computer applications in medicine have existed since the 1960s the
eHealth field, as we know it today, has emerged relatively recently and has
been characterised by considerable change over recent years.
For this
reason, we confined our searches initially to the period 1997–2007, and then
extended this to 2010. It was considered likely that information on major
historical considerations and seminal studies conducted before this time
period would have been identified by SRs appearing in this timeframe,
thereby allowing a deeper awareness of the evolution of the field to emerge.
Also of relevance was that a large amount of information from 1993 to the
start of the time period of interest would have been captured in our earlier
report on eHealth.10
2.3.2 Sensitivity and specificity
There is a tension between sensitivity and specificity in any search strategy
and especially in one that addresses a broad question. A sensitive search
strategy identifies all relevant information, whereas one that is specific
ensures that only articles of relevance are captured. Sensitivity and specificity
are influenced by a number of factors, these include: the time period covered;
the search terms used; and the combination in which these are used. There
is thus a trade-off between conducting an exhaustive search (with the
additional resources that this requires) and undertaking a search that may
miss some studies, but is unlikely to alter the overall strength of the evidence
or findings.12 Given concern that previous reviews may have missed large
amounts of relevant literature, we decided, as agreed with our External
Steering Group, to develop a highly sensitive search strategy because, even if
62
it was not possible to review all the relevant material identified in the course of
this project, the identification of this literature would nonetheless represent an
important resource to the wider academic community.
Our experiences and those of others2 reveal that quality improvement
research, particularly if related to enhancing safety, is poorly indexed within
bibliographic databases; as a result, broad search strategies using free text
and allied MeSH headings needed to be used. We built on taxonomies we
developed for our recent reviews, these including MeSH headings such as:
Equipment Safety; Information Systems (all sets, i.e. Hospital, Clinical
Laboratory, Clinical Pharmacy, Operating Room, Ambulatory Care); Medical
Informatics; Public Health Informatics; Clinical Informatics; Decision Support;
and Decision Aids.
Communication,
Telehealthcare.
Other search terms included, Electronic Clinical
Medical
Computing,
Medical
Records
Systems
and
These were then appropriately combined with terms that
focused searches on Quality of Healthcare, Safety and Errors.
We started our search using MEDLINE, which is considered the best starting
point to “get into” the literature.11 An initial list of MeSH terms was compiled by
our multi-disciplinary team relevant to the eHealth, quality and patient safety
fields. These terms were extracted from:
•
personal databases of relevant published literature already
available from our teams previous work
•
the AHRQ’s search strategy13 14
•
an in-depth manual MeSH tree exploration by two researchers (TB
& AB with input from JC, CP and the rest of the team); possible
relevant terms were explored with regard to their potential
relevance by reading the scope note which describes the MeSH
term
•
the keyword or index terms listings in our Reference Manager 11
(RefMan) database of SRs in the fields of eHealth, quality, patient
safety, and organisational and implementation issues.
63
The individual MeSH terms were independently scored by two reviewers
according to the contribution these made to the “netting” of the references
related to the key objectives of our project, this contribution being assessed
using a scoring algorithm (Table 2.1). Consensus was achieved by involving
a third researcher and the resultant file of relevant MeSH terms was then
circulated for comments and discussion to all other members of the team.
Table 2.1: Scoring algorithm for subjective exercise of scoring eHealth Medical
Subject Heading terms possibly relevant for our project
Score
Scope note
1
Stand alone MeSH term (not necessary to combine it with other terms) of
high priority for the project
2
MeSH term requires combining with other MeSH or free text terms to result
in high priority for the project
3
Combining MeSH term with other MeSH or free text terms still results in
Low priority for the project
4
MeSH term falls outside of project scope or remit (to exclude)
Whilst searching for relevant literature using MeSH terms we realised that
most of the eHealth field is also poorly indexed in bibliographic databases.
We therefore considered it important to also devise a comprehensive list of
free text terms so as to maximise the sensitivity of our search.
In order to create a broad, comprehensive list of free text terms to identify as
much as possible of the relevant literature, two researchers extracted
potential free text terms from:
•
our existing database of systematic reviews
•
online glossaries
•
webpages relating to patient safety and or eHealth (e.g. NHS CFH,
National Patient Safety Agency, AHRQ and the Virginia National
Centre for Patient Safety).
64
In addition, members of the team added terms which they considered to be
important based on their own previous research experiences. Finally, free
text terms were circulated for comments and discussion.
2.3.3 Databases searched
We searched the following eight electronic databases:
•
MEDLINE (from 1 January 1997 to August 2007)
•
EMBASE (from 1 January 1997 to August 2007)
•
The Cochrane Database of Systematic Reviews (from 1 January
1997 to August 2007)
•
Database of Abstracts of Reviews of Effects (from 1 January 1997
to August 2007)
•
The Cochrane Central Register of Controlled Trials (from 1 January
1997 to August 2007)
•
The Cochrane Methodology Register (from 1 January 1997 to to
August 2007)
•
Health Technology Assessment Database (from 1 January 1997 to
August 2007)
•
NHS Economic Evaluation (from 1 January 1997 to August 2007).
These searches were then extended to cover the period until 31 October
2010.
We also searched databases of research in progress or unpublished work:
•
The National Research Register (http://www.nihr.ac.uk/databases)
•
ClinicalTrials.gov (http://clinicaltrials.gov)
•
Current Controlled Trials (http://www.controlled-trials.com).
In order to make our search strategy sufficiently sensitive to avoid missing any
high quality evidence we used the Boolean operator ‘OR’ between all relevant
MeSH and free text terms relating to eHealth. We also did this with relevant
quality, safety, organisational and implementation terms. In the final stage,
we combined those two searches with Boolean operator ‘AND’ (IT search
65
strategy ‘AND’ patient safety and quality, organisational and implementation
issues search strategy). Although this approach inevitably yielded a large
number of articles that ultimately proved irrelevant, this formative phase
allowed us to develop a highly sensitive search strategy that helped to
overcome a key limitation that systematic reviewers have hitherto faced, i.e.
suboptimal indexing of papers on eHealth. Appendix 1 details our final search
strategy.
2.4 Criteria for considering reviews
2.4.1 Types of studies
Systematic reviews, health technology assessments and RCTs were the main
study designs of interest. We did, however, also draw on studies using other
study designs such as descriptive or qualitative studies to understand broader
contextual considerations, particularly in relation to assessing the acceptability
of interventions.
2.4.2 Types of participants
In keeping with the commissioning brief, we were interested in reviews
focused on reports of studies evaluating the impact of IT used by any type of
a healthcare professional (e.g. doctor and nurse) or allied professional (e.g.
receptionists and administrators).
Reviews that reported studies focusing
exclusively on the use of IT by patients and or their carers were not eligible for
inclusion.
2.4.3 Types of interventions
To be eligible for inclusion, studies needed to address eHealth applications.
We therefore included studies assessing the following interventions:
•
use of computers in information exchange
•
electronic health records
•
computer-based history taking systems
•
electronic booking systems and electronic referral systems
•
computerised decision support systems
•
artificial intelligence in healthcare (relevant to computerised decision
support systems)
66
•
computerised reminders in clinical practice
•
computer-aided detection or diagnosis in medical imaging
•
decision support for ePrescribing and other orders
•
telehealthcare
•
human and organisational factors related to computing in healthcare.
In addition, studies needed to address at least one of the following: impact on
safety; quality; or organisational, implementation or adoption considerations.
Consequently, although studies aiming to describe the trends in the literature
informed our work they were excluded as they did not assess impact on the
parameters of interest to the report.
Although within the broad field of eHealth, we excluded reports that were
outside the focus of the commissioning brief, these included studies of:
•
computer-assisted and any other type of IT enhanced surgery
•
computer-assisted therapy predominantly directed at or used by
patients
independently
or
under
supervision
of
a
healthcare
professional (e.g. smoking cessation, cognitive behaviour therapy17)
•
telemedicine (i.e. inter-professional communication)18
•
eLearning (the use of electronic technology and media to deliver,
support and enhance learning and teaching)19
•
consumer health informatics (patient-oriented eHealth)
•
information literacy20-22
•
point-of-care testing without CDSS or with CDSS directed at patients.23
•
public health surveillance systems24
•
references not having an a priori IT focus or objective, but that reported
on IT post retrieval of references.25 26
We subsequently found that many studies employed systematic review
methodology to simply describe the past and or current state of the literature
for medical or health informatics in general,27 or in particular geographical
setting28 or describe eHealth applications for a particular clinical domain
without assessing patient, practitioner or organisational impact.29 30 Similarly,
67
we found a number of studies using systematic review methodology to assess
study quality,31 32 such as use and validity of outcome measurement,33 validity
of economic analysis34 35 or appropriate use of statistical analysis.36 We found
that systematic reviews had also been conducted to provide a taxonomic
description of a particular eHealth application37-39 such as who is using the
application40 or how language regarding a field is being used.41 Some
systematic reviews did not provide sufficient information on included studies
to determine relevance.42
43
Finally, some highly relevant (non-systematic)
literature reviews had to be excluded from data abstraction and critical
appraisal due to one or more inclusion criteria not being met.44-54
2.4.4 Types of outcomes
The focus of this review were studies that were concerned with assessing the
impact of eHealth applications on a wide range of quality of care and or
patient safety indicators, as well as factors related to implementation and
adoption of eHealth applications.
2.5 Methods of the review
Two reviewers independently assessed the potential relevance of all titles and
abstracts identified from the electronic searches. Full text copies of all articles
judged to be potentially relevant from the titles and abstracts were retrieved.
Two reviewers then independently assessed these retrieved articles for
inclusion. Consensus on the final list of included studies was reached, with
any disagreements about particular reviews being resolved by discussion.
Reference lists of included reviews were manually searched.
2.6 Criteria for considering articles for this review
In the process of search strategy development we considered a number of
different methodological filters exploring the types of studies these yielded
and the volume of this literature, e.g. the Cochrane Effective Practice and
Organisation of Care Group filter. In the end, we decided that filters produced
by SIGN55 gave the best balance between sensitivity and specificity. Using
these filters, we also uncovered a number of studies using other designs and
these were categorised as described above. Although we did not apply any
68
language restrictions to electronic searches, we did not undertake critical
appraisal or data extraction on studies not published in English. Similarly
critical appraisal and data extraction were not performed on HTAs.
2.6.1 Selection criteria for systematic reviews
In order to select relevant SRs we applied methodological filters for SRs
devised by SIGN.55 This approach has been validated and is used in the
development of a number of evidence-based guidelines (see Appendix 1 for
methodology filters).
In order to be considered a SR, the publication needed to:
•
explicitly state that the publication is a SR in the title or text; or
•
search the literature in a systematic way in order to answer one or
more clear questions or describe a given topic; and
•
apply inclusion and or exclusion criteria or perform quality assessment
2.7 Quality assessment
We reviewed a range of quality assessment tools, but found that these were
all less than ideal for our purposes, the specific issues of relevance being the
narrow range of methodological approaches being considered and the failure
to pay sufficient attention to contextual considerations. We therefore adapted
available instruments56 to produce an instrument that was suitable for
assessment of SRs of eHealth (see Appendix 2) and this was then used to
assess the methodological quality of all included SRs.
The quality of all
eligible studies was assessed by two independent reviewers.
2.8 Data extraction
We developed a customised data extraction form to meet our specific needs
(see Appendix 2). Where appropriate, quantitative and qualitative data were
extracted from the references identified, however, in view of the breadth of
literature to be summarised and the limited timescale of the project, this took
the form of descriptions of the results of secondary research, e.g. SRs of the
effectiveness or cost-effectiveness of CDSSs, rather than attempts to
combine the results of individual studies for statistical analysis.
69
2.9 Synthesis and interpretation of identified quality and safety,
introduction, implementation and sustainability issues and features of
eHealth systems into conceptual frameworks
Multiple factors may combine to influence the success of eHealth applications
and for this reason such interventions can be difficult to evaluate.57 While
compiling available evidence, we aimed to explore the frontiers of knowledge
on quality and safety of eHealth applications by evaluating functional
capabilities and characteristics of eHealth solutions according to evidence
from studies.
Where possible though, we also used information from
computer simulation58 and conceptual models59 that assessed the effects that
could
be
expected
from
each capability
in
the
proposed
clinical
environment.60-62 This additional consideration of conceptual models was
important as we know that reporting of errors in healthcare is poor63 and many
may conceivably be envisaged to happen even if not reported. Therefore,
throughout the process of collection of evidence we aimed to identify (and
build new where non-existent) conceptual frameworks for evaluating eHealth
solutions based on their functional capabilities, functional characteristics and
simulations.64-66 We commented on the relevance of the evidence-base within
the specific context of the NHS67 aiming to incorporate findings, wherever
possible, into an appreciation of the skills, knowledge, experience, attitudes
and values of people (clinicians, healthcare managers, leaders, and patients);
as well as the characteristics of tools (such as adaptiveness), environmental
factors, tasks, goals and their inter-relationships.68
Importantly, we reported the evidence that is mounting about features that are
independent predictors of improved clinical practice69
70
and predictors of
successful implementation.70-75 For example, flexibility in incorporating
information from diverse sources and adaptability to varied practice settings
are likely to be the quality criteria by which CDSSs are judged in the future.76
We also aimed to consider broader factors, for example, whether external
incentives are being used sufficiently in promotion of eHealth innovations.77
70
Our multi-disciplinary group, representing the disciplines of clinical care,
informatics, patient safety, epidemiology, social science, psychology, medical
education and health policy, sought, in regular Project Steering Group
meetings, to engage with the data being generated and to contextualise the
findings within broader considerations relating to healthcare systems’ reforms,
developing technologies and policy deliberations on ways of encouraging
professionals to embrace innovations to improve the quality and safety of
primary and secondary medical care.
We followed three main steps in conducting a narrative synthesis of these
data: (1) developing a preliminary synthesis of the findings of included
studies; (2) exploring relationships in the findings; and (3) assessing the
robustness of the synthesis produced.7 Depending on the findings from this
last step, we would, if necessary, undertake further searches for evidence as
described above. We aimed as far as possible for transparency78 in this
essentially qualitative approach to seeking “saturation” in finding new
evidence or new dimensions to evidence.
In evaluating the overall strength of evidence, we considered the World Health
Organization’s Health Evidence Network (WHO HEN) system for public health
evidence79 (adapted from Greenhalgh et al.).80 This classifies evidence as:
•
Strong: consistent, good quality, plentiful or generalisable
•
Moderate: consistent and good quality
•
Limited to none: inconsistent or of poor quality.
Further, we judged key elements defined by the AHRQ in rating the strength
of findings on the basis of considerations relating to:
•
Quality: the aggregate of quality ratings for individual studies,
predicated on the extent to which bias was minimised
•
Quantity: number of studies, sample size or power, and magnitude
of effect
•
Consistency: for any given topic, the extent to which similar findings
are reported using similar and different study designs.
71
This classifying of the strength of evidence is presented in the Executive
Summary.
While covering a vast territory of eHealth, we selected and developed
exemplary case studies to develop a far richer, more context specific account
of evidence in relation to key technologies and issues that have the potential
to be of particular interest to the DH/NHS CFH (Chapters 9, 11, 12 and 18),
which involved engaging with the primary evidence in considerably more
detail. We have furthermore also conducted Cochrane systematic reviews
(see for example, Chapters 14 and 15) and are actively engaged in a number
of other related systematic reviews (see Appendix 7), the findings from which
have contributed to this overview of the literature.
References
1. Sheldon TA. Making evidence synthesis more useful for management and
policy-making. J Health Serv Res Policy 2005; 10 Suppl1: 1-5.
2. Grimshaw J, McAuley LM, Bero LA, Grilli R, Oxman AD, Ramsay C et al.
Systematic reviews of the effectiveness of quality improvement strategies and
programmes. Qual Saf Health Care 2003; 12: 298-303.
3. McAuley L, Pham B, Tugwell P, Moher D. Does the inclusion of grey literature
influence estimates of intervention effectiveness reported in meta-analyses?
Lancet 2000; 356: 1228-31.
4. Pham B, Platt R, McAuley L, Klassen TP, Moher D. Is there a "best" way to
detect and minimize publication bias? an empirical evaluation. Eval Health Prof
2001; 24: 109-25.
5. Kaplan B. Evaluating informatics applications--clinical decision support systems
literature review. Int J Med Inform 2001; 64:15-37.
6. The Cochrane Collaboration. Cochrane Effective Practice and Organisation of
Care Group, 2010.
7. Shepperd S, Lewin S, Straus S, Clarke M, Eccles MP, Fitzpatrick R, et al. Can
we systematically review studies that evaluate complex interventions. PLoS
Medicine 2009; 6: e1000086.
8. Lilford RJ, Mohammed MA, Braunholtz D, Hofer TP. The measurement of
active errors: methodological issues. Qual Saf Health Care 2003; 12: 8ii-12.
9. Battles JB, Lilford RJ. Organizing patient safety research to identify risks and
hazards. Qual Saf Health Care 2003; 12: 2ii-7.
10. Pagliari C, Sloan D, Gregor P, Sullivan F, Kahan JP, Detmer D et al. Literature
review and conceptual map of the field of eHealth. Final report to the SDO
programme. 2004.
11. Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of
Interventions Version 5.0.2 [updated September 2009]. The Cochrane
Collaboration, 2009. Available from: www.cochrane-handbook.org (last
accessed 4/1/11).
72
12. National Institute for Clinical Excellence. Guide to the Technology Appraisal
Process. London: National Institute for Clinical Excellence, 2004.
13. Chaudhry B, Wang J, Wu S, Maglione M, Mojica W, Roth E et al. Systematic
review: impact of health information technology on quality, efficiency, and costs
of medical care. Ann Intern Med 2006; 144: 742-52.
14. Goldzweig CL, Towfigh A, Maglione M, Shekelle PG. Costs and benefits of
health information technology: new trends from the literature. Health Aff
(Millwood) 2009; 28: w282-93.
15. World Health Organization. More than words - International Classification for
Patient Safety. Geneva: WHO, 2007.
16. Pawson R, Tilley N. Realistic Evaluation. London: Sage, 1997.
17. Kaltenthaler E, Brazier J, De NE, Tumur I, Ferriter M, Beverley C et al.
Computerised cognitive behaviour therapy for depression and anxiety update: a
systematic review and economic evaluation. Health Technol Assess 2006; 10:
iii,xi-iii,168.
18. Pagliari C, et al. EH1 E-Health Scoping Exercise. Review of the Traditional
Research Literature.
19. University of Bristol. eLearning: Using technology to support learning.
http://www.bris.ac.uk/medical-school/staff/elearning/ (last accessed 4/1/11).
20. Hersh WR, Hickam DH. How well do physicians use electronic information
retrieval systems? A framework for investigation and systematic review. JAMA
1998; 280: 1347-52.
21. Pluye P, Grad RM, Dunikowski LG, Stephenson R. Impact of clinical
information-retrieval technology on physicians: a literature review of
quantitative, qualitative and mixed methods studies. Int J Med Inform 2005; 74:
745-68.
22. Saranto K, Hovenga EJ. Information literacy-what it is about? Literature review
of the concept and the context. Int J Med Inform 2004; 73: 503-13.
23. Fermann GJ, Suyama J. Point of care testing in the emergency department. J
Emerg Med 2002; 22: 393-04.
24. German RR. Sensitivity and predictive value positive measurements for public
health surveillance systems. Epidemiol 2000; 11: 720-27.
25. Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW.
Deficits in communication and information transfer between hospital-based and
primary care physicians: implications for patient safety and continuity of care.
JAMA 2007; 297: 831-41.
26. Shojania KG, Duncan BW, McDonald KM, Wachter RM, Markowitz AJ. Making
health care safer: a critical analysis of patient safety practices. Evid Rep
Technol Assess (Summ) 2001; i-668.
27. Ammenwerth E, de KN. An inventory of evaluation studies of information
technology in health care trends in evaluation research 1982-2002. Methods Inf
Med 2005; 44: 44-56.
28. Mendis K. Health informatics research in Australia: retrospective analysis using
PubMed. Inform Prim Care 2007; 15: 17-23.
29. Dinakarpandian D, Lee Y, Dinakar C. Applications of medical informatics in
allergy/immunology. Ann Allergy Asthma Immunol 2007; 99: 2-9.
30. Duvvuri VR, Jianhong W. Information and communication technology
developments in asthma management: a systematic review. Ind J Med Sci
2007; 61: 221-41.
31. de Keizer N, Ammenwerth E. The effects and quality of IT evaluation studies:
Trends in 1982 - 2002. AMIA Annu Symp Proc 2005; 186-90.
32. de Keizer NF, Ammenwerth E. The quality of evidence in health informatics:
How did the quality of healthcare IT evaluation publications develop from 1982
to 2005? Int J Med Inform 2008; 77: 41-9.
73
33. Friedman CP, Abbas UL. Is medical informatics a mature science? A review of
measurement practice in outcome studies of clinical systems. Int J Med Inform
2003; 69: 261-72.
34. Eisenstein EL, Ortiz M, Anstrom KJ, Crosslin DR, Lobach DF. Economic
evaluation in medical information technology: why the numbers don't add up.
AMIA Annu Symp Proc 2006; 914.
35. Eisenstein EL, Ortiz M, Anstrom KJ, Crosslin DR, Lobach DF. Assessing the
quality of medical information technology economic evaluations: room for
improvement. AMIA Annu Symp Proc 2006; 234-38.
36. Chuang JH, Hripcsak G, Jenders RA. Considering clustering: a methodological
review of clinical decision support system studies. Proc AMIA Symp 2000; 146150.
37. Berlin A, Sorani M, Sim I. A taxonomic description of computer-based clinical
decision support systems. J Biomed Inform 2006; 39: 656-67.
38. Sim I, Berlin A. A framework for classifying decision support systems. AMIA
Annu Symp Proc 2003; 599-603.
39. Cruz-Correia RJ, Vieira-Marques PM, Ferreira AM, Almeida FC, Wyatt JC,
Costa-Pereira AM. Reviewing the integration of patient data: how systems are
evolving in practice to meet patient needs. BMC Med Inform Decis Mak 2007;
7: 14.
40. Garritty C, El EK. Who's using PDAs? Estimates of PDA use by health care
providers: a systematic review of surveys. J Med Internet Res 2006; 8: e7.
41. Oh H, Rizo C, Enkin M, Jadad A. What is eHealth (3): a systematic review of
published definitions. J Med Internet Res 2005; 7: e1.
42. Gurses AP, Xiao Y. A systematic review of the literature on multidisciplinary
rounds to design information technology. J Am Med Inform Assoc 2006; 13:
267-76.
43. Sargent DJ. Comparison of artificial neural networks with other statistical
approaches: results from medical data sets. Cancer 2001;91:1636-1642.
44. Coiera E, Westbrook J, Wyatt J. The safety and quality of decision support
systems. Methods Inf Med 2006; 45(Suppl 1): 20-25.
45. Corley ST. Electronic prescribing: a review of costs and benefits. Top Health Inf
Manage 2003; 24: 29-38.
46. de Lusignan S. The barriers to clinical coding in general practice: a literature
review. Med Inform Internet Med 2005; 30: 89-97.
47. Hanson CW, III, Marshall BE. Artificial intelligence applications in the intensive
care unit. Crit Care Med 2001; 29: 427-35.
48. Kaushal R, Barker KN, Bates DW. How can information technology improve
patient safety and reduce medication errors in children's health care? Arch
Pediatr Adolesc Med 2001; 155: 1002-07.
49. Kuperman GJ, Gibson RF. Computer physician order entry: benefits, costs, and
issues. Ann Intern Med 2003; 139: 31-39.
50. Kuperman GJ, Bobb AM, Payne T, Avery AJ, Gandhi TK, Burns G et al.
Medication-related clinical decision support in computerized provider order
entry systems: a review. J Am Med Inform Assoc 2007; 14: 29-40
51. Papshev D, Peterson AM. Electronic prescribing in ambulatory practice:
promises, pitfalls, and potential solutions. Am J Manag Care 2001; 7: 725-36.
52. Stroetmann VN, Thierry J-P, Dobrev A. eHealth for Safety: Impact of ICT on
Patient Safety and Risk Management. 2007.
53. Thursky K. Use of computerized decision support systems to improve antibiotic
prescribing Expert Rev Anti Infect Ther 2006; 4: 491-507.
54. Zhu Y, Williams S, Zwiggelaar R. Computer technology in detection and staging
of prostate carcinoma: a review. Med Image Anal 2006; 10: 178-99.
55. Scottish Intercollegiate Guidelines Network (SIGN) methodology filters.
Available from: www.sign.ac.uk/methodology/filters.html (last accessed 4/1/11).
74
56. Oxman AD. Systematic reviews: Checklists for review articles. BMJ 1994; 309:
648-51.
57. Lorenzi NM, Riley RT. Organizational issues = change. Int J Med Inform 2003;
69:197-203.
58. Fernando B, Savelyich BSP, Avery AJ, Sheikh A, Bainbridge M, Horsfield P et
al. Prescribing safety features of general practice computer systems: evaluation
using simulated test cases. BMJ 2004; 328: 1171-72.
59. Panniers TL, Feuerbach RD, Soeken KL. Methods in informatics: using data
derived from a systematic review of health care texts to develop a concept map
for use in the neonatal intensive care setting. J Biomed Inform 2003; 36: 23239.
60. Avery AJ, Sheikh A, Hurwitz B, Smeaton L, Chen YF, Howard R et al. Safer
medicines management in primary care. Br J Gen Pract 2002; 52(Suppl): S17S22.
61. Sim I, Gorman P, Greenes RA, Haynes RB, Kaplan B, Lehmann H et al.
Clinical decision support systems for the practice of evidence-based medicine.
J Am Med Inform Assoc 2001; 8: 527-34.
62. Wyatt J, Spiegelhalter D. Evaluating medical expert systems: what to test and
how? Med Inform (Lond) 1990; 15: 205-17.
63. Sheikh A, Hurwitz B. Reducing error, improving safety. Log of errors is needed.
BMJ 2000; 321: 505.
64. Anderson JG, Jay SJ, Anderson M, Hunt TJ. Evaluating the capability of
information technology to prevent adverse drug events: a computer simulation
approach. J Am Med Inform Assoc 2002; 9: 479-90.
65. Bell DS, Cretin S, Marken RS, Landman AB. A conceptual framework for
evaluating outpatient electronic prescribing systems based on their functional
capabilities. J Am Med Inform Assoc 2004; 11: 60-70.
66. Martin P, Haefeli WE, Martin-Facklam M. A drug database model as a central
element for computer-supported dose adjustment within a CPOE system. J Am
67. Arai L, Roen K, Roberts H, Popay J. It might work in Oklahoma but will it work
in Oakhampton? Context and implementation in the effectiveness literature on
domestic smoke detectors. Inj Prev 2005; 11: 148-51.
68. Small SD. Medical device-associated safety and risk: surveillance and
stratagems. JAMA 2004; 291: 367-70.
69. Grol R, Grimshaw J. From best evidence to best practice: effective
implementation of change in patients' care. Lancet 2003; 362: 1225-30.
70. Kawamoto K, Houlihan CA, Balas EA, Lobach DF. Improving clinical practice
using clinical decision support systems: a systematic review of trials to identify
features critical to success. BMJ 2005; 330: 765-68.
71. Butler KA, Esposito C, Hebron R. Connecting the design of software to the
design of work. Commun ACM 1999; 42: 38-46.
72. Durieux P. Electronic medical alerts -- so simple, so complex. N Engl J Med
2005; 352: 1034-36.
73. Handler JA, Feied CF, Coonan K, Vozenilek J, Gillam M, Peacock PR, Jr. et al.
Computerized physician order entry and online decision support. Acad Emerg
Med 2004; 11: 1135-41.
74. Holbrook A, Xu S, Banting J. What factors determine the success of clinical
decision support systems? AMIA Annu Symp Proc 2003; 862.
75. Zheng K, Padman R, Johnson MP, Diamond HS. Understanding technology
adoption in clinical care: Clinician adoption behavior of a point-of-care reminder
system. Int J Med Inform 2005; 74: 535-43.
76. Purcell GP. What makes a good clinical decision support system. BMJ 2005;
330: 740-41.
75
77. Casalino L, Gillies RR, Shortell SM, Schmittdiel JA, Bodenheimer T, Robinson
JC et al. External incentives, information technology, and organized processes
to improve health care quality for patients with chronic diseases. JAMA 2003;
289: 434-441.
78. PLoS Medicine Editors. Many reviews are systematic but some are more
transparent and completely reported than others. PLoS Med 2007; 4: e147.
79. World Health Organization Regional Office for Europe. Health Evidence
Network (HEN). 2007.
80. Greenhalgh T, Robert G, Macfarlane F, Bate P, Kyriakidou O. Diffusion of
innovations in service organizations: systematic review and recommendations.
Milbank Q 2004; 82: 581-629.
76
Chapter 3
The National Programme for Information Technology,
NHS Connecting for Health and the Information
Revolution
Summary
•
The National Programme for Information Technology was one of the most
comprehensive, ambitious and expensive eHealth-based overhaul of
healthcare delivery ever undertaken.
•
As a transformation programme for the NHS that underpinned system
reform, it was conceived as being far more than just the implementation of
information technology.
•
This Programme had its origins in the 1998 Department of Health strategy
Information for Health, which committed the NHS to lifelong electronic
health records for everyone with round-the-clock, on-line access to patient
records and information about best clinical practice for all NHS clinicians.
•
Officially launched in 2002, the Programme aimed over a 10-year period
to create the infrastructure, tools and environment through which it is
possible to deliver:
o a longitudinal electronic patient record (from “cradle to grave”)
accessible to multiple users throughout the NHS; this (i.e. NHS Care
Records Service) together with the dedicated National Network for the
NHS (N3) and the national database on which these records will be
held (the Spine); represents the backbone to the Programme
o a service through which prescriptions can be transferred electronically
from general practitioners and other prescribers to pharmacists
(Electronic Prescriptions Service) and which will in time integrate with
the
NHS
Care
Records
Service
(Electronic
Transmission
of
Prescriptions)
o An electronic appointment booking service enabling patients to
electronically book hospital appointments (Choose & Book).
77
• The Programme was subsequently expanded to include, amongst other
things:
o Picture Archiving and Communication System
o GP2GP, which is a system that enables electronic transfer of patient
records between GP practices when patients change their practice
o Quality Management Analysis System, which automates assessment
of GP practice performance against criteria in the new GP contract.
•
Whilst these were the headline deliverables of the Programme, our
scoping of the field identified other related eHealth projects or
applications, which, although officially falling within the remit of the now
defunct
National
Knowledge
Service
(such
as
ePrescribing
and
computerised decision support tools), were also within the broad remit of
NHS Connecting for Health and were therefore also closely interconnected with the delivery of the Programme. More recently, the
Programme expanded its remit yet further to include telehealthcare
initiatives within its fold.
•
Originally managed directly by the Department of Health, oversight of the
Programme transferred in 2005 to NHS Connecting for Health–a newly
created arm’s length body. More recently, both the Programme and NHS
Connecting for Health have been absorbed into the Department of
Health’s Informatics Directorate.
•
Foremost amongst the roles of NHS Connecting for Health is responsibility
for the national procurement of systems and services that were needed to
ensure delivery of the national Programme.
•
Given the extremely high level of public expenditure, the Programme has
attracted considerable public, professional, legal, financial, political and
international scrutiny.
•
The election of the new coalition government has in effect marked the end
of the original Programme; it’s recent strategy document, Liberating the
NHS, and White Paper, An Information Revolution, signify a clear move
78
away from a centralised “replace all” to a more locally owned and
delivered “connect all” model for health information technology.
3.1 Introduction
The National Programme for Information Technology (NPfIT) remains one of the
most ambitious and expensive IT-based overhaul of a health service ever
undertaken. In this chapter, we aim to give an overview of the Programme
being delivered by NHS Connecting for Health (NHS CFH) and the Department
of Health (DH). To this end, we describe the key applications that NHS CFH is
implementing and the structures that have been created to deliver these.
Evidence of the impact of those on quality and safety of care and how they can
be best implemented is discussed in later chapters in this report (Section 3).
We also reflect on recent changes in the health care IT strategy that have come
into force following the formation of the coalition government in May 2010.
3.2 NHS Connecting for Health
There are presently many different computer systems in the NHS in England of
variable quality and age; there is furthermore no national means to efficiently,
securely and confidentially transfer healthcare information between different
NHS locations.
In July 2004, following the Review of its Arms Length Bodies,1 the Department of
Health (DH) announced the creation of a new organisation, combining
responsibility for the delivery of the NPfIT with the management of the IT-related
functions of the NHS Information Authority (NHSIA), which had previously been
earmarked for closure. This newly created organisation, NHS CFH, thus had
within its remit the national procurement of critical IT healthcare systems and
then ensuring the maintenance, development and effective delivery of these IT
products and services. This remit continues following the recent reintegration of
the NPfIT (and NHS CFH) into the Department of Health’s Informatics
Directorate. Thus, whilst the core aim of NHS CFH is supporting the delivery of
79
NPfIT, it is also responsible for delivering a number of other programmes of
work that were previously the responsibility of the NHSIA (see Box 3.1).
Established in April 2005 as the national IT procurer for the health service, NHS
CFH was charged with providing integrated IT infrastructure and systems for the
NHS in England.2
It is, for example, responsible for electronically connecting
1.3 million NHS staff (of whom around 100,000 are doctors, 390,000 nurses
and 120,000 other healthcare professionals) and giving patients access to their
personal healthcare information. This body has thus, in a very real sense, been
laying the technical infrastructure for the NHS.
NHS CFH has also provided the policy focus for the DH on IT considerations for
the NHS.
This has included shaping the strategic infrastructure to ensure
integration where necessary and, where choices are available, setting the
standards required for local IT applications. This is made clear on NHS CFH’s
website, which states that: ‘NHS Connecting for Health operates through a
mixed economy of staff drawn from across the NHS, civil service, academia and
the private sector. It utilises this rich diversity of experience—encompassing
management, IT, clinical and medical skills—to deliver the national Programme
and thus improve services to patients.’
Box 3.1 Services and applications being delivered by NHS CFH (for more
information on each of those see NHS CFH’s website)
•
Addressing
•
Automatic identification and data capture (AIDC)
•
Blood safety tracking pilot
•
Choose and Book
•
Clinical Dashboards
•
Data Standards
•
Demographics
•
Deployment support
•
Education, Training and Development (ETD)
80
•
Electronic Prescription Service (EPS)
•
ePrescribing
•
GP Support
•
Health and Social Care Integration Programme
•
HealthSpace
•
Implementation
•
Informatics Capability Development (ICD)
•
Information Governance (IG)
•
Map of Medicine
•
N3 - The National Network
•
NHS Business Partners
•
NHS Care Records Service (NHS CRS)
•
NHS Infrastructure Maturity Model (NIMM)
•
NHS Interoperability Toolkit
•
NHSmail
•
NHS Pathways
•
NHS Strategic Tracing Service (NSTS)
•
Pathology Messaging
•
Picture Archiving and Communications System (PACS)
•
Professionalising Health Informatics (PHI)
•
Prison Health IT
•
Registration Authorities and Smartcards
•
Research Capability Programme
•
Secondary Uses Service (SUS)
•
Specialty Systems Project
•
Spine
•
Summary Care Records (SCR)
•
Systems & Service Delivery
Following the election in May 2010, the new coalition government integrated
NHS CFH into the DH’s Informatics Directorate. The publication of its NHS
strategy, Equity and Excellence: Liberating the NHS,3 and the subsequent
launch of its White Paper, Liberating the NHS: An Information Revolution,4 has
81
resulted in a shift in strategic focus for NHS CFH from a centralised to a more
locally led, pluralist, approach to implementing IT.
3.3 Background to the National Programme for Information Technology
The Programme had its origins in the 1998 DH strategy Information for Health,5
which committed the NHS to lifelong electronic health records (EHRs), ensuring
round-the-clock on-line access to patient records, and information to support
best clinical practice for all healthcare professionals working within the NHS.
Following the development and publication of the NHS Plan,6 a more detailed
supporting document, Building the Information Core: Implementing the NHS
Plan,7 published in January 2001, outlined the information and IT systems
needed to deliver the NHS Plan and support patient-centred care and services.
In March 2001, Derek Wanless, a commissioner with the Statistics Commission
was asked to examine future trends affecting the health service in the UK over
The Wanless Report,8 published in April 2002, had
the next two decades.
several key recommendations for IT in the NHS. These included: a doubling
and ring-fencing of IT spending; stringent, centrally-managed national standards
for data and IT; and better management of IT implementation in the NHS,
including a national programme. The Wanless Report coincided with the
publication of Delivering the NHS Plan,9 which developed the vision of ‘…a
service designed around the patient’, offering more choice of where and when
they accessed treatment.
In June 2002, following the Wanless Report and Delivering the NHS Plan, the
Department of Health published its new strategy for developing IT in the NHS
Delivering 21st Century IT Support for the NHS–A National Strategic
Programme.10 This strategy laid out the first steps for the Programme, including
the creation of a ministerial taskforce and the recruitment of a Director General
for the NPfIT.
control
Formally established in October 2002, it had a mandate to
specification,
procurement,
resource
management,
performance
82
management and delivery of the IT agenda centrally and implement modern,
integrated IT infrastructure and systems for all 8500 general practices and their
community services, and the approximately 300 NHS hospitals in England by
2010.
The original scope is presented in Figure 3.1.
The significantly
expanded scope is in contrast presented in Figure 3.2.
Figure 3.1: Original scope for the National Programme for Information
Technology in 2002
Source: Connecting For Health (2007) Reprinted with permission from NHS Connecting for Health. Crown
copyright
Figure 3.2: NHS Connecting for Health and National Programme for Information
Technology overview in February 2007
83
Source: NHS Connecting For Health (2007) Reprinted with permission from NHS Connecting for Health.
Crown copyright.
3.4 Programme scope, function and key activities
The NPfIT’s (and NHS CFH and the DH’s) scope continues to evolve, as does
the terminology associated with it, such that it is difficult for an outside observer
(or indeed those “inside” the organisation) to have a full picture of the
Programme’s activities. Broadly, it comprises clinical change, clinical systems,
and the underlying infrastructure for the NHS in England, which now serves
over 50 million people.
A key aim of the NPfIT, as originally conceived, was to support clinicians in
providing care by giving them better access to patient information and
supporting them in decision-making whenever and wherever this support is
needed. The Programme was seen as an essential element in delivering New
Labour’s NHS Plan6 to create the infrastructure for facilitating improved patient
care by enabling clinicians and other NHS staff to increase the efficiency,
effectiveness and safety of healthcare provision. As presented in Figure 3.1,
the core original deliverables of the Programme are:
•
An electronic NHS Care Records Service (NHS CRS), which consists of
a nationally transferable Summary Care Record (SCR) and a more local
Detailed Care Record (DCR). An essential component of NHS CRS is
the Spine—a central storage and communication service for records that
also supports other core systems and services.
•
An electronic prescription service (EPS), this being implemented through
the Electronic Transmission of Prescriptions (ETP) programme, which
allows for the digital transmission of prescriptions from GPs and other
prescribers to pharmacies. Details of dispensed prescriptions are also
sent electronically from pharmacies to the Prescription Pricing Authority,
the organisation that reimburses dispensers, i.e. community pharmacies
for the medication they have supplied to patients.
84
•
An electronic booking service, Choose and Book, which aims to make it
easier and faster to book convenient and accessible hospital
appointments for patients.
•
A National Network for the NHS (N3)—the largest virtual private network
(VPN) in Europe—which provides the IT infrastructure to meet NHS
needs with fast, reliable, high-bandwidth connectivity.
As noted above, this original list of deliverables, however, expanded to include
the following:
•
GP2GP, which is a system designed to allow direct secure transfer of
patient records between GPs when patients change practice.
•
The Quality Management and Analysis System (QMAS), which collects
data on quality of care provided by general practices, measured against
national Quality and Outcomes Framework (QOF) targets described in
the Revisions to the General Medical Services Contract.11
•
HealthSpace has been developed with a wide scope including support of
Choose and Book. This is a secure online personal health organiser,
which is accessible to anyone aged over 16 and living in England.
•
Picture Archiving and Communications Systems (PACS), which enables
images such as x-rays and scans to be stored electronically and then
viewed through any computer or screen linked to the PACS system.
•
NHSmail provides an email service for NHS staff accessible using web
browsers or commonly used email clients.
•
A linked Directory Service contains details of all NHS staff, including a
user interface to facilitate searching for people and browsing of the NHS
organisational structure. Directory and NHSmail services also have an
integrated outbound fax service and outbound short message service
(SMS).
Below we consider in more detail some of the above described core
deliverables of the Programme.
85
3.4.1 NHS Care Records Service
English GPs and primary care providers already routinely use EHRs (see
Chapter 6). Whilst useful, these records are however currently only available
from local primary care providers (i.e. the practice with which the patient is
registered). The NHS CRS in contrast aims to provide a live, interactive patient
record service, which is accessible to healthcare professionals at all times,
irrespective of wherever they work in the NHS.12 It consists of a locally held
Detailed Care Record (DCR) and a nationally transferable SCR.
The SCR was first introduced into selected ‘Early Adopter’ practices in 2007
and had been rolled-out to an estimated 400+ GP practices (out of 8,000) by
September 2010. Through the use of an electronic SCR service, clinicians can
access key details from a patient’s medical record, such as allergies, current
prescriptions, date of birth, and address. The SCR is however just the first step
toward what is seen as a service to provide safe, secure access to up-to-date
detailed healthcare records for every patient in England. The plan is to enable
clinicians to access patients’ records securely using a Smartcard, when and
where needed, via a nationally maintained information repository. Despite
widely acknowledged problems (including concerns about the opt-out consent
model, delays in implementation and limited patient interest)
with the
implementation of the SCR and HealthSpace, the importance of patients having
direct access to and the ability to contribute to their own medical records has
also been emphasised by the coalition government.13
While the online care record should eventually be—depending on individual
access rights—available in full to NHS staff who care for the patient, patients
will also for the first time have access to a summary of their healthcare records
through HealthSpace. NHS CFH developed an extensive public awareness
campaign to introduce this radical new way of sharing and storing patient
information to the public, this involves, amongst other things, information
leaflets,14
15
and a patient video.16 In an attempt to allay concerns regarding
86
possible breaches of patient confidentiality, the NHS published its commitments
to maintain security of data in its NHS Care Records Guarantee (see Chapter
18 for a fuller discussion of these issues).17 An initial independent evaluation of
HealthSpace has recently reported.18
The Spine
The Spine is the national database of key information about patients' health and
care. It forms the quintessential core of the NHS CRS, but will also support
other key elements of the Programme such as Choose and Book and ETP,
each of these using the Spine's messaging capabilities as part of their own
services. Key components making up the Spine include:
•
Personal Demographics Service (PDS), which facilitates unique linkages
of patient data from different sources. As the name implies, this contains
patient’s demographic information, including a unique NHS Number,
name, address and date of birth.19 The PDS does not however contain
any clinical data or information that may be considered particularly
sensitive, such as on ethnicity or religion.
•
Secondary Uses Service (SUS), which provides timely, anonymous
patient data for non-clinical care purposes. This includes commissioning,
public health, policy and planning and research (e.g. monitoring trends
of use of service).20 SUS aims to support national initiatives, such as
Payment by Results (PbR).
Where possible, SUS captures data
automatically through the Programme’s NHS operations systems. The
initial content for SUS will be person-specific, building on existing flows
such as commissioning datasets, cancer waiting times, clinical audit,
central returns and supporting demographic data. It has scope in the
future however to include non-person-specific data such as finance,
workforce and estates.
•
Access Control Framework, which controls access to clinical records of
patients and registers and authenticates all users storing type of
relationship between healthcare professionals and patients, as well as
87
patient preferences on information sharing (for example, whether
sensitive information is restricted from sharing).21
National Network for the NHS
The National Network for the NHS, better known as N3, connects all NHS
organisations and provides the NHS’s IT infrastructure, network services and
broadband connectivity requirements.22 The networking solutions provided by
N3 has now delivered the systems and services that enable the fast, secure
sharing of information, files and data between NHS sites (see Chapter 5 for
further details on health information exchange and interoperability). N3 is one of
the largest virtual private networks (VPN) in the world.
Fast access to up-to-date patient records, the streamlining of clinical practice
and a reduction in administrative tasks are all expected from the combination of
the latest networking technology and the bespoke applications of the
Programme. N3 provides support for NHS organisations in implementing new
services, such as the use of video-conferencing for appointments with
consultants. It is intended that it will also offer substantial savings on the cost of
telephony by enabling NHS organisations to converge their voice and data
networks.
3.4.2 Choose and Book
Driven by the Government’s agenda to give public and patients more choice,
Choose and Book has been introduced as a national electronic referral service
to allow patients a choice of place, date and time for their first out-patient
appointment. Patients can, if appropriate, choose a hospital or clinic, and book
their appointment to see a specialist either in conjunction with a member of the
practice team or, if more convenient, from home either by telephone or over the
Internet.23
When a GP refers a patient to a specialist, Choose and Book helps identify
relevant clinics together with details about availability.
After considering
88
individual preferences and clinical requirements, members of the primary care
team can assist patients in making their appointment at the time of referral.
Patients then receive confirmation of the time, date and location of their
appointment. Some patients may want or need more time to consider their
choices and, if so, they can take the appointment request letter away with them
and book the appointment later, either online or over the phone. Choose and
Book thus allows them to discuss options with family members and make any
necessary arrangements before scheduling the appointment.
The rollout of Choose and Book began with early adopter sites in the summer of
2004, before being rolled-out more widely in 2005 and 2006.
Its
implementation has now been completed and the majority of patients are now
referred for their initial consultant assessment via Choose and Book.
Some independent evaluations of Choose and Book have now been
published.24 25
3.4.3 ePrescribing
In 2003, NPfIT commenced a stakeholder engagement and planning process in
relation to perhaps one of the most important eHealth applications, namely
implementation of electronic prescribing (hereafter referred to as ePrescribing)
in hospitals. ePrescribing is defined by NHS CFH as ‘…the utilisation of
electronic systems to facilitate and enhance the communication of a
prescription, aiding the choice, administration or supply of a medicine through
decision support and providing a robust audit trail for the entire medicines use
process’.26 As defined, NHS CFH’s vision of ePrescribing clearly incorporates
computerised decision support system functionality (CDSS) (see Chapters 8
and 10 for detailed discussions on CDSS and electronic prescribing).
While UK GPs have been using ePrescribing for nearly two decades (with
varying degree of functionality, but constantly increasing in sophistication), this
has not been the case within the hospital sector in the UK.27
Diversity of
89
different clinical specialty requirements and complexity of prescribing in
hospitals have been seen as major barriers to implementation.
ePrescribing should, once implemented, facilitate wider improvements in
prescribing and administration processes, including reductions in paperwork,
improved audit trails for medication and enhanced communication (for example,
between hospital departments and pharmacies).
ePrescribing, with its built-in functionalities—such as decision and knowledge
support, alerts for contra-indications, allergic reactions and drug interactions,
formulary guidance or management, training and guidance for prescribers,
reminders, and audit trails of medication administration—represents one of NHS
CFH’s key activities designed to reduce risk of iatrogenic harm and thereby
improve patient safety. Ultimately, the goal is to integrate it seamlessly with
EHRs and other computer systems.
3.4.4 Electronic Prescription Service
As described above, virtually all UK GPs are already using ePrescribing and
use this technology to issue approximately 3.4 million prescriptions on each
working day. Given that the annual number of prescriptions issued in England
is expected to grow at about five per cent per year, it is clear why the ability to
automate repeat prescribing has proven so popular with GPs.
Building on these existing primary care computer capabilities, EPS introduces
important new add-on functionality.
Instead of printing ‘electronically
prescribed’ prescriptions (i.e. prescribed with the support of ePrescribing
system), GPs will be in a position to electronically send the prescription to the
dispensing pharmacist. The new system should lead to significant time savings,
especially insofar as it enables the more efficient processing of repeat
prescriptions, which now make up about 70 per cent of all prescriptions issued.
The sending of repeat prescriptions using EPS will in due course happen (semi) automatically.
90
These electronic prescriptions are received by the EPS, a secure information
hub, from which they can then be downloaded by participating dispensers. True
paperless prescribing will occur in a later release, when patients will be able to
choose (or “nominate”) a particular dispensing contractor, i.e. pharmacy to
automatically receive their electronic prescriptions.
In time, the electronic
submission of reimbursement claims by these dispensers will also be supported
by the EPS system.
The EPS is being introduced in primary care settings, such as GP practices and
community pharmacies, across England, and will continue to expand to other
settings in which primary care prescribing takes place. The first iteration of the
EPS, known as Release 1, which is mostly invisible to patients and end users,
went live in February 2005.28 Progress is behind schedule, however; Release 2,
which will be in two main phases, has now begun: Stage one will be the
'transition' phase, and stage two will be the 'full EPS' phase.29
30
In the
‘transition’ phase, prescribers and dispensers will gradually begin replacing
most paper prescriptions with electronic prescriptions. This is when intended
EPS-related benefits should really begin to be demonstrated as it becomes
possible for patients to nominate dispensers to receive their electronic
prescriptions and for dispensaries to electronically submit reimbursement
claims. Some hand-signed FP10 prescription forms will continue to be required
in this phase as only those dispensers working with nominated prescriptions will
be able to accept electronically-signed prescriptions. By the ’full EPS’ phase,
most sites will be using Release 2. At this point, all within-scope prescriptions
(e.g.
controlled drugs). will be able to be sent, signed and received
electronically— whether nominated or not.
The plan is that in time, the EPS will be integrated with NHS CRS. Information
on what has been prescribed and dispensed will be able to be automatically
recorded in the patient’s care record. Access to more accurate (or entirely new)
information about what has been prescribed and dispensed will be of great
91
assistance to the healthcare professionals who need it (for example, a doctor in
an accident and emergency department or a GP a patient visits while away from
home), thus helping to improve standards of care.
NHS Connecting for Health Evaluation Programme has commissioned an
evaluation project to evaluate the impact of the new service on quality and
safety of care, which will report later in 2011.31
3.4.5 Picture Archiving and Communications Systems
Picture Archiving and Communications Systems (PACS) capture, store,
distribute and display static or moving digital images such as x-rays or scans,
thereby potentially enabling more efficient diagnosis and treatment. The ability
to store digital images will form an essential part of every patient’s electronic
record, thereby potentially removing the need to print on film and to file or
distribute images manually.
These images can then be sent and viewed at one, or across several, NHS
locations. This should in turn result in increased capacity of diagnostic services
and should furthermore speed up the time to diagnosis.
For patients, this
should also mean fewer delays, fewer wasted appointments due to lost or low
quality images, and less retesting, which in turn should reduce the total lifetime
radiation dose. Patient care also has the potential to benefit as clinicians and
care teams have the ability to work together viewing common information
across one or more locations. PACS has been procured to provide full access
to digital images in NHS organisations throughout England and its
implementation is now complete. An initial evaluation of PACS has reported.32
3.4.6 The NHSmail Email and Directory Service
Historically, there have been many different local email systems operating in the
NHS. The quality and reliability of these services varied substantially and they
incurred substantial costs.
In addition, none of the services were secure
enough to allow the transmission of patient information resulting in information
92
being sent frequently via mail or fax incurring further costs for paper, printing
and postage as well as slowing down the process.
The NHSmail service was launched in October 2004 to replace these disparate
email services with a standardised platform that was capable of being used for
the transmission of patient information. NHSmail provides a central, secure
email service, thus reducing the overall email costs to the NHS and providing a
swift and secure means of exchanging information across the NHS.33
The NHSmail platform is a centrally-funded service that is free at the point of
use, and offers a helpdesk that is available to support users 24 hours a day,
seven days a week. It preserves patient confidentiality and is protected by both
anti-virus and anti-spam software. NHSmail is the only system approved by the
DH, British Medical Association (BMA) and the Royal College of Nursing (RCN)
for NHS users to securely exchange patient data within the NHS. NHSmail
aims to improve workflow by allowing authorised users to log-on from
anywhere, at any time, giving them access to the NHS Directory, with contact
details for over one million NHS staff. NHSmail also includes features that allow
users to send free SMS and fax messages directly from the system, and to
share calendars and mailboxes with other NHSmail users.33 It has been
upgraded and is now based on Microsoft’s Exchange 2007, but still only
provides limited storage space (20Mb).
3.4.7 Support for primary care
GP Systems of Choice
This is the NHS CFH scheme through which the NHS funds the provision of GP
clinical IT systems in England. GP Systems of Choice (GPSoC) allows practices
and PCTs to benefit from a range of quality assured GP clinical IT systems
purchased from existing suppliers who are contracted to work within the NPfIT.
Practices can choose between systems provided by their local service provider
or by suppliers that are contracted to offer systems on the GPSoC Framework.
93
GPSoC introduces standards which will improve the quality of service that
practices receive from their GPSoC framework supplier.
The Quality Management and Analysis System (QMAS)
The new GMS contract was introduced in April 2004. A key component was the
Quality and Outcomes Framework (QOF) of national achievement targets
describing how GP practices would be rewarded financially, based on their
achievement in up to four domains: clinical; organisational; patient experience;
and additional services.
To support the QOF and the GMS contract, the Programme developed a single,
national IT system, known as Quality Management and Analysis System
(QMAS). QMAS tracks the performance of each practice, comparing it against
QOF goals allowing for the payment of financial incentives based on quality of
care.34 The system is also used to compute national disease prevalence rates
for a range of conditions.35 QMAS last underwent a major upgrade in August
2006, in support of the 2006–07 GMS contract.34
GP2GP
When patients transfer from the care of one GP to another, it is clearly
important that their medical records be transferred as swiftly as possible.
Ensuring that medical records arrive in time for the patient’s first consultation
may have significant clinical benefits, as it has been estimated that not having
this information could threaten the quality of care in over 50 per cent of
consultations.36 Using the traditional paper-based system, the transfer of patient
records may take weeks or months leaving clinicians without the information
they need to provide the best possible care for patients.36
GP2GP is designed to address this problem. It enables the transfer of the
electronic component of a GP patient health record. When a patient registers
with a new practice, GP2GP will move the patient’s current EHR from their
previous general practice to their new one. On receipt of the EHR, the new
94
practice will undertake certain housekeeping activities, e.g. authorise current
repeat prescriptions listed in the EHR, and will have the clinically important
medical history in hand at the time of the patient’s first consultation (see
Chapter 5).36 The GP2GP system includes a feature called “autosend”, which
allows the EHR to be sent without any action on the part of the sending GP.
Not only does this cut down on the administrative cost to the sender, but it also
ensures that the patient’s new GP will receive the records quickly, within
minutes or hours.37
The GP2GP system met its March 2007 implementation goal of achieving rollout in 500 practices and is currently engaged in a nationwide rollout among
practices meeting certain 'entry and readiness criteria’.36 Current estimates
suggest that 60 per cent of practices are now using GP2GP with an average of
10,000 transfers taking place each week.38
Apart from the aforementioned NPfIT deliverables, NHS CFH is involved in a
number of
other work-streams
and
deliverables
with
regards
to
IT
implementation.
3.5 An information revolution?
The formation of the new coalition government in May 2010–with its mandate to
usher in a new austerity phase for public services–has resulted in a change in
direction for NPfIT (see Box 3.2). Whilst the details of the new direction have
yet to be agreed, it seems clear from the overall strategy document, Equity and
Excellence: Liberating the NHS, and the consultation White Paper, Liberating
the NHS: An Information Revolution, that central IT expenditure will be
substantially scaled back, but there will remain a commitment to maintaining the
centralised infrastructure of the NPfIT, in particular N3, the SCR, and the
Spine.3
4
There is also a clear commitment to expanding the role of
telehealthcare initiatives allowing healthcare to be delivered remotely, and
enabling self-care.
The key underpinning thrust of the consultation paper
however is for more localised IT decision-making and responsibility, and the far
95
greater release of data to support all aspects of healthcare delivery and
research.
3.6 Conclusions
NHS CFH and its national Programme are complex ambitious endeavours that
have impacted on many aspects of NHS care provision. Whilst many of the
infrastructure-related aspirations have been delivered, the deliverables that
have the potential to impact more substantially on day-to-day patient care such
as the SCR, DCR, HealthSpace and EPS have proved more challenging to
deliver. Whether the change in philosophy from a “replace all” to a “connect all”
approach ushered in by the change of government proves more successful
remains to be seen. Initially, the core aspirations of this modernisation agenda
relate to improving data storing and management and supporting professional
decision-making; more recently, this focus has expanded to also include the
delivery of care from a distance. We focus on these three issues in the next
chapter and explore the potential these have to impact on the quality and safety
of healthcare provision.
96
Box 3.2 DH statement on the future of the NPfIT, 9 September 201039
‘A Department of Health review of the National Programme for IT has concluded
that a centralised, national approach is no longer required, and that a more locallyled plural system of procurement should operate, whilst continuing with national
applications already procured.
A new approach to implementation will take a modular approach, allowing NHS
organisations to introduce smaller, more manageable change, in line with their
business requirements and capacity. NHS services will be the customers of a more
plural system of IT embodying the core assumption of ‘connect all’, rather than
‘replace all’ systems. This reflects the coalition government’s commitment to ending
top-down government and enabling localised decision-making.
The review of the National Programme for IT has also concluded that retaining a
national infrastructure will deliver best value for taxpayers. Applications such as
Choose and Book, Electronic Prescription Service and PACS have been delivered
and are now integrated with the running of current health services. Now there is a
level of maturity in these applications they no longer need to be managed as
projects but as IT services under the control of the NHS. Consequently, in line with
the broader NHS reforms, the National Programme for IT will no longer be run as a
centralised national programme and decision making and responsibility will be
localised.’
References
1. Department of Health. Reconfiguring the Department of Health's Arm's Length
Bodies. London: HMSO, 2004.
2. Department of Health. Arm's length bodies review. Available from:
http://www.dh.gov.uk/en/Aboutus/Deliveringhealthandsocialcare/Armslengthbodie
s/DH_4105578 (last accessed 4/1/11).
3. Department of Health. Equity and Excellence: Liberating the NHS. London:
HMSO, 2010.
4. Department of Health. Liberating the NHS: An Information Revolution. London:
HMSO, 2010.
5. Department of Health. Information for health an information strategy for the
modern
NHS.
Available
from:
97
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
http://www.dh.gov.uk/en/PublicationsAndStatistics/LettersAndCirculars/HealthSer
viceCirculars/DH_4005016 (last accessed 4/1/11).
Department of Health. The NHS Plan: A Plan for Investment, a Plan for Reform.
London: Department of Health, 2000.
Department of Health. Building the Information Core: Implementing the NHS Plan.
London: Department of Health, 2001.
Wanless D. The Wanless Report. London: HMSO, 2002.
Department of Health. Delivering the NHS Plan: next steps on investment, next
steps on reform. London: Department of Health, 2002.
Department of Health. Delivering 21st century IT support for the NHS: national
strategic programme. London: Department of Health, 2002.
Revisions to the GMS contract 2006/07 Delivering investment in general practice.
Available
from:
http://www.bma.org.uk/employmentandcontracts/independent_contractors/genera
l_medical_services_contract/NewDEs0706.jsp (last accessed 4/1/11).
Robertson A, Cresswell K, Takian A, Petrakaki D, Crowe S, Cornford T, et al.
Implementation and adoption of nationwide electronic health records in secondary
care in England: qualitative analysis of interim results from a prospective national
evaluation. BMJ 2010; 341: c4564.
NHS Connecting for Health. Review agrees value of a core Summary Care
Record http://www.connectingforhealth.nhs.uk/newsroom/news/valueofscr (last
accessed 3/1/11).
NHS Connecting for Health. Patient Information leaflet: The NHS Care Records
Service Better information for better, safer care. Available from:
http://www.stokepct.nhs.uk/pdfs/scr/easy_read_summary_leaflet.pdf
(last
accessed 4/1/11).
NHS
Connecting
for
Health.
Available
from:
http://www.nhscarerecords.nhs.uk/faqs/ (last accessed 4/1/11).
NHS Connecting for Health. NHS Care Records Service Patient Video. Available
from:
http://www.connectingforhealth.nhs.uk/v/?v=nhscfhcrsvideo.mp4&w=320&h=255
(last accessed 4/1/11).
National Information Governance Board for Health and Social Care. The NHS
Care
Record
Guarantee.
Available
from:
http://www.nigb.nhs.uk/guarantee_guarantee.pdf (last accessed 4/1/11).
Greenhalgh T, Hinder S, Stramer K, Bratan T, Russell J. Adoption, non-adoption,
and abandonment of a personal electronic health record: case study of
HealthSpace. BMJ 2010; 341: c5814.
NHS Connecting for Health. The Personal Demographics Service. Available from:
http://www.connectingforhealth.nhs.uk/systemsandservices/demographics/pds
NHS Connecting for Health. Secondary Uses Service. Available from:
http://www.connectingforhealth.nhs.uk/systemsandservices/sus/ (last accessed
4/1/11).
NHS
Connecting
for
Health.
Spine
Factsheet.
Available
from:
http://www.connectingforhealth.nhs.uk/resources/systserv/spine-factsheet
(last
accessed 4/1/11).
NHS Connecting for Health. N3. Available from: http://www.n3.nhs.uk/ (last
accessed 4/1/11).
Choose
and
Book
Website.
Available
from:
http://www.chooseandbook.nhs.uk/staff/started/getting-started-overview
(last
accessed 4/1/11).
98
24. Rashid M, Abeysundra L, Mohd-Isa A, Khan Y, Sismeiro C. Two years and 196
million pounds later: where is Choose and Book? Inform Prim Care 2007; 15:
111-19.
25. Rabiei R, Bath PA, Hutchinson A, Burke D. The National Programme for IT in
England: clinicians' views on the impact of the Choose and Book service. Health
Informatics J 2009; 15:167-78.
26. NHS Connecting for Health. ePrescribing FAQs. Available from:
http://www.connectingforhealth.nhs.uk/systemsandservices/eprescribing/faqs
27. Crowe S, Cresswell K, Avery AJ, Slee A, Coleman JJ, Sheikh A. Planned
implementations of ePrescribing systems in NHS hospitals in England: a
questionnaire study. JRSM Short Rep 2010; 1: 33.
28. NHS Connecting for Health. Electronic Prescription Service (EPS). Available
(last
from: http://www.connectingforhealth.nhs.uk/systemsandservices/eps
accessed 4/1/11).
29. NHS Connecting for Health. Introducing release 2. Available from:
http://www.connectingforhealth.nhs.uk/systemsandservices/eps/introduced/releas
e2 (last accessed 4/1/11).
30. NHS Connecting for Health. EPS Release 2 national rollout. Available from:
http://www.connectingforhealth.nhs.uk/systemsandservices/eps/staff/implementati
on/release2/rollout (last accessed 4/1/11).
31. Lichtner V, Petrakaki D, Hibberd R, Venters W, Cornford A, Barber N. Mapping
stakeholders for system evaluation - the case of the Electronic Prescription
Service in England. Stud Health Technol Inform. 2010;160(Pt 2):1221-5.
32. Collin S, Reeves BC, Hendy J, Fulop N, Hutchings A, Priedane E. Implementation
of computerised physician order entry (CPOE) and picture archiving and
communication systems (PACS) in the NHS: quantitative before and after study.
BMJ 2008; 337: a939.
33. NHS
Connecting
for
Health.
About
NHSmail.
Available
from:
http://www.connectingforhealth.nhs.uk/systemsandservices/nhsmail/about
(last
accessed 4/1/11)
34. NHS Connecting for Health. What is QMAS? Available from:
http://www.connectingforhealth.nhs.uk/systemsandservices/gpsupport/qmas (last
accessed 4/1/11).
35. Connecting for Health. QMAS IT - QMAS getting ready for the end of the year.
Available
from:
http://www.connectingforhealth.nhs.uk/systemsandservices/gpsupport/qmas/qma
sendofyear.pdf (last accessed 4/1/11).
36. NHS Connecting for Health. GP2GP - Electronic patient record transfer. Available
from:
http://www.connectingforhealth.nhs.uk/systemsandservices/gpsupport/gp2gp/impl
ementation/3914.pdf (last accessed 4/1/11).
37. NHS Connecting for Health. Explaining autosend. Available from:
http://www.connectingforhealth.nhs.uk/systemsandservices/gpsupport/gp2gp/how
doesitwork/autosend (last accessed 4/1/11).
38. NHS
Connecting
for
Health.
GP2GP.
Available
from:
http://www.connectingforhealth.nhs.uk/systemsandservices/gpsupport/gp2gp/pro
gress (last accessed 4/1/11).
39. Department of Health. The future of the National Programme for IT Available
from:
http://www.dh.gov.uk/en/MediaCentre/Pressreleases/DH_119293
(last
accessed 4/1/11).
99
Chapter 4
Exploring, describing and integrating the fields of
quality, safety and eHealth
Summary
•
eHealth is a relatively new and rapidly evolving field and so many of the
concepts, terms and applications are still evolving.
•
There is furthermore no agreed definition of eHealth, with some
researchers using this to relate primarily to the area of consumer
informatics, whereas others use it more generically to refer to any of the
ways in which information technology can be employed to improve
delivery of healthcare.
For the purposes of this report, we considered it
important to use an inclusive definition and chose to use Eysenbach’s
definition, as adapted by Pagliari, as the basis for our work:
characterises not only a technical development, but also a new way of
improve
healthcare locally,
regionally
and
worldwide
by using
information and communication technology.’
•
Whilst the number of eHealth applications is very large, these can
conceptually be divided into three broad domains relating to key activities
they support:
o storing, managing and sharing data
o informing and supporting clinical decision-making
o delivering expert professional and/or consumer care remotely.
•
Most frameworks of quality currently in use incorporate the following key
dimensions of care: effectiveness of treatments; appropriateness of the
means of delivery; acceptability; efficiency; and equity.
100
•
Whilst there are no internationally agreed definitions of patient safety,
adaptations of the National Patient Safety Agency’s definition of Patient
Safety Incidents is increasingly being used. This, in its original definition,
states that a ’…patient safety incident is any unintended or unexpected
incident which could have harmed or did lead to harm for one or more
patients being cared for by the NHS.’
•
There are a number of patient safety taxonomies; however, our scoping of
this literature found that the Joint Commission on Accreditation of
Healthcare Organizations Patient Safety Event Taxonomy and the related
World Health Organization’s Classification are the most comprehensive
and clinically relevant in that they incorporate five key areas:
o impact of medical error
o type of processes that failed
o domain, i.e. the setting in which an incident occurred
o cause or factors leading to the safety incident
o prevention and mitigation factors to reduce risk of recurrence
and/or improve outcomes in the case of a further incident.
•
Integrating the fields of eHealth, quality and safety clearly demonstrates
the numerous ways in which information technology has the potential to
improve the effectiveness and efficiency of many facets of healthcare
delivery. This includes, for example, helping clinicians to access
comprehensive information on their patients, aiding monitoring of their
conditions and the treatments being issued, reducing inappropriate
variability in healthcare delivery, and proactively identifying and alerting
clinicians to threats to patient safety.
•
The integration of these domains however also highlights ways in which
introduction of new eHealth applications can result in unintended
consequences and inadvertently introduce new risks.
101
4.1 Introduction
In order to help identify the focus of our work, during the formative phase of
work we undertook mapping of the domains of eHealth, quality and safety,
looking particularly for areas of intersection between these fields of enquiry.
This proved challenging for several reasons, not least that eHealth, quality and
safety are relatively novel and at times contested fields of enquiry. This chapter
summarises our attempts at this underpinning conceptual work and provides the
framework within which we interpreted the findings chapters reported in Section
2 of this volume.
The analytic framework that we have developed thus draws on the following
main strands:
•
Exploration of notions and constructs of quality and its assessment
•
A classification of patient safety risks, derived from the World Health
Organization’s (WHO) International Classification for Patient Safety1 and
a systematic overview of safety taxonomies by the Joint Commission on
Accreditation of Healthcare Organizations2
•
The core applications of eHealth, represented within the conceptual
model developed by Pagliari et al.3
•
The headline deliverables of NHS Connecting for Health (NHS CFH) and
other key elements of the National Programme for Information
Technology (NPfIT) (See Chapter 3).
These strands encompass a number of cross-cutting themes around
behavioural and organisational change, usability and interoperability of
technology and its evaluation.
4.2 Quality of healthcare
Modern medicine faces important challenges. Never before has scientific
development impacted so profoundly on human health and our knowledge on
how to impact on disease and health grown so rapidly.
Several hundred
thousand scientific articles are now published each year that in one way or
102
another propose improvements to ways of improving health and/or the delivery
of healthcare. Since the first randomised controlled trial (RCT) conducted over
60 years ago, the number has now grown to over 10,000 annually.4 5 This new
evidence in turn creates a constant need for constant questioning of the
evidence underpinning medicine and, in many cases, change.
Despite these advances in our ability to use new medicines, procedures and
approaches to improve the health of individuals and populations, there is a
growing body of knowledge that demonstrates that translating these insights
into practice frequently fails.6-10 Over the last decade, hundreds of authoritative
studies have clearly revealed the very considerable variations in the
performance of healthcare professionals and indeed health systems;6-10 the net
effect of this variation is that, even in countries with mature health systems,
typically only about 50 per cent of the population receive optimal standards of
care.11 Moreover, there is the continuing paradox of the inverse care law,
describing how those most in need are typically least likely to receive care.12 13
There is a long history of interest in the quality of care.14 Over forty years ago,
Avedis Donabedian,15 widely considered the father of the quality movement,16 17
invigorated focus on quality of healthcare as a measure of the pursuit of best
care. In a broader sense, the focus on quality is a result of a complex
convergence of scientific, political, economic and social issues. But despite this
long history of focusing on quality considerations, the response by the medical
profession has tended to be mixed.18
Maximising patient safety is an essential component of the wider quality
improvement agenda. Although these topics are often considered separately,
reports such as the Institute of Medicine’s Crossing the Quality Chasm8 and To
Err is Human7 closely aligned safety and quality, clearly emphasising the
potential of information technology (IT) to improve both the quality and safety of
care. Of note is that analyses of safety incidents have in a sense served as a
catalyst to the integration of the fields of informatics, human factors engineering,
103
and cognitive and social psychology, bringing these within the fold of the quality
agenda.19
4.2.1 Defining and measuring the quality of healthcare
Experts have struggled to formulate a concise, meaningful, and generally
applicable definition of quality as it pertains to healthcare.18 As with safety
(discussed below), there remains no single agreed definition of quality, although
many have attempted to elucidate the concept and many broad frameworks for
understanding it have been proposed.15 20-31 While trying to capture the essence
of this elusive construct, all definitions agree that quality care is that which helps
an individual and/or population to maximise their welfare, duration of life, and
which leads to desired health outcomes. For example, Donabedian defines it in
terms of the structures, processes and outcomes of care;15 Maxwell20 in terms
of accessibility, effectiveness, efficiency, acceptability, equity and relevance;
and Schuster in terms of the underuse, misuse and overuse of services.23
Others have presented a detailed list of condition-specific quality of care
indicators.30 The simple and widely cited framework developed by Campbell et
al. draws on several of these characterisations and considers quality in terms of
its relevance to the individual—access and effectiveness (of clinical care and
inter-personal care) and the public or populations—access, effectiveness,
equity, efficiency and cost.25
We have integrated these within the conceptual
map shown as Figure 4.1, which highlights the core domains of effectiveness,
efficiency, equity, acceptability and access, illustrating their inter-dependencies
(shown by the arrows and repetition of sub-terms).
This integration also
recognises the central role of effectiveness in both the quality and safety
debate, echoing the assertion that ‘…quality-improvement efforts based on
evidence of effectiveness are likely to be more readily embraced and may save
more lives than will efforts to improve safety that cleave to the concept of
accidental death and lack a solid evidence-base.’19 Cost is subsumed within
efficiency, although the effectiveness, equity, accessibility and acceptability of
care also have downstream cost implications. One of the most widely adopted
definitions of quality is the Institute of Medicine’s, which defines quality as the
104
‘…degree to which health services for individuals and populations increase the
likelihood of desired health outcomes and are consistent with current
professional knowledge.’8 This is the definition we have adopted in our report.
Figure 4.1 Map of quality concepts
Efficiency
Streaming processes
Increasing access (to data)
Reducing costs
Acceptability
Enhanced patient
satisfaction
communication,
empowerment
Effectiveness
Clinical appropriateness
(evidence-base + tailoring)
Clinical impact
Equity
Enhanced access
Reduced variation
Access
To services, knowledge,
data, people, support
105
Measures and indicators of quality of care
Lilford et al. highlight how:32
‘The distinction between a measure of quality and an indicator of quality is
important. Generally speaking we have very few real measures of quality. For
example, post-operative length of stay is a measure of the patient's hospital stay,
but only an indicator of quality, e.g. a patient's long stay might represent
postoperative complications or poor discharge arrangements. Thus, the term
indicator is preferable.’
Quality indicators assess aspects of care associated with patient outcomes in
a variety of domains. Broadly speaking, indicators can be summarised into
those that assess the technical provision of care and the interactions between
patients and providers. Concepts such as effectiveness, access, capacity,
safety, patient-centredness, equity and disparities are commonly used to
describe the technical or inter-personal aspects of care.
Measuring the quality of care or the impact of interventions aiming to improve
quality is thus far from simple33 and needs adjusting for, amongst other
things, treatment refusal.34 A further complication comes from the observation
that different methods of measurement of the quality of care do not
necessarily provide the same answers.32
35
This complexity (and in some
case vagueness) of definitions of quality means that researchers and
practitioners often find it difficult to understand and relate to it. The very
language and terminology of quality of care leaves many healthcare
professionals somewhat bewildered. Terms such as desired patient
outcomes, patient-reported outcome measures, criteria, standards, profiles,
observed and expected mortality, case-mix and case-severity adjustments,
charts, tables, leagues, quality control, continuous quality management,
quality improvement, quality assurance are often poorly understood.
Moreover, what constitutes quality often changes depending on the goal of
the measurement.36
106
One overarching benchmark for quality however, as defined by the Institute of
Medicine, is provision of care according to current best knowledge. Broadly
speaking, measures of quality can be classified into those that measure
processes and those that measure outcomes. As a process, the quality of
care can be measured according to whether patients were offered
recommended services. For example, a study of processes found that quality
of care increased as the number of long-term conditions experienced by
patients increased.37
Process measures are really only valid when there is a proven relationship
with health outcomes; such process measures are often termed “intermediate
clinical outcomes” because they mediate the association between the process
and outcomes of care. For example, the relationship between glycaemic
control and risk of diabetes-related complications is well established and so
HbA1C is in people with diabetes a valid process measure of quality. Another
commonly measured process measure of quality of care is rate of admissions
to hospital.38
For example, a high rate of emergency or unplanned
admissions to hospital is generally considered to be a reflection of suboptimal
community-based patient care; and admission rates for so-called “ambulatory
care sensitive conditions” have been proposed as a marker of the quality care
received by patients with these conditions in the community. However, in high
income countries such as the UK, which have well-developed health systems
offering universal coverage, the impact of socio-economic factors (such as
deprivation) and health behaviours (such as smoking) are likely to outweigh
the influence of healthcare factors on unplanned hospital admissions for longterm illnesses.39
Outcome measures of quality are on the other hand only appropriate to study
if they are modifiable as a result of a process that aims to improve this
outcome. So deaths from causes that are not modifiable cannot be
considered measures of quality of care. Alternative measures, such as
premature deaths from long-term illnesses (for example, deaths from
107
coronary heart disease or stroke in people under 65 years of age) have been
proposed as markers of the quality of health systems. In-hospital death rates
in various forms, (e.g. standardised hospital mortality ratios) have also been
proposed as markers of quality or safety. However, their use has been
criticised because in-hospital deaths will be strongly influenced by variations
in hospital case-mix. Furthermore, the methods used to standardise inhospital mortality are strongly influenced by the coding of secondary
diagnoses in hospital records and this can introduce artefactual errors and
bias in the estimates of standardised hospital mortality rates.40
There is also now a clear recognition of the importance of patients’ or
consumers’ perspectives into what constitutes quality healthcare. Language
has adapted to reflect patient-centredness and researchers now routinely talk
about “desired health outcomes”, (i.e. those desired by the patient and not
necessarily doctor) and “patient expectations”.
Variability in measured quality of care poses a great challenge for policymakers,
healthcare professionals
and patients
alike
with
outcomes
’…influenced by definitions, data quality, patient case-mix, clinical quality of
care and chance.’ 32 For example, an association between a higher volume of
activity and improved outcomes in hospital care is now supported by evidence
from more than 300 studies across a wide range of procedures and
conditions.41 Nevertheless, evidence is far from uniform.42-44 The clinical and
policy significance of these findings is complicated by the methodological
shortcomings of many studies. Differences in case mix and processes of care
between high- and low-volume providers may explain part of the observed
relationship between volume and outcome. That said, it is far from clear how
policymakers should act on the finding that high volume of activity is
associated with better outcomes. This research has also been largely
hospital-based, and the association between provider size or volume of
activity and outcomes in primary care is less well understood.
108
While measurable differences do not necessarily translate into clinically
meaningful differences in patients' lives,45
46
measuring quality of care is
nonetheless an essential prerequisite to quality improvement. Only by
measuring outcomes of quality improvement initiatives and interventions can
we know whether they actually result in an improvement. The UK Quality and
Outcomes Framework (QOF) is an example of a national policy initiative
aimed at measuring the quality of primary care.47 In 2006, in an attempt to
improve the quality of care, the US Congress signed a pact with the American
Medical Association which pledged to develop over 100 standard measures
of doctors performance that will be reported to the federal government.48
4.2.2 Improving quality of care
Problems with quality have traditionally been classified into three areas:
shortage
of
technical and/or interpersonal competence;
overuse
or
unnecessary or inappropriate use of services; and underuse of or lack of
access to needed and appropriate services. Appropriateness is thus at the
heart of defining quality of care (see Figure 4.1).49 Yet defining what is
appropriate is far from easy50;51 and indicators of quality of care cannot simply
be transferred between countries or indeed clinical settings as variation in
professional culture or clinical practice may be important confounders.52
Improving quality of care is complex and includes approaches such as:
implementation of evidence-based practice (where possible catalysed by the
development of clinical practice guidelines for example);53 leadership
commitment;54
continuous
quality
improvement;55
continuing
medical
education and professional development;56 regulation; assessment and
accountability; competition, continuing professional development;57
and feedback;
59
58
audit
and patient empowerment through, for example, creating
public pressure by publishing performance data.60 61 The UK is now seen as
an international leader in investigating the impact of financial incentives to
drive quality improvement initiatives, which overall seem to have relatively
modest short-term effects.47;62 An example of a national initiative is The
109
100 000 Lives Campaign and building on this, the 5 Million Lives
Campaign.63;64
Although attempts to improve quality are numerous, the strength of the
evidence in support of most of these initiatives is either weak or absent.6;65
We lack an understanding of which approaches are most appropriate for what
types of improvement in what settings and of the determinants of successful
performance change and none of the popular models for improving clinical
performance appear to be superior.6 Furthermore, it has only recently been
recognised that quality and quality improvement initiatives, as well as
schemes to improve patient choice, need to include mechanisms to also
measure and monitor potential harms.46
4.3. Safety
Approximately 850,000 medical errors occur in NHS hospitals every year
resulting in an estimated 40,000 deaths.66;67 These are, by any standard,
shocking statistics and reflect an unacceptable state of healthcare.68
Safety is an intricate feature of high quality care although it has only begun to
receive attention as a quality parameter in recent years.
The Institute of
Medicine’s landmark reports on healthcare safety and quality, To Err Is
Human7 and Crossing the Quality Chasm8 and equivalent reports in the UK66
69
have articulated a broad agenda for quality and safety improvement in
healthcare that needs to be based on evidence,70 recognising the current
unacceptable state of identifying, analysing and learning from medical
mishaps.71
4.3.1 Defining the safety of healthcare
One of the most fundamental challenges of patient safety is the variety of
different concepts underlying this broad construct and the lack of a single
agreed definition; the range and number of classification schemes72-81
characterising incident reporting systems and patient safety research reflects
110
these underlying tensions. Diverse understanding of patient safety related
terminology82-84 impacts importantly on how we are able to learn from
potentially dangerous events85 when attempting to interpret or synthesise the
results of published studies, which address somewhat different theoretical
propositions, (e.g. avoidance of error, risk prevention, reporting or
investigation of incidents) and use different outcome measures. Patient safety
encompasses aspects such as harm to patients, incidents that lead to harm,
the antecedents or processes that increase the likelihood of incidents, and the
attributes of organisations that help guard against harm and enable rapid
recovery when risk escalates.86
Most definitions of patient safety and medical error recognise that
organisational factors interact with human factors to facilitate and mitigate
errors.87 There is, however, a tension between focusing on individual
practices and seeing safety primarily as a systems’ problem.88 The latter
approach focuses on developing systems that prevent errors and, equally
importantly, ensures that clinicians provide effective care.89;90
Shojania et al. have defined a patient safety practice as ‘…a type of process
or structure whose application reduces the probability of adverse events
resulting from exposure to the healthcare system across a range of diseases
and procedures.’91
In conceptualising patient safety we found it helpful to focus on its goal.
Battles and Lilford92 specify how the ‘…goal of patient safety is to reduce the
risk of injury or harm to patients from the structure and process of care. This
can be accomplished by eliminating or minimising unintended risks and
hazards associated with the structure and process of care’ and ‘…a vision for
patient safety would be zero health care associated injuries or harm.’
In an attempt systematically to organise and represent patient safety theory
we considered various other definitions and constructs of safety related
111
terminology.72-81;93 What became evident is that in the past efforts to define
and classify “errors” or “mistakes” were somewhat theoretically and
methodologically flawed.74 Contributions to learning from risk-averse
industries such as aviation94-96 have been seminal in recognising key safety
principles, such as: that errors can never be completely eradicated and
generally derive from faulty system design not from negligence; that major
safety incidents are only the ”tip of the iceberg” of procedures and processes
that indicate possibilities for organisational learning; and most importantly that
prevention of safety incidents should be an ongoing process based on open
and full disclosure and reporting. The key objective of designing safe systems
is to make it difficult for the individual to err and to design systems that absorb
errors, mistakes, slips and lapses that inevitably occur due to a characteristic
of human nature—fallibility.96
Among many safety models Reason's Swiss Cheese Model has become the
dominant paradigm for analysing medical errors and patient safety incidents
(See Figure 4.2).97 98 The holes in the defences arise for two reasons: active
failures (e.g. shortage in communication or improper ventilation technique);
and latent conditions (e.g. inadequate patient monitoring, or inadequate
staffing skills mix). Nearly all adverse events involve a combination of these
two sets of factors.99
Vincent et al. have adapted and further developed Reason’s model and
describe the people who are directly involved as the inheritors, rather than the
instigators of an accident sequence (see Figure 4.3).80 These frameworks
facilitated systematic and conceptually driven approaches to organisational
risk assessment.
112
Fig 4.2 The Swiss Cheese Model of accident causation—model of how
defences, barriers, and safeguards may be penetrated by an accident
trajectory
with permission from British Medical Journal.
Figure 4.3 The anatomy of an organisational accident
Reprinted from Vincent et al. (1998) with permission from British Medical Journal.
When we approach quality and safety as intricately related we then see how
safety incidents can lie either in the structure or the process of care. Figure
4.4 illustrates the process as occurring within the structure of healthcare.92
113
Figure 4.4 A structure and process model for patient safety based on
Donabedian
Reprinted with permission from J Battles
92
Quality and Safety in Health Care
Recognising the problem of multiple definitions of patient safety related
concepts, a number of organisations have advocated the use of a standard
taxonomy and terminology.101 The National Patient Safety Agency’s (NPSA)
definition of Patient Safety Incidents states that a ‘…patient safety incident is
any unintended or unexpected incident which could have harmed or did lead
to harm for one or more patients being cared for by the NHS.’ More recently
the WHO’s World Alliance for Patient Safety has produced a conceptual map
and classification scheme in collaboration with a number of high profile
international organisations, including the
NPSA.1 This builds on the
taxonomy developed by the Joint Commission on Accreditation of Healthcare
Organisations (JCAHO), on behalf of the WHO, which was based on
systematic review and expert consensus.74
75
Homogeneous elements of
models—which comprise terms and the relationships between terms that
make up the building blocks of a classification scheme—were categorised
into five complementary root nodes or primary classifications (See Box 4.1).
For our review we found the combination of the original JCAHO taxonomy2
and the WHO conceptual map and classification scheme1
102
to be most
useful for understanding the potential role of eHealth solutions in facilitating
114
safer patient care, as well as their unintended consequences (See Box 4.1).
These differentiate types of problem (communication, patient management,
clinical practice), cause (organisational, technical, human), impact (medical
and non-medical), issues for prevention (accuracy, communication, alarms)
and domain (setting, target, people).
In so doing, they recognise that
adverse events may result from multiple systemic features operating at
different levels, such as the task, the team, the work environment and the
organisation.80 They also recognise the importance of psychological and
human factors in the nature, mechanisms and causes of error and the fact
that liability to error is strongly affected by the context and conditions of work,
as has been argued by Leape and others.79 These are further discussed
below when combined and integrated with eHealth map in Section 4.5.
Box 4.1 Five complementary primary classifications of patient safety
developed by the JCAHO
1. Impact: The outcome or effects of medical error and systems failure,
commonly referred to as harm to the patient.
2. Type: The implied or visible processes that were faulty or failed.
3. Domain: The characteristics of the setting in which an incident occurred
and the type of individuals involved.
4. Cause: The factors and agents that led to an incident.
5. Prevention and mitigation: The measures taken or proposed to reduce
incidence and effects of adverse occurrences.
2
Chang et al. (2005) Reproduced with permission from International Journal Quality Health Care
4.3.2 Improving safety of healthcare
There are two main approaches to tackling the problem of human fallibility:
the person and the system approaches.98
Trying to
find the causes for
patient safety incidents is often messy and multi-faceted, resisting a clean,
simple fix. In essence, approaches to safety incidents centre on examining
the chain of events that led to an accident or near miss and considering all
acts of those who were in any way involved, and then, crucially, by looking
115
further back at the working conditions of staff and the organisational context
of the incident trying to understand why it occurred.103
A number of authors have emphasised the limitations of voluntary reporting
by healthcare providers as the principal means for detection of safety
incidents.104 For example, among all types of medical errors, cases in which
the wrong patient undergoes an invasive procedure warrant special attention.
However, such major errors tend to be under-reported.105 As a response to
high profile policy reports66;69 that urged the health service to improve patient
safety, and in an attempt to facilitate the process of learning from mistakes,
the NPSA’s National Reporting and Learning System was created. This, for
the first time nationally, encouraged the systematic reporting by healthcare
workers, patients and carers of patient safety incidents, including “near
misses”, which did not result in actual harm.106 It is recognised that a system,
which does not criminalise mistakes107 is needed and the focus on learning
from safety incidents.108 There is also a need to learn how to accurately
submit information to better prioritise, organise and streamline event
analysis.109 Experience from the aviation industry shows that as reporting
rises, the number of serious events begins to decline. Paradoxically, in the
short term at least, an increase in reporting of patient safety incidents has
been a clear sign that the NPSA has been successful in beginning to promote
an open and fair culture in which we can all learn from the mistakes of others–
i.e. the first steps in the process of creating an ”organisation with a
memory”.106
We consider how data from the NRLS have been useful in
helping to think about how to reduce drug allergy-related errors in Chapter 12.
There have furthermore been a number of other important reports
demonstrating clear opportunities from this database, which now contains
well over five million entries.
Yet reporting systems are far from fully effective in monitoring more serious
violations, usually only providing information after a violation has caused
some harm.110 New approaches and methods are needed to study violations
116
in healthcare and IT will play a key role in these; for example by monitoring, in
real-time, variability in quality—this can identify threats to patient safety111 or
through tools, such as event monitoring and natural language processing that
can inexpensively detect adverse events in clinical databases.112 Before we
describe in more detail how eHealth, quality and safety interact we briefly
consider key conceptual considerations in the field of eHealth.
4.4. Defining the field of eHealth
Drivers for increased use of IT in healthcare are multiple,113 but can be
summed up in one overarching message: the increased pervasiveness and
use of IT in all areas of personal, business and public life.
eHealth for healthcare delivery is now one of the main priority areas for most
developed and indeed developing countries and the stakes for eHealth are
high. During the 58th World Health Assembly in 2005, the ministers of health
of the 192 member states of the United Nations (UN), approved the so-called
eHealth Resolution WHA58.28eHealth (a resolution is the highest legal entity
possible in the United Nations system).114 For an extract from the resolution
see Box 4.2. The WHO now also runs a Global Observatory for eHealth
(http://www.who.int/goe/en/), this underscoring the importance now being
attached to technology to help realise the much needed health improvement
agenda internationally.
Such strong support for eHealth, in spite of scepticism of some agencies and
member states, shows global political commitment that has translated into
unprecedented investment in IT in healthcare.
eHealth is a new term that in a very simple way aims to describe, that
something now happens or is assisted by IT, that it happens “electronically”.
For example, like “e” before mail to indicate that mail is now sent
electronically, rather than by post, eHealth refers to the increasing use of IT to
help in the delivery of healthcare.
117
This “e” has however assumed other connotations–for example, notions of
modernity and progress, and representing a drive institute a paradigm shift in
healthcare that translates into increased efficiency, enhanced quality and
safety, and public empowerment and involvement in the pursuit of health.115
The term eHealth has over 50 different definitions.116 It is now widely used in
different domains of society from academic117;118 to policy.113;119 Questions
remain about how the differing concepts and understandings of the term
“eHealth” are used by and impact on different stakeholders.116
Nevertheless, there is a trend to use eHealth in an encompassing way for any
use of IT related to health. A recent analysis of the literature suggests the
concept is best understood as the application of predominantly networked
digital information technologies to support the organisation and delivery of
care.3 It thus sits within the broader field known as health informatics, which
in addition encompasses the fields of hardware, software development and
other technological issues.
For this review, we considered it important to use an inclusive definition of
eHealth (but which nonetheless framed this as a subset of health informatics
as the broader more technological issues were beyond the scope of this
enquiry) and thus chose to use Eysenbach’s definition,
as adapted by
Pagliari, as the basis for our work:115;117
characterises not only a technical development, but also a new way of
improve health care locally, regionally, and worldwide by using information
and communication technology.’
118
The conceptual map shown in Figure 4.5 is derived from an extensive review
of eHealth research and commentary, conducted for the NHS Service
Delivery and Organisation (SDO) Programme by Pagliari et al.3 It represents
a synthesised overview of the important application domains within eHealth,
set within the context of the broader field of medical (or health) informatics
applications. This illustrates the overlap between the two topics (which are
often treated as being equivalent), but also emphasises the pervasive role of
networked information and communication technologies to eHealth, as
distinct from more independent applications of computers and digital devices.
In addition, it reveals the overlap with related subfields such as biomedical
informatics, eBusiness, eLearning and public health informatics, with the
generic areas of computing and telecommunications, and medical equipment.
Development of this map drew on the Medical Informatics Scientific Content
Map developed by the International Medical Informatics Association120 and
structural elements of Medline’s MeSH tree for Medical Informatics.
Box 4.2 An excerpt from the United Nations eHealth Resolution
‘The Fifty-eighth World Health Assembly…Noting the potential impact that advances
in information and communication technologies could have on health-care delivery,
public health, research and health-related activities for the benefit of both low- and
high-income countries; Aware that advances in information and communication
technologies have raised expectations for health;
…
Stressing that eHealth is the cost-effective and secure use of information and
communications technologies in support of health and health-related fields, including
health-care services, health surveillance, health literature, and health education,
knowledge and research,
1. URGES Member States:
(1) to consider drawing up a long-term strategic plan for developing and
implementing eHealth services in the various areas of the health sector, including
health administration…
(2) to develop the infrastructure for information and communication technologies for
health as deemed appropriate to promote equitable, affordable, and universal access
119
to their benefits, and to continue to work with information and telecommunication
agencies and other partners in order to reduce costs and make eHealth successful;
…
(6) to establish national centres and networks of excellence for eHealth best
practice, policy coordination, and technical support for health-care delivery, service
improvement, information to citizens, capacity building, and surveillance;
(7) to consider establishing and implementing national electronic public-health
information systems and to improve, by means of information, the capacity for
surveillance of, and rapid response to, disease and public-health emergencies’
The core applications of eHealth, according to this map, are summarised as:
•
Storing and managing data: this refers to electronic health records
(EHRs) and records systems, including clinical and administrative data
(and stored images), at both the patient and population levels.
•
Informing and supporting decisions: this includes applications to aid
professionals, namely clinical decision support tools and systems, with
on-line access to research evidence, guidelines and professional
development tools.
It also includes consumer health informatics
applications such as on-line patient information, decision aids, targeted
educational interventions or therapy, personal health records and
online peer support.
•
Providing expertise or care at a distance: this includes, firstly,
telemedicine applications, which chiefly concern professional-toprofessional interaction for advice or case conferencing and also
remote access to health records and decision support. It also includes
telecare applications, which represent an increasingly important part of
the eHealth landscape and concern the provision of remote
professional-patient communication and support (telehealthcare), as
well as the use of technology to promote supported self-care in the
home (including self-monitoring, education etc.).
Each of these three areas overlaps with the others, to a greater or lesser
extent,
and
there
is
considerable
interplay between
the
individual
120
applications, reflecting both the complexity of the area and the pervasive
theme of networked data and technologies.
For example, telecare
applications may integrate decision aids and online information, in addition to
telemonitoring and remote clinician advice, while the effectiveness of both
computerised decision support systems (CDSSs) and telecare is highly
dependent on valid electronic health records.
Likewise, computerised
provider (or physician) order entry (CPOE) is conceptually related to the EHR,
since it concerns electronic documentation and data transfer, but it also often
integrates CDSS, which may in turn accommodate both prompts for
guideline-compliant prescribing, error alerting and or linkage to incident
reporting systems. While digital devices, per se, are not within the scope of
eHealth, the application of networked devices linked to an EHR is for
example. This might include radio-frequency identification (RFID) bracelets
linked to patient records and CPOE.
Figure 4.5 has been elongated to fit the page and the upper, bold, arrow is
designed
to
illustrate
the
overlap
between
the
outer
two
circles.
Nevertheless, the display is useful in that it emphasises the central
importance of the EHR in maximising the benefits of eHealth applications for
the effective organisation and delivery of healthcare.
Figure 4.5 also
recognises the value of the EHR and related applications for research,
disease surveillance and health service planning, as well as the overlap with
electronic knowledge repositories, which are critical to the development of
effective decision support.
121
Public
health
information
systems
Mobile access
to records;
evidence &
CDS
Medical
conferencing;
clinical email
Systems biology; genomics
Bioinformatics
Genetic
patient data
[Adapted from
Pagliari, 2005]
Medical
tools,
Equipment
Home & ambulatory
disease monitoring
& self-management
support
Patient-provider
email; internet
consultations
Remote
interventions (e.g.
telepsychiatry,
telesurgery)
Care
Administrative
systems
e-Business
Monitoring/planning
(e.g. bed occupancy, stock,
workflow, costs,). Transacting
(e.g. ordering, billing)
Administrative
(e.g. for audit,
purchasing, billing,
tracking service
utilisation
etc.)
Diagnostic or
treatment advice
from subject
experts (e.g.
telepathology)
Expertise
& Knowledge
Delivering Expertise &
Care at a Distance
Pervasive eHealth Theme: Networked Digital Information and Communication
Education & Library Sciences
eLearning
Knowledge
repositories
e.g. NELH
Integrated records,
supporting multiple
stakeholders
Medical images
Records
Systems
Clinical
(e.g. for capturing,
displaying, sharing,
linking or exchanging patient-specific
data; or populating
decision support)
Storing & Managing
Data
eHealth
Medical Informatics
Computerised medical
equipment e.g. scanners,
Use of digital
chemical assay tools
equipment for patient care
vital signs sensors
e.g. robotics, bionics, imaging,
& monitors,
remote disease monitoring,
drug dispDigital
patient tracking
ensers
devices linked to
(e.g. RFID)
clinical networks,
EHR, decision support
[Described in terms of its applications]
Data
Patient-specific
records, supporting
care of individuals.
Population-based
data, aiding research,
policy & planning.
Administrative data,
aiding organisational
& business processes.
e.g. e-Management of
clinical trials. Use of EHR databases for epidemiological research.
GIS for disease surveillance.
Health promotion via WWW
Supporting
Professionals
Case-specific
diagnostic or treatment
advice based on patient
data & expert
knowledge/evidence
Automated prompts &
reminders for guideline
compliant prescribing,
screening or reporting
+ safety alerts
Electronic guidelines,
research reports
& CME tools
Public Health
Supporting
Patients/Citizens
On-line information
on health/lifestyle or
illness/treatments
Tools for balancing
risks/benefits/preferences to aid choices
Targeted educational
interventions
Personal Health
Records & selfmanagement aids
On-line support
networks
Informing &
Supporting Decisions
Emerging ICT (egeg WWW, wireless,
satellite, DTV, GRID, broadband, GIS, robotics
Healthcare IT
virtual reality, mobile/ wearable/
systems, hardware, software,
implanted devices, pervasive
computing, computational networks, data integration tools, PACS,
information architectures, specific devices
imaging, nanoICT
technology, (eg smartcards, mobile phones
applications for
GPRS PDAs), bar-codes, RFID,
NHS-Net etc.
health providers, patients
& citizens
Computing &
Telecommunications
Figure 4.5 A conceptual map of eHealth
122
4.4.1 Cross-cutting issues for eHealth
In addition to defining and elucidating the applications of eHealth, Pagliari et
al.’s review for NHS SDO identified a number of cross-cutting issues that
reflect common themes of published research and commentary on the topic.
These include:
•
managing change (individual & organisational behaviour, public
perception)
•
integrating appropriate evaluation (tailoring to technology stage &
research questions)
•
optimising human-computer interaction (usability & accessibility)
•
integrating data and systems
(standards & interoperability;
data
quality & clinical coding)
•
ethico-legal issues (e.g. data privacy and governance, digital health
inequalities)
•
quality improvement (clinical benefits, patient enablement, efficiency,
cost, safety) and
•
patient safety (reducing error and risks, preventing harm).
This review addressed all of these themes, with the exception of ethico-legal
issues, for which the focus on systematic reviews of effectiveness is less
suitable.
Particular attention was paid to quality and safety, human and
organisational issues affecting the technology adoption, socio-technical
factors affecting the usability of technology, challenges for evaluation and
maximising the communicative and integrative power of technologies and
data.
4.4.2 NHS Connecting for Health and eHealth
Our tracing of NHS Connecting for Health (NHS CFH; see Chapter 3) reveals
that it is a highly complex group of sub-programmes, only a fraction of which
are represented within the headline deliverables typically cited.
It has
common strands with major quasi-independent initiatives, such as the
National Knowledge Service and NHS Direct Online, while there is
considerable internal complexity in the form of multiple design and
123
specification programmes around IT architectures, standards and individual
applications; supplier procurement and control mechanisms, and education
and user engagement. Chapter 3 maps the key deliverables of NHS CFH
and areas of synergy and overlap with other aspects of the National
Programme for Information Technology (NPfIT). These elements are now
distributed throughout all three main eHealth applications domains, although
the primary focus has been on delivering electronic applications that support
NHS staff and organisations. A core component of NHS CFH is the NHS
Care Records Service (NHS CRS), the primary objective of which was to
deliver an integrated EHR, which will in due course support other applications
such as appointment booking (Choose and Book) and the electronic transfer
of prescriptions (ETP).
The EHR is also central to the effective use of
CDSSs, which NHS CFH and the National Knowledge Service have been
collaborating to develop. The National Knowledge Service and NHS Direct
are engaged in knowledge delivery to professionals and patients, NHS Direct
is using CDSS algorithms for triage and to support patient self-care support.
The transfer of clinical email and images supports the exchange of expertise
and clinical care over geographic boundaries. The importance of internetbased data transfer and communications, as facilitated by New National
Network (N3), also identifies these programmes with common definitions of
eHealth.
While the primary focus of NHS CFH has been professional and
organisational interventions, a number of elements are consumer-oriented
(most obviously HealthSpace). In addition, while telehealthcare was originally
outside the remit of the programme, areas such as remote disease monitoring
and patient access to GP records are also being considered by the
Programme. Patient access to GP records has already been made available
independently of NHS CFH by one system supplier (EMIS) through the EMIS
Access system (https://www.emisaccess.co.uk/).
Figure 4.6 shows the application domains at the centre of the larger eHealth
map and is intended to illustrate the areas addressed by the commissioning
brief and this report.
124
Figure 4.6 Areas of eHealth prioritised in the research commissioning brief
Informing &
Supporting Decisions
Supporting
Patients/Citizens
On-line information
on health/lifestyle or
illness/treatments
Tools for balancing
risks/benefits/preferences to aid choices
Targeted educational
interventions
Personal Health
Records & selfmanagement aids
On-line support
networks
Supporting
Professionals
Case-specific
diagnostic or treatment
advice based on patient
data & expert
knowledge/evidence
Automated prompts &
reminders for guideline
compliant prescribing,
screening or reporting
+ safety alerts
Electronic guidelines,
research reports
& CME tools
Storing & Managing
Data
Data
Patient-specific
records, supporting
care of individuals.
Population-based
data, aiding research,
policy & planning.
Administrative data,
aiding organisational
& business processes.
Integrated records,
supporting multiple
stakeholders
Medical images
Records
Systems
Clinical
(e.g. for capturing,
displaying, sharing,
linking or exchanging patient-specific
data; or populating
decision support)
Administrative
(e.g. for audit,
purchasing, billing,
tracking service
utilisation
etc.)
Delivering Expertise &
Care at a Distance
Expertise
& Knowledge
Diagnostic or
treatment advice
from subject
experts (e.g.
telepathology)
Medical conferencing;
clinical email
Mobile access
to records;
evidence;
CDS
Care
Patient-provider
email; internet
consultations
Remote
interventions (e.g.
telepsychiatry,
telesurgery)
Home & ambulatory
disease monitoring &
self-management
support
125
4.5 Integrating the concepts of quality, safety and eHealth
As described previously, the concepts of quality and safety are closely
related, and arguments for the power of IT to improve healthcare quality and
safety have been well made (although research evidence has yet to catch up
with the vision). There is also potential for technology to introduce new risks,
which can be rooted in different aspects of the hardware, software or
networks (e.g. reliability and usability), the users (including cognitive and
psychological factors) or the teams and organisations in which they are used
(e.g. safety culture).121
The links between technology, quality and safety are also reflected in the
conclusions of the recent SDO eHealth review,3 which summarised the
headline benefits as in Box 4.3.
Box 4.3 Headline benefits of eHealth, as discussed by SDO
Improving healthcare access by
•
helping to alleviate barriers to effective healthcare introduced by physical
location or disability
•
facilitating consumer empowerment for self-care and health decision-making.
Improving healthcare quality by:
•
aiding evidence-based practice
•
tailoring care to individuals, where IT enables more informed decision-making
based on evidence and patient-specific data
•
improving transparency and accountability of care processes & facilitating
integrated and shared care
•
reducing errors and increasing safety.
Improving healthcare cost-efficiency by
•
streamlining healthcare processes, reducing waiting times and waste
•
improving diagnostic accuracy & treatment appropriateness.
3
Source: Pagliari et al. (2005)
126
eHealth applications have the potential to improve the safety of patient care in
many ways; for example, by ensuring that records are legible, less likely to be
lost and more likely to reflect an accurate patient identity; by improving
diagnostic accuracy, supporting evidence-based prescribing and flagging
dangers through decision support; by aiding organisational learning through
incident reporting systems, or by improving the capacity of patients to mitigate
risks through identifying errors and maximising self-care.122
However,
introducing new technology also comes with risks attributable to the hardware
or software itself or its misuse by healthcare personnel.123;124 Hence, it is
appropriate for this project to define the possible risks and safety benefits of
eHealth applications and this requires an understanding of the frameworks by
which patient safety has been traditionally understood and classified. Our
efforts have been aided by multiple sources, including the comprehensive
2001 review of the safety and quality literature by the US Agency for
Healthcare Research and Quality (AHRQ).91
Key elements of the safety and quality frameworks, discussed in this chapter,
are triangulated with the core application domains from the eHealth map in
Figure 4.7, illustrating the relevance of safety considerations to all aspects of
technology design, development and deployment.
127
Figure
4.7
Integrated
maps
of
quality,
safety
and
128
eHealth
4.5.1
Integrated
Joint
Commission
on
Accreditation
of
Healthcare
Organizations safety classification scheme and eHealth applications action
Below we further integrate earlier introduced (See section 4.4) conceptual
maps and classification schemes of safety from JCAHO2 and WHO1;102 with a
more detailed description of integration with those of eHealth. Figure 4.8
presents the core concepts within the JCAHO taxonomy, which have been
annotated to illustrate issues for eHealth.
(More elaborate organisational
charts exist, but this level is most easily triangulated with issues for eHealth).
129
Management
IT skills. Process
redesign. NKS
Protocols/Processes
e-protocols (e.g.
referral), coding &
QoF, IT standards
e.g. training &
supervision around safety
Knowledge Transfer
eg documentation,
incentives, standards
e.g. around information
sharing or event reporting
Culture
e.g. staffing/training,
safety financing
Facilities
IT usability,
accessibility,
dependability,
+ QI loops
e.g. design, malfunction
obsolescence
SYSTEMS
EHR/ e-forms
Remote advice for
professional;
remote patient
consultation
EHR/Coding
Patient websites
Promote change &
innovation
IT planning &
procurement
Organisational
Documentation
Consent Process
Advice or Interpretation
Information quality
Communication
Negligence
Recklessness
Intentional rule violations
CAUSE
Lack of tools/procedures for
tracking, auditing, alerting
Technical
IT skills, ability to
recognise good &
poor quality info.
Use of unsecure ecomms for sensitive
info.
Knowledge-based
(interpretation)
Rule-based
(retrieval)
Skills based
(execution)
Interpretation of
data (images,
reports, records)
information,
evidence, decision
support
130
Data
extraction
Data entry,
software use
External factors
Clinical Decision Support; CPOE,
e-prescribing, operating room
systems, bar-coding & RFID
tagging (e.g. blood, patients),
EHR, eLearning, audit
Practitioner
HUMAN
(Error)
Post-intervention
Inaccurate prognosis
Intervention
Incorrect procedure enacted
Correct procedure omitted
Clinical Performance
(Practitioner level)
Pre-Intervention
Questionable diagnosis
Patient factors
Patient
e-comms.
e.g. lab,
booking,
referral.
EHR
eg e-referral, EHR
systems
Referral or consultation
Tracking or follow-up
Patient Management
(Organisation level)
TYPE OF PROBLEM
Figure 4.8: Core features of the JCAHO Patient Safety Taxonomy—
annotated to reflect sample eHealth solutions and issues
CDS (alerts, alarms,
reminders) linked to EHR,
supporting by high quality
coding.
Integrated EHR; better end
comms (e.g. lab, referral, 2
opinion)
Bar Coding, EHR, NHS
Number, Record Linkage,
RFID
Due to faulty infor-mation,
interventions or
confidentiality breaches
Figure 4.8 continues…
Morbidity and
mortality due to
poor CDSS, ecomms or EHR
integrity
Confidentiality &
negligence suits
Poor adoption of
technology
Cost to modify or
replace IT & re-train
staff
Legal settlements.
Lost bed days &
waiting times
131
4.5.2 Detailed map of the impact on quality and safety of healthcare by
ePrescribing
According to Barber et al., errors
related to medicines’ management are
probably the most common type of medical error in both primary and
secondary care in the UK.125 Of all types of medicines’ management errors—
prescribing, dispensing, administration, monitoring, repeat prescribing126—
prescribing errors are typically the most serious.125 Thus from a combined
conceptual map of quality, safety and eHealth here we present in detail how a
specific application, namely ePrescribing can proactively and reactively
improve quality and safety of prescribing.
The International Classification for Patient Safety (ICPS) presents a
comprehensive conceptual framework developed to be used in conjunction
with other processes, systems and maps (see Figure 4.9). Based on the
patient safety and eHealth conceptual frameworks discussed above, we
present a list of problems related to medication incidents, contributing factors
(staff and environment) and a detection process that may mitigate the
process of prescribing, prevent or reduce the likelihood of problems
happening, and assist in the recognition of these errors (see Table 4.1).
132
Figure 4.9 should be read together with Table 4.1. Legend: The solid lines
enclose the 10 major classes of the ICPS and represent the semantic relationships
between them. The dotted lines represent the flow of information. Incident type is
medication incident and action taken to reduce risk is use of ePrescribing.
National Patient Safety Agency International classification of Patient Safety – permission applied for.
133
Table 4.1 This should be read together with Figure 4.9. It presents problems leading
to medication safety incidents, contributing prescriber, communication and
environment factors, detection processes and finally theoretical modes of action of
ePrescribing application that influence all the former.
134
Medication
safety
incidents*
Wrong Patient
Wrong Drug
Wrong
Dose/Strength
of Frequency
Wrong
Formulation or
Presentation
Wrong Route
Wrong Quantity
Contraindication
Omitted
Medicine
or
Dose
Adverse Drug
Reaction
Contributing prescriber factors
Cognitive factors
Perception/understanding
Knowledge-based/problem
solving
Rule-based
Slip/lapse
error/absentmindedness/
forgetfulness
Technical error in execution
(physical)
Failure to synthesise/act on
available information
Performance factors
Distraction/inattention
Fatigue/exhaustion
Behavior/violation
Non-compliance
Routine violation
Risky behavior
Reckless behavior
Problem
with
substance
abuse/Use
Sabotage/criminal act
Communication factors
Paper-based
Verbal
Contributing
environment
factors
Remote/long
distance
from
service
Detection process
Error recognition
By change in patient’s status
By system/environment
Change/alarm
By a count/audit/review
Proactive risk assessment
Theoretical modes of action of ePrescribing as
a pro-active and reactive application to
improve quality and safety of prescribing
1. Improved identification
2. Improved legibility
3. Automatic checks on previous response to the
drug
4. Visual or audible alerts and warnings related
to drug–drug interactions
5. Automatic
checks
for
individual’s
characteristics
(e.g.
drug-allergies,
comorbidities, weight, gender or ethnicity) that
may influence the choice of drug, dose,
strength of frequency, formulation or
presentation, route or quantity
6. Automatic checks for results of clinical
investigations (e.g. laboratory values of kidney
function tests or blood pressure) that may
influence the choice of drug, dose, strength of
frequency, formulation or presentation, route
or quantity
7. Automatic checks on previous response to the
drug
8. Automatic checks on duplicate therapies
9. Automatic check on the amount of medication
prescribed at last visit notifying about under
and overuse of medication and reducing
likelihood of under and overprescribing
10. Automatic barring of too high (dangerous)
doses according to patient’s characteristics
(single, daily or life dose limits)
11. Advice regarding evidence-based alternative
first-choice drug(s) from the same or similar
class, drug dose, strength of frequency,
formulation or presentation, route or quantity
12. Linking to algorithms emphasising (offering as
a first choice when a drug is selected) costeffective drug, dose, strength of frequency,
13. Advice regarding the cost of medication and
cheaper equivalent drug(s) from the same or
similar class, dose, strength of frequency,
14. Reminders about corollary orders
15. Access to information about previously
prescribed medication (anytime & anywhere)
16. Reminders on drug guidelines
17. Faster response to clinicians’ prescriptions
18. Reduced patient data re-keying
19. Link to formulary—full information and
knowledge-base
about the
medication
prescribed
20. Instant provision of information about
formulary-based drug coverage including onformulary alternatives and co-pay information
21. Improved communication amongst prescribers
and dispensers (e.g. call back queries, instant
reporting that item is out of stock, alerts for
unfilled, unrenewed prescriptions)
22. Automatic monitoring of orders and audit
23. Shorter process turn-around time—transit
time to dispensing site, time till first dose,
prescription renewal or refill
24. Reduction in use of paper
25. Data are available for immediate analysis
including post-marketing reporting, drug
utilisation review, etc.
135
26. Possibility of remote or long-distance
prescribing
27. Electronic access to checklists, protocols and
or policies
136
4.5.3 An example of a scenario to illustrate the potential of eHealth to improve
the quality and safety of laboratory results reporting
As a further example, a detailed analysis of potential theoretical impact of
eHealth on patient safety is presented in Table 4.2. This provides a scenario
of eHealth’s impact on the quality and safety of clinical processes using the
example of laboratory results reporting.
137
Table 4.2 Potential hazards in laboratory result reporting
Action
Potential hazard
Doctor
orders • Unnecessary testing (e.g. previous
test (either by
recent tests not obvious on
computerised
computer screen). Recent hospital
provider order
out-patient investigation results not
entry (CPOE) or
available to GP, no available online
paper form, or
protocol to advise on appropriate
verbally
tells
testing)
patient to get x • May not enter correct patients
checked)
details, e.g. wrong patient called
up on computer scheme ( “similar
name”)
• Electronic ordering, incorrect test
ordered or test left out, patient not
informed to go for test, no follow up
of uncompleted tests.
• Patient misunderstands instruction
• Appropriate test does not get
done
Nurse
calls • Wrong patient presents (e.g.,
patient for test
mishears name, nurse does not
check name)
• “Similar name” problem as above
• Misidentified specimen sent to
lab
Form completed • Wrong labels picked up and applied
sample bottles
by accident
labelled
• Wrong labels printed because of
similar names
• Misidentified specimen sent to
lab
Blood drawn
• Wrong sample bottle used
• Sample arrives late at lab and is old
resulting in inaccurate analysis
• Unsuitable specimen
Samples
and
forms arrive at
laboratory
•
•
•
•
Possible solution
Shared
patient
records
(primary and secondary
care).
Regular
integration
of
results into GP system
Clinician education to use
patient date of birth for
calling up records.
Test
request
routinely
recorded on computer and
regularly audited against
completion
Clinician education, patient
asked to check details on
specimen label
Clinician education, patient
asked to check details on
specimen label
Easily
available
online
advice.
Computer alarm warning
nurse that time for last
blood test is passed
request
routinely
Lost samples before reaching lab, Test
recorded
on
computer
and
clinician unaware that sample has
regularly audited against
not arrived.
completion
Lost specimen
Error at data entry level if form not Bar-coding of all specimens
bar-coded
Misidentified sample analysed
138
Table 4.2 Potential hazards in laboratory result reporting continued…
Action
Analysis
performed.
Result verified
Posting
of
laboratory
results—printed,
emailed or webbased.
Results pulled into
practice computer
system, scanned
and
entered
manually,
or
entered
automatically from
web by software
Results checked
by clinician, on
paper,
scanned
document
or
electronic entry
Potential hazard
• Lab-based
calibration, poor
reagents, etc.
Possible solution
errors—e.g.
storage of
•
•
•
•
Test request routinely
recorded on computer
and regularly audited
against completion.
Clinician education in
mail box hygiene
Patient asked to check
for result
• Results entered into wrong Administration education.
Use of date of birth/NHS
patient’s records
• Result
unavailable
for number for data entry.
intended patient
• Patient mistreated because of recorded on computer
and regularly audited
erroneous result
against completion.
Printed lab result goes missing
email not noticed
Mailbox full
Wrong clinician/surgery gets
results (delay)
• Confidentiality risk
• Clinician does not read mail (e.g.
on
holiday
no
alternative
provision made)
• Abnormal result not noticed, due
to lack of knowledge, poor
highlighting of abnormal results,
fatigue, information overload,
multiple presentation of different
lab results and set overlooked
• Abnormal result missed
Clinician/admin
education.
Regular
check on mail boxes
(especially
holiday/illness) to ensure
completion.
Double checking of all
“normal” results
Presentation of tests
individually on screen
rather than layered
139
Table 4.2 Potential hazards in laboratory result reporting continued…
Action
Potential hazard
Possible solution
Patient informed
• Patient does not call for result
• Patient is told some results recorded on computer
normal, not aware that further and regularly audited
completion.
potentially abnormal results are against
Patient
informed
is part of
to follow
trail
to
be
• Patient is told wrong result audit
(misidentification, wrong record completed
Clinician education on
called up)
of
date
of
• Confidentiality risk, by phone , use
birth/postcode/password
post or in person
• Patient does not get message before giving information
Follow
up
of
abnormal result
•
•
•
•
Integration
with
computer alerting
systems
•
•
(phone/email/letter
about
abnormal result)
Patient does not understand
significance of result and does
not follow-up as intended
Patient does not attend followup of abnormal result, system
not in place to detect this
Patient attends follow-up. One
abnormal result noted by
clinician but does not notice
second abnormality
Important result not followed
up
Result scanned or added as
free text, not in searchable form
detectable by computer alerting
system
Inappropriate drug prescribed
despite
biochemical,
abnormality.
Patient action required
(e.g. send for) logged in
computer. Regular audit
to
check
attendance
when sent for.
All
abnormal
results
highlighted in journal text
Clinician/administration
education
140
4.6 Conclusions
There are clear parallels in the concepts of quality, safety and eHealth. Each of
these areas has suffered from variability in definitions, epistemologies and
terminologies, which represent obstacles to the synthesis of existing research.
Effectiveness is a core construct of the quality concept, which also accommodates
efficiency (and cost), access, equity and accessibility.
Prevention of harm and
identification of error and risks lie at the centre of the safety concept, and human and
organisational factors (including safety culture) are seen as important contributory
factors.
Networked data and communications are a pervasive theme within the
concept of eHealth and EHRs a core application area.
Tools for informing and
supporting decisions, and aiding clinical practice and patient care at a distance are
also central eHealth domains. The multiple objectives of NHS CFH overlap with
those of several other NHS technology programmes; for example, NHS Direct. The
primary focus of NHS CFH is the delivery of networks and tools for supporting
healthcare professionals and organisations; however, elements of the Programme
are consumer-oriented and there have been recent moves towards integrating
remote healthcare within the scope of the programme. Nevertheless, the scope of
this project’s commissioning brief is firmly located within the former domain.
Evidence of effectiveness is crucial for motivating the adoption of new eHealth
applications by clinicians and a primary objective of the review is to synthesise
available evidence on the outcomes of eHealth interventions and identify best
practice implementation and adoption strategies.
References
1. World Health Organization. International Classification for Patient Safety. Version 1.0
for Use in Field Testing 2007 - 2008. Geneva: WHO.
2. Chang A, Schyve PM, Croteau RJ, O'Leary DS, Loeb JM. The JCAHO patient safety
event taxonomy: a standardized terminology and classification schema for near
misses and adverse events. Int J Qual Health Care 2005; 17: 95-105.
3. Pagliari C. Literature review and conceptual map of the field of eHealth. Final report
to the SDO programme, 2004.
4. Editor. Streptomycin in pulmonary tuberculosis (Reprinted from BMJ 1948:ii:790-1).
BMJ 1998; 317: 1248.
5. Editor. Fifty years of randomised controlled trials. BMJ 1998; 317.
6. Grol R. Improving the quality of medical care: Building bridges among professional
pride, payer profit, and patient satisfaction. JAMA 2001; 286: 2578-85.
141
7. Institute of Medicine. To Err Is Human: Building a Safer Health System. Washington,
DC, National Academy Press, 2000.
8. Institute of Medicine. Crossing the Quality Chasm: A New Health System for the 21st
Century. Washington, D.C., National Academy Press, 2001.
9. Institute of Medicine. Patient Safety: Achieving a New Standard for Care.
Washington, D.C., National Academy Press, 2003.
10. Institute of Medicine. Preventing Medication Errors. Washington, D.C., National
Academy Press, 2007.
11. McGlynn EA, Asch SM, Adams J, Keesey J, Hicks J, DeCristofaro A et al. The
Quality of Health Care Delivered to Adults in the United States. N Engl J Med 2003;
348: 2635-45.
12. Tudor Hart J. The inverse care law. Lancet 1971;297:405-412.
13. Watt G. The inverse care law today. Lancet 2002;360:252-254.
14. Spiegelhalter DJ. Surgical Audit: Statistical lessons from Nightingale and Codman. J
R Stat Soc Ser A Stat Soc 1999;162: 45-58.
15. Donabedian A. Evaluating the quality of medical care. Milbank Mem Fund Q 1966;
44: 166–203.
16. Best M, Neuhauser D. Avedis Donabedian: father of quality assurance and poet.
Qual Saf Health Care 2004; 13: 472-73.
17. Donabedian A. The quality of medical care. Science 1978; 200: 856-64.
18. Blumenthal D. Quality of care -- What is it?- Part one of six. N Engl J Med 1996; 335:
891-94.
19. Brennan TA, Gawande A, Thomas E, Studdert D. Accidental deaths, saved lives, and
improved quality. N Engl J Med 2005; 353: 1405-09.
20. Maxwell RJ. Quality assessment in health. BMJ 1984; 288: 1470-72.
21. Howie JGR, Heaney D, Maxwell M. Quality, core values and the general practice
consultation: issues of definition, measurement and delivery. Fam Pract 2004; 21:
458-68.
22. Berwick DM. Quality comes home. Ann Intern Med 1996; 125: 839-43.
23. Schuster MA, McGlynn EA, Brook RH. How good is the quality of health care in the
United States? Milbank Q 1998; 76: 517-63.
24. Engels Y, Campbell S, Dautzenberg M, van den Hombergh P, Brinkmann H,
Szecsenyi J et al. Developing a framework of, and quality indicators for, general
practice management in Europe. Fam Pract 2005; 22: 215-22.
25. Campbell SM, Roland MO, Buetow SA. Defining quality of care. Soc Sci Med 2000;
51: 1611-25.
26. Berwick DM. Errors today and errors tomorrow. N Engl J Med 2003; 348: 2570-72.
27. Starfield B. New paradigms for quality in primary care. Br J Gen Pract 2001; 51: 30309.
28. Grol R, Baker R, Moss F. Quality improvement research: understanding the science
of change in health care. Qual Saf Health Care 2002; 11: 110-11.
29. Brook RH, McGlynn EA, Shekelle PG. Defining and measuring quality of care: a
perspective from US researchers. Int J Qual Health Care 2000; 12: 281-95.
30. Reeves D, Campbell SM, Adams J, Shekelle PG, Kontopantelis E, Roland MO.
Combining multiple indicators of clinical quality: an evaluation of different analytic
approaches. Med Care 2007; 45: 489-96.
31. Pronovost PJ, Berenholtz SM, Needham DM. A framework for health care
organizations to develop and evaluate a safety scorecard. JAMA 2007; 298: 2063-65.
32. Lilford R, Mohammed MA, Spiegelhalter D, Thomson R. Use and misuse of process
and outcome data in managing performance of acute medical care: avoiding
institutional stigma. Lancet 2004; 363:1147-54.
33. Kassirer JP. The quality of care and the quality of measuring it. N Engl J Med 1993;
329: 1263-65.
34. Lilford R, Girling A, Stevens A, Almasri A, Mohammed MA, Braunholtz D. Adjusting
for treatment refusal in rationing decisions. BMJ 2006; 332: 542-44.
142
35. Kerr EA, Hofer TP, Hayward RA, Adams JL, Hogan MM, McGlynn EA et al. Quality
by any other name?: A comparison of three profiling systems for assessing health
care quality. Health Serv Res 2007; 42: 2070-87.
36. Pugh MJV, Berlowitz DR, Montouris G, Bokhour B, Cramer JA, Bohm V et al. What
constitutes high quality of care for adults with epilepsy? Neurology 2007; 69: 202027.
37. Higashi T, Wenger NS, Adams JL, Fung C, Roland M, McGlynn EA et al.
Relationship between Number of Medical Conditions and Quality of Care. N Engl J
Med 2007; 356: 2496-04.
38. Bottle A, Gnani S, Saxena S, Aylin P, Mainous AG, III, Majeed A. Association
Between quality of primary care and hospitalization for coronary heart disease in
England: national cross-sectional study. J Gen Intern Med 2008; 23: 135-41.
39. Calderón-Larrañaga A, Carney L, Soljak M, Bottle A, Partridge M, Bell D, et al.
Association of population and primary healthcare factors with hospital admission
rates for chronic obstructive pulmonary disease in England: national cross-sectional
study. Thorax 2010 ePub ahead of print
40. Lilford R, Pronovost P. Using hospital mortality rates to judge hospital performance: a
bad idea that just won't go away. BMJ 2010; 340: c2016.
41. Halm EA, Lee C, Chassin MR. Is volume related to outcome in health care? A
systematic review and methodologic critique of the literature. Ann Intern Med 2002;
137: 511-20.
42. Pitches D, Mohammed M, Lilford R. What is the empirical evidence that hospitals
with higher-risk adjusted mortality rates provide poorer quality care? A systematic
review of the literature. BMC Health Serv Res 2007; 7: 91.
43. Millett C, Car J, Eldred D, Khunti K, Mainous AG, III, Majeed A. Diabetes prevalence,
process of care and outcomes in relation to practice size, caseload and deprivation:
national cross-sectional study in primary care. J R Soc Med 2007; 100: 275-83.
44. Saxena S, Car J, Eldred D, Soljak M, Majeed A. Practice size, caseload, deprivation
and quality of care of patients with coronary heart disease, hypertension and stroke
in primary care: national cross-sectional study. BMC Health Serv Res 2007; 7: 96.
45. Heath I, Hippisley-Cox J, Smeeth L. Measuring performance and missing the point?
BMJ 2007; 335: 1075-76.
46. McDonald R, Roland M. Pay for performance in primary care in England and
California: comparison of unintended consequences. Ann Fam Med 2009; 7: 121-27.
47. Alshamsan R, Majeed A, Ashworth M, Car J, Millett C. Impact of pay for performance
on inequalities in health care: systematic review. J Health Serv Res Policy 2010; 15:
178-84.
48. Pear R. A.M.A. to Develop Measure of Quality of Medical Care. New York Times
2006 Feb 21.
49. Hodgson DC, Brierley JD, Cernat G, Bondy S, Slaughter PM, Pinfold SP et al. The
consistency of panelists' appropriateness ratings: do experts produce clinically logical
scores for rectal cancer treatment? Health Policy 2005; 71: 57-65.
50. Naylor CD. What is Appropriate Care? N Engl J Med 1998; 338: 1918-20.
51. Anderson GM, Brown AD. Appropriateness ratings: overuse, underuse, or misuse?
Lancet 2001; 358: 1475-76.
52. Marshall MN, Shekelle PG, McGlynn EA, Campbell S, Brook RH, Roland MO. Can
health care quality indicators be transferred between countries? Qual Saf Health
Care 2003; 12: 8-12.
53. Sheldon TA, Cullum N, Dawson D, Lankshear A, Lowson K, Watt I et al. What's the
evidence that NICE guidance has been implemented? Results from a national
evaluation using time series analysis, audit of patients' notes, and interviews. BMJ
2004; 329: 999.
54. Soumerai SB, McLaughlin TJ, Gurwitz JH, Guadagnoli E, Hauptman PJ, Borbas C et
al. Effect of local medical opinion leaders on quality of care for acute myocardial
infarction: a randomized controlled trial. JAMA 1998; 279: 1358-63.
143
55. Ferguson TB, Jr., Peterson ED, Coombs LP, Eiken MC, Carey ML, Grover FL et al.
Use of continuous quality improvement to increase use of process measures in
patients undergoing coronary artery bypass graft surgery: a randomized controlled
trial. JAMA 2003; 290: 49-56.
56. Peck C, McCall M, McLaren B, Rotem T. Continuing medical education and
continuing professional development: international comparisons. BMJ 2000; 320:
432-35.
57. Colton D. Quality Improvement in health care: conceptual and historical foundations.
Eval Health Prof 2000; 23: 7-42.
58. Chassin MR. Assessing strategies for quality improvement. Health Aff (Millwood)
1997; 16: 151-61.
59. Jamtvedt G, Young JM, Kristoffersen DT, O'Brien MA, Oxman AD. Does telling
people what they have been doing change what they do? A systematic review of the
effects of audit and feedback. Qual Saf Health Care 2006; 15: 433-36.
60. Marshall MN, Shekelle PG, Leatherman S, Brook RH. The public release of
performance data: what do we expect to gain? a review of the evidence. JAMA 2000;
283: 1866-74.
61. Werner RM, Asch DA. The unintended consequences of publicly reporting quality
information. JAMA 2005; 293: 1239-44.
62. Campbell SM, Reeves D, Kontopantelis E, Sibbald B, Roland M. Effects of pay for
performance on the quality of primary care in England. N Engl J Med 2009; 361: 36878.
63. Berwick DM, Calkins DR, McCannon CJ, Hackbarth AD. The 100 000 Lives
Campaign: Setting a goal and a deadline for improving health care quality. JAMA
2006; 295: 324-27.
64. McCannon CJ, Hackbarth AD, Griffin FA. Miles to go: an introduction to the 5 Million
Lives Campaign. Jt Comm J Qual Patient Saf 2007; 33: 477-84.
65. Brown CA, Belfield CR, Field SJ. Cost effectiveness of continuing professional
development in health care: a critical review of the evidence. BMJ 2002; 324: 652-55.
66. Department of Health. An Organisation with a Memory: Report of an expert group of
learning from adverse events in the NHS. London: HMSO, 2000.
67. Aylin P, Tanna S, Bottle A, Jarman B. How often are adverse events reported in
English hospital statistics? BMJ 2004; 329: 369.
68. Blendon RJ, DesRoches CM, Brodie M, Benson JM, Rosen AB, Schneider E et al.
Views of practicing physicians and the public on medical errors. N Engl J Med 2002;
347: 1933-40.
69. Department of Health, Chief Pharmaceutical Officer. Building a Safer NHS for
Patients: improving medication safety. London: HMSO, 2004.
70. Shojania KG, Duncan BW, McDonald KM, Wachter RM, Markowitz AJ. Making
Health Care Safer: a critical analysis of patient safety practices. AHRQ Publication
No. 01-E058. 2001. Rockville, MD, Agency for Healthcare Research and Quality
(AHRQ).
71. Leape LL. Reporting of medical errors: time for a reality check. Qual Saf Health Care
2000; 9: 144-45.
72. Elder NC. Classification of medical errors and preventable adverse events in primary
care: a synthesis of the literature. J Fam Pract 2002;51:927-932.
73. Dovey SM, Meyers DS, Phillips RL, Jr., Green LA, Fryer GE, Galliher JM et al. A
preliminary taxonomy of medical errors in family practice. Qual Saf Health Care
2002; 11: 233-38.
74. Chang A, Schyve PM, Croteau RJ, O'Leary DS, Loeb JM. The JCAHO patient safety
event taxonomy: a standardized terminology and classification schema for near
misses and adverse events. Int J Qual Health Care 2005; 17: 95-105.
75. Runciman WB, Baker GR, Michel P, Dovey S, Lilford RJ, Jensen N, et al. Tracing the
foundations of a conceptual framework for a patient safety ontology. Qual Saf Health
Care 2010; 19: e56.
144
76. Runciman WB, Helps SC, Sexton EJ, Malpass A. A classification for incidents and
accidents in the health-care system. J Qual Clin Pract 1998; 18: 199-211.
77. Hurwitz B, Sheikh A (eds.). Health Care Errors and Patient Safety. London: WileyBMJ, 2009.
78. Vincent C, Taylor-Adams S, Chapman EJ, Hewett D, Prior S, Strange P et al. How to
investigate and analyse clinical incidents: Clinical Risk Unit and Association of
Litigation and Risk Management protocol. BMJ 2000; 320: 777-81.
79. Leape LL. Error in medicine. JAMA 1994; 272:1851-57.
80. Vincent C, Taylor-Adams S, Stanhope N. Framework for analysing risk and safety in
clinical medicine. BMJ 1998;316:1154-1157.
81. Nebeker JR, Barach P, Samore MH. Clarifying Adverse Drug Events: A clinician's
guide to terminology, documentation, and reporting. Ann Intern Med 2004; 140: 795801.
82. Tamuz M, Thomas EJ, Franchois KE. Defining and classifying medical error: lessons
for patient safety reporting systems. Qual Saf Health Care 2004; 13: 13-20.
83. Elder N, Pallerla H, Regan S. What do family physicians consider an error? A
comparison of definitions and physician perception. BMC Fam Pract 2006; 7: 73.
84. Perneger T. The Swiss cheese model of safety incidents: are there holes in the
metaphor? BMC Health Serv Res 2005; 5: 71.
85. Sutcliffe KM. Defining and classifying medical error: lessons for learning. Qual Saf
Health Care 2004; 13: 8-9.
86. McLoughlin V, Millar J, Mattke S, Franca M, Jonsson PM, Somekh D et al. Selecting
indicators for patient safety at the health system level in OECD countries. Int J Qual
Health Care 2006; 18: 14-20.
87. Veazie PJ. An individual-based framework for the study of medical error. Int J Qual
Health Care 2006; 18: 314-19.
88. Leape LL, Berwick DM, Bates DW. What practices will most improve safety?:
evidence-based medicine meets patient safety. JAMA 2002; 288: 501-07.
89. Shojania KG, Duncan BW, McDonald KM, Wachter RM. Safe but sound: patient
safety meets evidence-based medicine. JAMA 2002; 288: 508-13.
90. Forster AJ, Shojania KG, van Walraven C. Improving patient safety: moving beyond
the "hype" of medical errors. Can Med Assoc J 2005; 173: 893-94.
91. Shojania KG, Duncan BW, McDonald KM, Wachter RM, Markowitz AJ. Making health
care safer: a critical analysis of patient safety practices. Evid Rep Technol Assess
(Summ ) 2001; i-668.
92. Battles JB, Lilford RJ. Organizing patient safety research to identify risks and
hazards. Qual Saf Health Care 2003; 12: ii2-ii7.
93. Boxwala AA, Dierks M, Keenan M, Jackson S, Hanscom R, Bates DW et al.
Organization and representation of patient safety data: current status and issues
around generalizability and scalability. J Am Med Inform Assoc 2004; 11: 468-78.
94. Wiegmann DA. Human error analysis of commercial aviation accidents: application of
the Human Factors Analysis and Classification system (HFACS). Aviat Space
Environ Med 2001; 72: 1006-16.
95. Helmreich RL. On error management: lessons from aviation. BMJ 2000; 320: 781-85.
96. Wilf-Miron R, Lewenhoff I, Benyamini Z, Aviram A. From aviation to medicine:
applying concepts of aviation safety to risk management in ambulatory care. Qual
Saf Health Care 2003; 12: 35-39.
97. Reason J. Human Error. Cambridge: Cambridge University Press, 1990.
98. Reason J. Human error: models and management. BMJ 2000; 320: 768-70.
99. Panesar SS, Cleary K, Sheikh A. Reflections on the National Patient Safety Agency's
database of medical errors. J R Soc Med 2009; 102: 256-58.
100. Vincent C, Taylor-Adams S, Stanhope N. Framework for analysing risk and safety in
clinical medicine. BMJ 1998; 316: 1154-57.
101. World Health Organization. Towards a Common International Understanding of
Patient Safety Concepts and Terms: Taxonomy and Terminology Related to Medical
145
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
Errors and System Failures. Geneva: World Health Organization, Department of
Health Service Provision, 2003.
World Health Organization. More than words - International Classification for Patient
Safety. Geneva: WHO, 2007.
Pringle M. Significant event auditing and root cause analysis. In: Hurwitz B, Sheikh A
(eds.). Health Care Errors and Patient Safety. London: Wiley-BMJ, 2009: 193-206.
Vincent C. Incident reporting and patient safety. BMJ 2007; 334: 51.
Chassin MR, Becher EC. The Wrong Patient. Ann Intern Med 2002; 136: 826-33.
Stephenson T. The National Patient Safety Agency. Arch Dis Child 2005; 90: 226-28.
Holbrook J. The criminalisation of fatal medical mistakes. BMJ 2003; 327: 1118-19.
Pierson S, Hansen R, Greene S, Williams C, Akers R, Jonsson M et al. Preventing
medication errors in long-term care: results and evaluation of a large scale webbased error reporting system. Qual Saf Health Care 2007; 16: 297-302.
Tuttle D, Holloway R, Baird T, Sheehan B, Skelton WK. Electronic reporting to
improve patient safety. Qual Saf Health Care 2004; 13: 281-86.
Amalberti R, Vincent C, Auroy Y, de Saint Maurice G. Violations and migrations in
health care: a framework for understanding and management. Qual Saf Health Care
2006; 15: i66-i71.
Harley M, Mohammed MA, Hussain S, Yates J, Almasri A. Was Rodney Ledward a
statistical outlier? Retrospective analysis using routine hospital data to identify
gynaecologists' performance. BMJ 2005; 330: 929.
Bates DW, Evans RS, Murff H, Stetson PD, Pizziferri L, Hripcsak G. Detecting
adverse events using information technology. J Am Med Inform Assoc 2003; 10: 11528.
Wyatt JC, Sullivan F. eHealth and the future: promise or peril? BMJ 2005; 331: 139193.
Healy JC. The WHO eHealth Resolution - eHealth for all by 2015? Methods Inf Med
2007; 46: 2-4.
Eysenbach G. What is e-health? J Med Internet Res 2001; 3: E20.
Oh H, Rizo C, Enkin M, Jadad A. What is eHealth (3): a systematic review of
published definitions. J Med Internet Res 2005; 7: e1.
Pagliari C, Sloan D, Gregor P, Sullivan F, Detmer D, Kahan JP et al. What is eHealth
(4): a scoping exercise to map the field. J Med Internet Res 2005; 7: e9.
Jones R, Rogers R, Roberts J, Callaghan L, Lindsey L, Campbell J et al. What is
eHealth (5): a research agenda for eHealth through stakeholder consultation and
policy context review. J Med Internet Res 2005; 7: e54.
Watson R. EU wants every member to develop a "roadmap" for ehealth. BMJ 2004;
328: 1155.
Lorenzi N. Medical informatics scientific content map. Available from:
http://www.imia.org/endorsed/2002_scientific_map.lasso (last accessed 3/1/11)
Pronovost P, Sexton B. Assessing safety culture: guidelines and recommendations.
Qual Saf Health Care 2005; 14: 231-33.
Bates DW, Gawande AA. Improving safety with information technology. N Engl J
Med 2003; 348: 2526-34.
Ash JS, Berg M, Coiera E. Some unintended consequences of information
technology in health care: the nature of patient care information system-related
errors. J Am Med Inform Assoc 2004; 11: 104-12.
Ash JS, Sittig DF, Campbell E, Guappone K, Dykstra RH. An unintended
consequence of cpoe implementation: shifts in power, control, and autonomy. AMIA
Annu Symp Proc 2006; 11-15.
Barber N, Rawlins M, Dean FB. Reducing prescribing error: competence, control,
and culture. Qual Saf Health Care 2003; 12: i29-i32.
Avery AJ, Sheikh A, Hurwitz B, Smeaton L, Chen YF, Howard R et al. Safer
medicines management in primary care. Br J Gen Pract 2002; 52: S17-S22.
146
SECTION 3
FINDINGS
147
Descriptive overview of results
Our initial searches revealed a total of 46,349 potentially eligible articles of which 67
were relevant systematic reviews. During the second phase of work, which included
an expansion into the field of telehealthcare, we identified a further 95 relevant
systematic reviews, bringing the total number to 162 systematic reviews (see
Appendix 3). These are listed in Appendix 4 and a critique of each individual review
is provided in Appendix 5.
The chapters in this section (Chapters 5-17) draw in the main on evidence from the
systematic reviews, which have been assessed using an adapted version of the
Critical Appraisal Skills Programme criteria (see Executive summary and Appendix
2), but we also, where necessary, draw on evidence from a broader body of trial,
technical, descriptive, qualitative and policy relevant work, all of which is referenced
in individual chapters, to help contextualise our findings.
148
Chapter 5
Summary
•
Effective and efficient sharing of clinical information is essential to the
future coherent development of health systems, which increasingly deliver
care to individual patients through an array of primary and secondary care
providers working in a range of locations.
•
The ideal in this respect is for professionals–and also patients if they so
wish–to have the ability simultaneously to access and seamlessly transfer,
contribute to and integrate clinical data from disparate sources.
•
Health information exchange and interoperability considerations aim to
provide a shared platform and syntax by which different types of systems
can safely and effectively communicate healthcare data.
•
Whilst there is as yet only limited research on the benefits and risks of
health information exchange and interoperability, investment in this
infrastructure is probably the key to realising economic returns associated
with other more directly patient relevant eHealth applications such as
electronic health record systems and ePrescribing.
•
Additional benefits will be gained through enabling reuse of coded data for
health planning, public health, research and audit.
•
Whilst increasing levels of interoperability are clearly desirable for many
reasons, greater centralised storage of sensitive data also inevitably
increases the risk of threats to data security and inadvertent breaches of
confidentiality.
•
Currently, most UK healthcare settings are characterised by relatively low
levels of interoperability capability, this being particularly true of the
hospital sector, where paper-based records are still widely used.
•
There is increasing recognition for the need to minimise the risks and
inefficiencies to care provision associated with the introduction of
technology that fails to integrate with existing IT systems.
149
•
The NHS has already begun to increase the potential for health
information exchange and interoperability through, for example, systems
such as LabLinks that send laboratory results electronically to general
practices, electronic transfer of discharge summaries and clinic letters, the
National Network for the NHS, the central Spine and associated NHS Care
Records Service; other developments include encouragement of common
operational standards such as Health Level Seven and co-operating with
the Continua Health Alliance.
•
Key outstanding issues that face healthcare systems in realising the
potential for seamless exchange of information include the need to
develop and deploy standard coding structures across all care settings
(e.g. using Systematized Nomenclature of Medicine-Clinical Terms),
facilitate integration of the increasing amounts of patient-generated data
(through HealthSpace, home sensors and telemetry devices, for example),
ensure that minimum standards of interoperability are implemented across
the increasingly decentralised NHS and improve secure audited access to
electronic health records to minimise the risks of threats to privacy.
•
Making progress in optimising the ability to share and integrate data
seems particularly relevant given the expressed government objective of
creating an “information revolution” within the NHS.
5.1 Introduction
Patients in most economically developed countries now commonly receive their
healthcare from a range of practitioners, many of whom work in different healthcare
settings.
For most people, this will include health centres, pharmacy practices,
dentists, and opticians, a variety of hospitals and laboratory services, and
occasionally ambulance services.
Guided self-monitoring through teleheathcare
may in the future also play an increasingly important role in care provision (see
Chapter 13).
150
All these services and systems hold electronic records for patients (and with
electronic self-monitoring the patients hold electronic information themselves).
However, such records are largely held in silos with limited or no ability to interact
with one another and little prospect for direct access by the patients and their
clinicians to the bulk of these records.
This situation can lead to wasted resources through, for example, double entry of
demographic data and, duplication of test requests, and increases the risk of errors
through delayed diagnosis, lost data and inaccurate data entry.
Effective and
efficient sharing of clinical information is therefore essential to the future
development of modern health systems.
This chapter aims to provide an overview of health systems characterised by high
levels of health information exchange and interoperability (HIEI) capability, review
the benefits that are likely to accrue from such developments and the possible risks
associated with prompting system change along these lines. The chapter is not
intended to be a detailed discussion of the technical infrastructures and ontologies
which underpin HIEI (for which interested readers are referred to Coiera’s
introduction1 or the review by Tanebaum2), but rather to provide a basis for
considering the key applications being introduced into the NHS, which are
considered in detail in subsequent chapters in this Section.
5.2 Definition, description and scope
5.2.1 Definition
Health information exchange and interoperability refers to the ability to access,
contribute to and integrate data from disparate sources. As defined by the Health
Information Security and Privacy Collaboration, it refers to:3
‘The ability of systems or components to exchange health information and to use the
information that has been exchanged accurately, securely, and verifiably, when and
where needed.’
151
5.2.2 Description
Health information exchange and interoperability allows clinicians and patients to
access necessary parts of the patient record. It does this by providing a shared
platform and syntax by which different types of systems can exchange data
seamlessly. However, the degree to which different systems can interact varies.
Walker et al. provide a helpful conceptually driven analytic framework and taxonomy
for HIEI, which is illustrated in Box 5.1.4 This four level taxonomy reflects the amount
of human involvement required, the sophistication of information technology (IT) and
the level of standardisation needed for different degrees of information sharing
capability between healthcare organisations.4
Box 5.1 A conceptual analytic framework and taxonomy for health information
exchange and interoperability
•
Level 1: Non-electronic data—no use of IT to share information (for example, ‘snail’
mail and meetings).
•
Level 2: Machine-transportable data—transmission of non-standardised information
via basic IT; information within the document cannot be electronically manipulated (for
example, fax or personal computer [PC]–based exchange of scanned documents,
pictures, or portable document format [PDF] files).
•
Level 3: Machine-organisable data—transmission of structured messages containing
non-standardised data; requires interfaces that can translate incoming data from the
sending organisation’s vocabulary to the receiving organisation’s vocabulary; usually
results in imperfect translations because of vocabularies’ incompatible levels of detail
(for
example,
email
of
free
text,
or
PC-based
exchange
of
files
in
incompatible/proprietary file formats, HL-7 messages (see Glossary).
•
Level 4: Machine-interpretable data—transmission of structured messages containing
standardised and coded data; idealised state in which all systems exchange
information using the same formats and vocabularies (for example, automated
exchange of coded results from an external lab into a provider’s EHR automated
exchange of a patient’s ‘problem list’).
4
Adapted from: Walker et al. (2005) Permission Requested
5.2.3 Scope
These considerations are clearly cross-cutting and of relevance to almost any
eHealth application, irrespective of whether this is used for storage, decision support
152
or support of remote delivery of care. Given the historic low-level of interoperability
between existing information systems in the NHS, there is the risk that this situation
will be compounded if due attention is not given to in the context of introducing new
applications.
Using this framework, it is clear that most systems in England will, despite the work
of NHS Connecting for Health (NHS CFH) and the introduction of the key National
Programme for Information Technology (NPfIT) deliverables (see Chapter 3), for the
foreseeable future be operating at Levels 1, 2 or 3. For example:
•
Transfer of records between GPs: Even though UK primary care has a long
history of using the electronic health record (EHR), changing general
practitioners (GPs) has typically necessitated the printing out of the patients’
record and the manual entry of a summary of this record into the new GP’s
system, this cumbersome process resulting in the loss of considerable
amounts of data (Level 1). Currently, an average general practice in England
deals with around 500 patient record transfers each year; inner-city and
university practices will deal with far more. Re-entering data creates a huge
administrative burden that raises the possibility of omitting important
information and transcription errors. More recently, for most practices certain
types of attachment (scanned files) can now be transferred, but these cannot
be easily integrated with different GP software (Level 2). As discussed in
Chapter 3, GP2GP, which permits communication between GP systems, is
now being rolled-out across England. The exchange takes place remotely
and automatically on completion of registration. This means that patients’
records, including, for example, full diabetes and blood pressure records, are
available within hours of registration and as a result many hours, and in some
cases, days of repeat data entry are avoided (Level 3). There are, however,
some problems that still need to be addressed, these including the fact that
the differences between the systems necessitates data “cleaning”; there is
furthermore the need to reorganise data, particularly when this involves
transfer of data between computer systems, but even when transfers occur
between practices that use the same computers systems if they do not
organise their data in exactly the same way.
153
•
Choose and Book: This eBooking system permits GPs and their patients to
access hospital appointment systems and make appointments at times to suit
them with the hospital of their choice (Level 3).
•
NHS Care Record Service (NHS CRS): The aim is that this will eventually
replace the current mix of electronic and paper records held throughout the
NHS allowing immediate access, with permission, to some aspects of the
patient’s record (the Summary Care Record (SCR)) from any part of the
country and more detailed clinical, laboratory and radiology records (Detailed
Care Record (DCR)) to local providers. This will in time allow the integration
of decision support functionality (Levels 3-4; see Chapters 3 and 6 for further
details).
•
Picture Archiving and Communications System (PACS): Until recently,
radiological images were printed on film and held in hospital records
departments. If patients needed to be seen at other locations (e.g. for a
specialist assessment), images had to be physically transported with many
not arriving in time, or being lost in transit, and resulting in repeat
investigations or delay in diagnosis.
The PACS, which has been widely
implemented by NHS CFH, now allows access to digital centrally stored
images within and between hospitals and eventually should permit full
interoperability and compatibility with other NPfIT services.
It permits, for
example, a doctor in an accident and emergency department to access
specialist opinion in another trust and images will, it is planned, also be
accessible in some primary care settings. The aim is that in due course,
PACS will integrate with the NHS CRS, thereby removing the cumbersome
barrier between images and other aspects of the patient record (Level 3-4).
•
ePrescribing and the Electronic Prescription Service (EPS): For most
patients, prescriptions are still generated by computer in their practice, printed
on paper and then taken to pharmacies where the same information is
entered into the pharmacy computing system to print labels and maintain a
database of their dispensing activity. The paper prescriptions are then sent
on to the pricing bureau to be again computer-entered for payment (Level 1).
The Electronic Prescription Service being implemented by NHS CFH should
allow for the transmission of prescriptions from GPs and other prescribers
154
directly to pharmacies, thus automatically populating their databases for
dispensing. Prescriptions are then also sent electronically by pharmacists to
the Prescription Pricing Authority (PPA), the organisation that reimburses
dispensers, i.e. community pharmacies for the medication they have supplied
to patients. The system provides a robust audit trail of prescribing activity,
reduces administrative effort and may reduce errors of transcription. Such
ePrescribing systems have the potential to decrease delay in the dispensing
of orders, reduce errors related to handwriting or transcription, allow order
entry on or off-site, error checking for duplicate or incorrect doses, and
facilitate inventory and posting of charges (Level 3; see Chapter 10).
•
Laboratory results: While laboratory data are digitally recorded and stored,
results are often printed out on paper and sent to those ordering the tests. In
many cases, these results are then scanned and entered into an electronic
record; the most relevant data are still in many cases still then manually
entered into, for example, the GP patient records (Level 1).
Recent
developments have provided access to all patients’ laboratory results across
several hospitals and practices (SCI store in Scotland)5 (Level 2) and results
for many practices in other parts of the UK may now be downloaded, for
example via LabLinks,6 directly into practice computer systems with
occasional need for modification (Level 3).
5.3 Theoretical benefits and risks
5.3.1 Benefits
The examples in the preceding section illustrate the types of benefits that more
integrated computer systems can yield. Walker et al., in a review of the potential of
HIEI to improve services in the United States (US),4 have summarised the broad
range of benefits that may be realised (see Box 5.2).
155
Box 5.2 The potential benefits of health information exchange and
interoperability
•
Interoperability between both freestanding and hospital-based outpatient
clinicians: This would enable computer-assisted reduction of redundant tests,
and it would reduce delays and costs associated with paper-based ordering and
reporting of results.
In addition, provider-laboratory connectivity would give
clinicians better access to patients’ longitudinal test results, eliminate errors
associated with reporting results orally, optimise ordering patterns by making
information on test costs readily available to clinicians, and make testing more
convenient for patient.
•
Connectivity between external radiology centres: This would reduce
redundant tests and would save time and costs associated with paper- and filmbased processes. Interoperability here could also improve ordering by giving
radiologists access to relevant clinical information, thereby enabling them to
recommend optimal testing; improve patient safety by alerting both the provider
and the radiologist to test contraindications; facilitate coordination of care and
help prevent errors of omission by enabling automated reminders when followup studies are indicated; and lessen adverse environmental impacts by reducing
the use of chemicals and paper in film processing.
•
Out-patient providers and pharmacies: Interoperability between out-patient
providers and pharmacies would reduce the number of medication-related
phone calls for both clinicians and pharmacists. It would also improve clinical
care by facilitating the formation of complete medication lists, thereby reducing
duplicate therapy, drug interactions and other adverse drug events, and
medication abuse. It could also enable automated refill alerts, offer clinicians
easy access to information about whether patients fill prescriptions, and
complete insurance forms required for some medications. In addition, it could
help identify affected patients in the event of drug recalls, uncover new side
effects, and improve formulary management.
•
Provider-provider connectivity: Has the potential to save time associated with
handling chart requests and referrals. Connectivity would reduce fragmentation
of care from scattered records and improve referral processes.
•
Provider connectivity to the US public health system: This would make
reporting of vital statistics and cases of certain diseases more efficient and
complete. However, the most important impact of public health interoperability
would almost certainly derive from earlier recognition of emerging disease
156
outbreaks and bio-surveillance, as it becomes easier to identify warning signs
and trends by aggregating data from many sources. Since robust quantitative
evidence about the value of HIEI in earlier recognition of disease and biosurveillance does not yet exist, we did not project value from these sources6.
•
Provider-payer transactions: These enjoy a relatively high degree of
standardisation,
largely
because
of
Health
Insurance
Portability
and
Accountability Act. Some transactions are highly automated, but others are not,
particularly in smaller organisations.
Most of the benefits they describe are also likely to apply to the NHS. In summary,
HIEI should save resources through the reduction of administration by ending repeat
data entry, preventing duplication of investigation, facilitating more targeted
investigation and improving access to records for front-line staff. This improved
access to more comprehensive records should also help obviate the problems
associated with access to the current fragmented records, which can, for example,
reduce the risks of errors that commonly occur when people move across healthcare
transition boundaries; e.g. being discharged from hospital back to primary care.
Walker et al. estimated the projected financial benefits in the US based on the
impact on laboratory services alone at $31.8 billion. In their analysis of aspects of
interoperability for which they were able to allocate:4
‘…dollar values, net savings from national implementation of fully standardised
interoperability between providers and five other types of organisations [they]
estimate that this could yield $77.8 billion annually, or approximately 5% of the
projected $1.661 trillion spent on US healthcare in 2003.’
Although a detailed analysis to quantify the patient safety benefits expected from the
NPfIT–which has had at its core the need to overcome historic problems with data
sharing–has not been conducted by NHS CFH, it believes, based on a limited
preliminary economic analysis that this could be worth several billions of pounds of
savings to the NHS. This estimate includes:
157
‘£2.5 billion as the human value of preventable fatalities from medication errors
arising from inadequate information about patients and medicines; a large proportion
of the £500 million spent each year on treating patients who are harmed by
medication errors and adverse reactions; a reduction in the payments by NHS Trusts
each year (approximately £430 million each year) for settlements made on clinical
negligence claims.'7
In discussing the potential impact on patient safety, Kaelber and Bates have
estimated that developments in HIEI should through optimising the availability of
information substantially reduce the incidence of prescribing errors and resulting
medication-related adverse drug events.8 Improved potential for continuity of care is
another important potential benefit of such developments.
The two detailed case studies by Blick (1997 & 2001) are also important in that they
provide insights into the potential gains associated with improved information
management in the fields of pathology services.9;10 In the first, he discusses the
need to focus on the broader non-analysis related quality considerations such as
specimen logging, tracking and the issuing of reports, which, based on simulations
with his advanced information management system, can greatly be improved.9 The
second paper highlights the difficulties posed by stand-alone point-of-care (POC)
systems in critical care medicine, these including the lack of recording of results in
patients’ electronic health records (EHRs) and the inability easily to bill for such
procedures.10 Such failings he hypothesises could again be reduced through use of
better designed and connected information management systems.
5.3.2 Risks
Movement across geographical boundaries
The ability to transmit data relies heavily on adequate identification of patients. In
England, patients now have a unique NHS Number.11 A unique Community Health
Index (CHI) number also exists in Scotland,12 therefore, even within the UK,
translation algorithms may be required when patients move between different
countries. This ability to track patients as they move has also been identified as a
major issue in countries that do not have unique identifiers.
158
Transcription errors
While HIEI should reduce the frequency with which data are keyed in and is hence
likely to reduce error, the potential to key in the wrong information into the wrong file
still exists. When transcription errors occur in systems interacting at Level 1 or Level
2, the necessary human interaction means that at least some check on data is
occurring. Obvious errors that a machine may not detect (for example, a pregnancy
test result wrongly attributed to a 65-year-old woman, or the contraceptive pill being
prescribed to a man) may go unnoticed unless plausibility checks are incorporated
into higher level systems.
Privacy and security
Security is a major concern. Any system which makes it easier for clinicians and
patients to access their data inevitably makes it easier for unauthorised individuals to
gain access. Similarly, the more people that have access to a patient’s record the
more risk there is that, whether by accident or design, confidential information will
leak.13
Casual access by clinical staff to unauthorised material is, for example, a long
recognised
challenge.14 Audit trails and personal logons which can identify
unauthorised entry are another option, but in the fast-paced environment of hospital
wards, GP surgeries and pharmacies it is easy for machines to remain logged-on
and unauthorised access to occur. While such breaches were always possible, HIEI
potentially increases the harm of such breaches.
System failure
The risk that failure in one component could have a major effect on the functioning of
the whole system increases exponentially as systems become more integrated. A
practice which keeps a local record will not, for example, be affected by a failure in
communication with a central server. Therefore, the HIEI system as a whole, and all
its components, must be reliable and be able to assure a uniform, satisfactory level
of service quality in addition to running being securely and regularly backed up, so
that organisations can rely on the overall system availability. It must provide for realtime access to information, particularly for urgent care specialties such as
emergency medicine and intensive care.
159
Considerations relating to Implementation
As previously mentioned, most UK healthcare settings are characterised by relatively
low levels of HIEI; this is particularly true of the hospital sector. Similar problems
exist in the US and solutions have tended to evolve locally, but this can create
subsequent problems with reluctance to yield ownership of a well-known local
system to become part of a larger more integrated system with thus no clear ways
for HIEI capabilities to easily expand beyond these initial projects.15 Likewise some
vendors who rely on interoperability of their products as a means of generating
revenue are likely to see the use of a common platform as a threat. There are
parallels here with general practice computing in the UK.
Improving HIEI to the desired level is likely to generate long-term revenue savings;
however, this will require considerable up-front investment in hardware and software
capabilities with early adopters paying most. There is considerable uncertainty in
some quarters about the extent to which the development of EHR should proceed to
before agreed systems of HIEI are developed with early adopters of such records
fearful that any subsequent standards agreement will leave them stranded with
obsolete systems. However, experts in the US strongly advocate the integrated
approach, expressing serious reservations about the potentially wasted energy in
developing local systems.15
5.4 Empirically demonstrated benefits and risks
5.4.1. Benefits
Given that this remains a relatively new field of empirical enquiry, we unsurprisingly
found only limited empirical evidence summarised in systematic reviews (see
Appendix 5) in well-circumscribed areas to draw upon. Below, we discuss the main
findings contained within these three systematic reviews (SR) and additional
selected original investigations.
In the SR by Georgiou et al, they considered the impact of computerised physician
order entry (CPOE) systems embedded in hospital pathology information
management systems.16 This review identified 19 controlled studies employing a
variety of designs, including randomised controlled trials (RCTs), quasi-experimental
160
designs and before-after studies. This review helpfully identified a suite of indicators
to assess the impact of these computerised systems on various aspects of pathology
service function.
Overall, however, they concluded that there are some data to
support the impact of these integrated systems on process measures such as a
reduction in the number of tests, the ordering of redundant tests and the costs of
these investigations, there is as yet little evidence of a beneficial impact on patient
outcomes such as length of stay, adverse events and mortality.
In a SR focusing on communication between hospitals and family practices Kripalani
et al. found that deficit in communication and information transfer at hospital
discharge are common, these needlessly jeopardising patient care.17 From their
critique of a large body of evidence (73 studies in total), they were able to identify 18
controlled investigations (which included three RCTs) focusing in the main on
comparisons between computer- and hand-generated discharge summaries.
Comparison between studies was challenging because of the range of outcome
measures used, but nonetheless overall they found that computer-generated
summaries were generated by hospital clinicians in a more timely fashion; there was
also some evidence of improved quality of information in these summaries and the
timing of the receipt of these summaries by family practitioners. There was however
no clear evidence that these improvements in the processes of care translated into
improved clinical outcomes for patients.
More recently, Fontaine et al have reported a SR on HIEI amongst primary care
providers.18 Their investigations identified a total of over 60 eligible reports, of which
39 were peer-reviewed publications. Twenty of these peer-reviewed publications
presented original data; however, only two of these studies employed RCT designs.
These original studies focused on a range of relevant outcome measures, but overall
the only robust finding was for evidence of increased efficiency in relation to
accessing relevant patient data generated from outside the practice (such as
pathology results) and staff time involved in processing claims. There was relatively
little in the way of clear evidence of positive impacts on either the quality or safety of
care. Importantly, they concluded that it was not as yet possible to demonstrate a
positive return on investment.
161
There have been early reports of regional attempts at developing interoperability
networks in the US, which provide helpful pointers to some of the challenges that
have needed to be overcome and structures and processes that can prove helpful in
making progress locally.
Halamka et al. have, for example, reported on their
experiences of establishing an interoperability network in Massachusetts, which
focused on the sharing of clinical and financial data.15
This has been a time-
consuming process, but one which the authors believe has been successful as
judged by the feasibility work, and the development of relevant structures and
policies that have helped to place this network on a stable footing. It is also claimed
that this has resulted in substantial cost savings in the processing of claims (from $5
to 25 cents per transaction).
Early successes in the feasibility and claims about the sustainability of creating data
exchange consortia were however also made by the Santa Barbara County Care
Exchange (SBCCE) scheme, and they too reported the ability to generate cost
savings was a key ingredient to the early success of this venture.19 However, this
scheme has recently disbanded for reasons that included a lack of true multistakeholder involvement, a pre-occupation in high-tech solutions at the expense of
meeting the everyday needs of healthcare providers, over-optimistic assumptions on
financial returns and the difficult question of how to balance the sharing of
information with privacy concerns.20;21;22
Taking a mainly European perspective, Dobrev et al have conducted an interesting
series of economic case studies of implementations of EHRs and ePrescribing
systems in 11 countries.23 These detailed investigations involved a combination of
literature reviews, qualitative interviews, analysis of cost and benefit data, and
economic modelling, from which they concluded that investment in eHealth requires
up-front investment from healthcare organisations, but that this in most cases results
in net economic gains to society (but not necessarily healthcare organisations). In
contrast with many of the claims made about eHealth, they conclude that this does
not therefore necessarily equate with a return on investment, i.e. the capital outlay
translates into improvements in the quality of care, but at a cost. Another important
conclusion and one that is of considerable relevance to the discussions in this
chapter is that these benefits are primarily mediated through enhanced inter162
operability, which results in the ability to reuse data for multiple ends such as audit,
research, planning and surveillance. Finally, the authors underscored the point that it
typically takes years for these benefits to accrue: investment in eHealth must not
therefore be seen as an “easy win”.
5.4.2 Risks
Johnson has in an unpublished study identified vulnerabilities in highly integrated
information systems in both the US (Veteran Affairs Administration system) and
England (Lorenzo), which led to system downtimes necessitating a return to paperbased models of care for which healthcare organisations and professionals were illprepared.24
He urges the need to take steps to minimise such risks and also for
contingency planning for scenarios in which downtimes occur.
The issue of consent and privacy is also important and has in England at least in
particular centred on consent models for the Summary Care Record. The response
of moving from an opt-out to opt-in model has only partially allayed concerns in this
respect.13;25;26
Also of relevance to HIEI is the ongoing issue of concerns relating to data quality in
EHRs, which typically serve as the lynchpin for integration between different
information systems.27 Inaccuracies and ambiguities in the core clinical record can,
as a result of information sharing, be magnified and result in additional work to
reconcile records.
Finally, as the discussion on the Santa Barbara21 and European Commission23
above highlight, there is likely to be a need for substantial upfront and net
expenditure to achieve effective HIEI with no guaranteed direct returns to healthcare
organisations or indeed society on such investments.
5.5 Implications for policy, practice and research
5.5.1 Policy
There are several features of UK practice which provide a positive platform for
increased HIEI, although important challenges remain.
This includes the single-
payer nature of the NHS, which in contrast to other health systems where there may
163
be a multitude of healthcare purchasers, both public and private. There is also, for
example, strong central government support for increasing information sharing.28
The central procurement of major IT by NHS CFH now allows for much greater
enforcement of standards nationally in relation to interoperability compared to
previously haphazard IT commissioning with numerous systems, although concerns
are being raised about a return to the previous picture resulting from the increasing
move to localisation in the post-New Labour era. The certification model now being
pursued in the US through the Certification Commission for Health Information
Technology is in this respect potentially of considerable relevance to the UK
situation.29
The adoption of Health Level Seven (HL7)–which is a volunteer organisation that
“…provides a framework (and standards) for the exchange, integration, sharing and
retrieval of electronic health information through defining standards, guidelines and
methodologies”32–by NHS CFH is a particularly important development.30 The HL7
version 3 messages and clinical document architecture are used by the NHS to
facilitate the transfer of information. Importantly, this standard is being adopted
internationally, for example, in Australia by the National EHealth Transition
Authority.31 Alternative systems such as OpenEHR32 also have strong support and
are gaining ground in many European countries.
While ensuring computable semantic interoperability between complex datasets will
continue to pose significant problems to healthcare IT,33 security and confidentiality
also remain a challenge and further developments need to be encouraged to
reassure both clinical staff and the public that information, widely available to many
different groups within the NHS, will remain secure.
5.5.2 Practice
One of the key challenges that face healthcare systems in realising the potential for
seamless exchange of information is the development and deployment of standard
coding structures across all care settings. Currently, in the UK, different codes are
used by hospitals and general practices. However, NHS CFH, with strong central
backing, has adopted the Systematized Nomenclature of Medicine-Clinical Terms
(SNOMED-CT) system.34;35 This is a common computerised language that will in due
164
course be used by all computers in the NHS to facilitate communications between
different healthcare professionals. It is a joint development between the NHS and
the College of American Pathologists to create an agreed terminology and is likely to
be widely adopted internationally. It has greater depth and coverage of healthcare
than other versions of clinical terms such as Read codes and will facilitate the
exchange of healthcare and clinical knowledge by clinicians, researchers and
patients worldwide.
One of the real opportunities for the future of healthcare, but an important challenge
to integrated care, is the growth of patient-generated data from telemetric devices.
The Continua Health Alliance is attempting to standardise protocols and improve
interoperability of such devices and NHS CFH is an important member of this
consortium.36
The current policy of encouraging HIEI through the rigorous setting of standards,
through contractual arrangements, for software and hardware intended to link with
the SCR has been shown to be effective, particularly in primary care computing
systems (for example, GP2GP) and such arrangements should continue. Part of the
reasons underpinning the success of GP2GP has been the close liaison with general
practice representative bodies such as the British Medical Association’s (BMA)
General Practitioners Committee (GPC) and the Royal College of General
Practitioners (RCGP) on the GP2GP Project Board, this dating back to the outset of
the project.
Such liaisons in general practice and other parts of the NHS are
important in facilitating the introduction of innovations such as HIEI (see Chapters 17
and 18).
5.5.3 Research
Further research to identify both the benefits and costs of HIEI and the barriers (for
example, concerns regarding confidentiality and commercial interests) and
facilitators to its implementation are required. Research on how patient-generated
data can easily be incorporated into the EHR and how these are then synthesised
and made use of is also necessary.
165
More generally, the UK occupies a strong international position in terms of access to
health and other related datasets and the expertise in analysing these data, whether
for public health surveillance, epidemiological work or health services research. The
relatively few linkages between the many stand-alone datasets however prevents
innovation in this area, necessitating complex and less than perfect algorithms to
undertake linkages; more widespread use of unique identifiers could help greatly in
helping to unlock the potential within these datasets and help to further the
information revolution that is now being called for.28
References
1. Coiera E. Guide to Health Informatics. 2nd ed. London: Hoder Arnold; 2003.
2. Tanebaum AS. Computer Networks. 4th ed. Englewood Cliffs, New Jersey: Prentice
Hall; 2002.
3. Health Information and Security Privacy Collaboration.
Available from:
http://healthit.hhs.gov/portal/server.pt?CommunityID=1206&spaceID=399&parentname=
&control=SetCommunity&parentid=&PageID=0&space=CommunityPage&in_hi_totalgrou
ps=1&in_hi_req_ddfolder=6652&in_ra_topoperator=or&in_hi_depth_1=0&in_hi_req_pag
e=20&control=advancedstart&in_hi_req_objtype=18&in_hi_req_objtype=512&in_hi_req_
objtype=514&in_hi_req_apps=1&in_hi_revealed_1=0&in_hi_userid=8969&in_hi_groupo
perator_1=or&in_hi_model_mode=browse&cached=false&in_ra_groupoperator_1=or&in
_tx_fulltext=health+information+exchange (last accessed 28/12/10).
4. Walker J, Pan E, Johnston D, Adler-Milstein J, Bates DW, Middleton B. The value of
health care information exchange and interoperability. Health Aff (Millwood) 2005; W510-W5-18.
5. The Scottish Government website. eHealth Strategy 2008-11.Available from:
http://www.scotland.gov.uk/Publications/2008/08/27103130/4 (last accessed 28/12/10).
6. Department of Health. Report of the Review of NHS Pathology Services in England. An
Independent Review for the Department of Health.
Available from:
http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGui
dance/DH_4137606.
7. National Audit Office. The National Programme for IT in the NHS.
http://www.nao.org.uk/publications/nao_reports/05-06/05061173.pdf
(last
accessed
28/12/10).
8. Kaelber DC, Bates DW. Health information exchange and patient safety. J Biomed
Inform 2007; 40: S40-5.
9. Blick KE. Decision-making laboratory computer systems as essential tools for
achievement of total quality. Clin Chem 1997; 43: 908-12.
10. Blick KE. The essential role of information management in point-of-care/critical care
testing. Clin Chim Acta 2001; 307: 159-68.
11. NHS
Choices.
The
NHS
Number.
Available
from:
http://www.nhs.uk/NHSEngland/thenhs/records/Pages/thenhsnumber.aspx
(last
accessed 28/12/10).
12. The
Scottish
Government.
Community
Health
Index
number
http://www.scotland.gov.uk/Topics/Health/NHS-Scotland/DeliveryImprovement/1835/1865/1852 (last accessed 28/12/10).
13. Anderson R, Brown I, Dowty T, Inglesant P, Health W, Sasse A. Database State.
London: Jospeh Rowntree Reform Trust, 2009.
166
14. Silvester N. Doctor who hacked into Prime Minister's health records escapes
http://www.dailyrecord.co.uk/news/healthprosecution.
Available
from:
news/2010/01/10/doctor-who-hacked-into-prime-minister-s-health-records-escapesprosecution-86908-21955907/ (last accessed 27/12/10).
15. Halamka J, Overhage JM, Ricciardi L, Rishel W, Shirky C, Diamond C. Exchanging
health information: local distribution, national coordination. Health Aff (Millwood) 2005;
24: 1170-79.
16. Georgiou A, Williamson M, Westbrook JI, Ray S. The impact of computerised physician
order entry systems on pathology services: A systematic review. Int J Med Inform 2006;
76: 514-29.
17. Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW. Deficits in
communication and information transfer between hospital-based and primary care
physicians: implications for patient safety and continuity of care. JAMA 2007; 297: 83141.
18. Fontaine P, Ross SE, Zink T, Schilling LM. Systematic review of health information
exchange in primary care practices. J Am Board Fam Med 2010; 23: 655-70.
19. Brailer DJ, Augustinos N, Evans L, Karp S. Moving toward electronic information health
exchange: Interim report on the Santa Barbara County Data Exchange. Prepared for
California Healthcare Foundation. 2003.
20. Brailer DJ. From Santa Barbara to Washington: a person's and a nation's journey toward
portable health information. Health Aff (Millwood). 2007; 26: w581-8.
21. Miller RH, Miller BS. The Santa Barbara County Care Data Exchange: what happened?"
Health Affairs 2007; 26: w568–w580.
22. Vest JR, Gamm LD. Health information exchange: persistent challenges and new
strategies. JAMIA 2010; 17: 288-94.
23. Dobrev A, Jones T, Stroetmann V, Stroetmann K, Vatter Y, Peng K. Interoperable
eHealth is Worth it: Securing Benefits from Electronic Health Records and ePrescribing.
Bonn/Brussels: European Commission, 2010.
24. Johnson CW. Case studies in the failure of healthcare information systems. Personal
communication, 2009.
25. Greenhalgh T, Stramer K, Bratan T, Byrne E, Mohammad Y, Russell J. Introduction of
shared electronic records: multi-site case study using diffusion of innovation theory. BMJ
2008; 337: a1786.
26. Greenhalgh T, Stramer K, Bratan T, Byrne E, Russell J, Potts HW. Adoption and nonadoption of a shared electronic summary record in England: a mixed-method case study.
BMJ 2010; 340: c3111.
27. Thiru K, Hassey A, Sullivan F. Systematic review of scope and quality of electronic
patient record data in primary care. BMJ 2003; 326: 1070.
28. Department of Health. Liberating the NHS: An Information Revolution. London, 2010
29. Certification Commission for Health Information Technology. Available from:
http://www.cchit.org/ (last accessed 28/12/10).
30. Health Level 7. Available from: http://www.hl7.org.uk/ (last accessed 27/12/10)
31. National E-Health Transition Authority. http://www.nehta.gov.au/ (last accessed
28/12/10).
32. OpenEHR. Available from: http://www.openehr.org/home.html (last accessed 28/12/10).
33. Mead CN. Data interchange standards in healthcare IT--computable semantic
interoperability: now possible but still difficult, do we really need a better mousetrap? J
Healthc Inf Manag 2006; 20: 71-78.
http://www.ihtsdo.org/nc/our-standards/snomed34. SNOMED-CT.
Available
from:
ct/?sword_list%5B%5D=snomed (last accessed 28/12/10).
35. Simpson CR, Anandan C, Fischbacher C, Lefevre K, Sheikh A. Will Systematized
Nomenclature of Medicine-Clinical Terms improve our understanding of the disease
burden posed by allergic disorders? Clin Exp Allergy 2007; 37: 1586-93.
36. Continua Health Alliance. Available from: http://www.continuaalliance.org/index.html (last
accessed 28/12/10).
167
Chapter 6
Electronic health records
Summary
•
The electronic health record is a complex construct encompassing
digitised patient and healthcare records and the information systems into
which these are embedded.
•
A range of terms has been used to describe different types of electronic
health record and there is, at yet, no one agreed taxonomy. Electronic
health records systems vary in their complexity, with some being closely
coupled to other eHealth functionalities such as decision support. These
sources of variation are evident in published evaluations, presenting a
challenge to evidence synthesis to inform policymaking.
•
Electronic health records are at the heart of eHealth implementation plans
around the world. The long-term vision is for fully integrated, longitudinal,
patient records, supported by universal data and interoperability
standards,
and
high
level
communication
and
decision
support
technologies. This is some way from being realised.
•
Electronic health records have the potential to improve the accessibility,
completeness and accuracy of healthcare documentation and support
clinical information exchange across teams. Electronic health record
systems that integrate decision support can aid clinical decision-making
and guideline compliance, enhance health promotion activities and reduce
medical errors and adverse drug events. Those that are accessible to
patients may enable self-management and shared decision making.
•
Many of the practical benefits of electronic health records, such as
improved care quality, patient safety and public health, may arise from the
secondary uses of data for audit, surveillance, planning and research.
•
Despite the many theoretical benefits of electronic health record systems,
previous studies have yielded mixed evidence of their effectiveness and
there have been relatively few long-term economic appraisals.
168
•
Theoretical benefits pertaining to savings in time and cost have yet to be
substantiated by empirical evidence although there is some evidence of
enhanced data accuracy, completeness, legibility and accessibility.
•
Achieving the successful adoption of electronic health records depends
greatly on understanding the context in which they are to be implemented,
including characteristics of the organisation and the work processes of
potential users, as well as on the quality of technical support and training
provided and the extent to which decision support is integrated.
•
The quality of data recorded in electronic health records varies widely, due
to a range of socio-technical factors surrounding individual users,
variations in the mechanisms and practice of data coding, and
compatibility of the different systems used. Standardised and widely
accepted measures of data quality in electronic health records are lacking,
but the introduction of the Systematized Nomenclature of Medicine-Clinical
Terms has potential to improve the consistency of clinical coding across
the NHS.
•
The sensitive nature of the personal information held in electronic health
records calls for clear technical and governance measures to ensure
appropriate access and safeguard privacy. While patients generally find
clinical uses of electronic health records acceptable, secondary uses of
privileged health data by third parties arouse more concern.
6.1 Introduction
The implementation of electronic health records (EHR) is a core objective of
national and organisational eHealth strategies worldwide. This is driven by a
recognition of the potential benefits of EHR for healthcare quality and
efficiency, as well as their central importance for the delivery of eHealth
systems and services, as illustrated in the conceptual map in Chapter 4.1
This chapter reviews core concepts surrounding EHR, considers the
theoretical benefits and risks associated with their introduction, and reviews
169
the empirical evidence demonstrating their impacts on quality, safety and
organisational efficiency.
6.2.1 Definitions
The EHR is a complex construct encompassing different types of digitised
medical and healthcare records and the information systems into which these
are embedded. Many terms have been used to describe different aspects of
the EHR and it has been difficult to achieve universal consensus on a single
taxonomy.
The EHR has been defined in terms of: data, e.g.: “an individual patient's
medical record in digital format”;2 systems, e.g. “the set of components that
form the mechanism by which electronic health records are created, used,
stored, and retrieved. It includes people, data, rules and procedures,
processing and storage devices, and communication and support facilities”;3
and databases e.g. “a repository of information regarding the health of a
subject of care, in computer processable form”.3 More comprehensive
definitions recognise all of these components.a Examples of prominent
European, United States (US) and international policy definitions are provided
in Box 6.1. These are broadly compatible and tend to emphasise the role of
the EHR as an integrator of patient information across time and between
providers, although, to an extent, this objective still represents a vision rather
than a reality.
a
EHR literature has also been categorised in terms of its philosophical underpinnings
19
170
Box 6.1: Definitions of electronic health records used by key UK, US
and International Organisations
“The concept of a longitudinal record of a patient’s health and healthcare to combine
information from primary healthcare with periodic care from other institutions”
[Source: UK Department of Health]4
“A longitudinal collection of patient-centric, healthcare information, available across
providers, care settings, and time. It is a central component of an integrated health
information system.” [Source: US Institute of Standards & Technology]5
“Repositories of electronically maintained information about individuals’ lifetime
health status and healthcare, stored such that they can serve the multiple legitimate
users of the record.” [Source: EC, Information Society and Media]6
“Contains all personal health information belonging to an individual; is entered and
accessed electronically by healthcare providers over the person’s lifetime; and
extends beyond acute inpatient situations including all ambulatory care settings at
which the patient receives care.” [Source: WHO]7
“A repository of information regarding the health status of a subject of care in
computer processable form, stored and transmitted securely, and accessible by
multiple authorised users. It has a standardised or commonly agreed logical
information model which is independent of EHR systems. Its primary purpose is the
support of continuing, efficient and quality integrated health care and it contains
information which is retrospective, concurrent, and prospective.” [Source: ISO]3
A longitudinal electronic record of patient health information generated by one or
more encounters in any care delivery setting. Included in this information are patient
demographics, progress notes, problems, medications, vital signs, past medical
history, immunizations, laboratory data, and radiology reports. The EHR automates
and streamlines the clinician's workflow. The EHR has the ability to generate a
complete record of a clinical patient encounter, as well as supporting other carerelated activities directly or indirectly via interface—including evidence-based
decision support, quality management, and outcomes reporting.” [Source: HIMSS]8
Box 6.2 illustrates some of the more specific terms that have historically been
used to describe different types of electronic records in healthcare. These are
not mutually exclusive and the landscape is complicated by widespread
variation in the uses of particular terms to describe equivalent concepts.b
There is nonetheless an important conceptual distinction to be made between
b
For example, electronic medical record in the US, electronic health record in
Canada, electronic patient record in England and electronic health care record in
Europe.
171
digitised patient-specific records gathered in one setting (whether episodic or
longitudinal) and those that draw together information from multiple settings,
as in the integrated, longitudinal model referred to above. Terms used to
describe the former often include the words ‘patient’ or ‘medical’, whilst there
is an emerging international consensus around the use of ‘electronic health
record’ to reflect the broader concept. In light of the semantic complications,
we have elected to use EHR as a generic term in this report, as others have
also begun to do, but we remain mindful of the underlying complexity.9
It is worth noting the link between the concept of health information exchange
(HIE), which concerns systems that enable the transfer of data between
providers, and the concept of the EHR. While, on the one hand, the former
involves communications and the latter data, the systems into which EHR are
set are very often equivalent to those described under the banner of HIE. This
conceptual link is illustrated in this extract from a recent commercial
document:
“… EHR relates to something much broader – namely, a system-wide health
record that takes data through a Health Information Exchange. The EHR
contains all patient data and follows the individual wherever he or she goes.
The HIE pulls in data relating to that patient from the insurance company or
government agency, from hospitals and physician offices, labs, pharmacies
and other sources of clinical and administrative data. Then, using clinical best
practices from organizations such as the American Medical Association, it
analyzes the patient information and derives alerts and recommendations to
turn the data into actionable information. As a result, a fragmented healthcare
environment is transformed into a single portal where data are consolidated
and
recommendations
are
provided
at
the
point
of
care.”10
172
Box 6.2 Examples of Terms Used to Describe Types of Electronic Health
Record11-15
•
Electronic Medical Record (EMR),16
(EPR),
21 22 23
17 18 19 20
Electronic Patient Record
Computerised Medical Record (CMR),22
24
Computerised
Patient Record (CPR),25 Computer-Based Medical Record (CBMR),26
Computer-Based Patient Record (CBPR),27 Digital Medical Record28
29
(sometimes reserved for web-based records), Automated Health Records
(AHR),30
31
and Automated Medical Records (AMR).32 Largely equivalent terms
used to describe digitally stored patient records, including those created on
computer or transcribed or scanned from paper records (AHR, AMR). These
usually refer to information from single providers (e.g. a GP practice, a diabetes
clinic, a hospital), but are often used interchangeably with the broader term EHR.
In the UK, the terms Summary Care Record (SCR) and Emergency Care
Summary (ECS) have been used to refer to an abbreviated patient record held
nationally for use in emergency or out-of-hours settings, which is typically derived
from the electronic primary care record,33
34 35
whilst Detailed Care Record
(DCR) is used to describe more detailed records held in one setting (i.e.
EMR/EPR).36 The term Continuity of Care Record (CCR) describes a standard
for a summarised EHR that can be accessed by and added to by different health
professionals managing a patient, so as to support integrated and concurrent
care. 37 38
•
Electronic Health Record (EHR), Electronic Care Record (ECR),39
Integrated Care Record (ICR),40 Electronic Health Care Records (EHCR):12
A record that contains information from multiple providers of the patient’s care.
It varies in how the information is integrated (e.g. centralised data storage
versus linkage to federated data stores), how much information is integrated
(detailed or summary), and the providers of the information (e.g. an integrated
diabetes care record vs. a more generic shared care record). Longitudinal
health record (LHR) is used to describe the EHR when available over time
41
42
•
Interoperable Electronic Health Record (IEHR):43 44 This term has been used
to describe an EHR that can interface with a range of records systems and
databases, and with tools for decision-making and communication.
173
Box 6.2 continued…
•
Personal Health Record (PHR):45 Sometimes also referred to as Electronic
Patient Carried Medical Records (EPCMR) or Patient Medical Records
(PMR),46 Personal Electronic Health Record (PEHR),47 Electronic Personal
Health Records (ePHR).48 These refer to records that are accessible by the
patient themselves. These vary in locus of control (patient or provider as author
of content or controller of access), medium (e.g. SmartCard, PC, web), functional
complexity (e.g. inclusion of patient information, prescription renewal, booking)
and extent of integration with provider-held EHR systems. In Australia, the term
Person-Controlled Electronic Health Record (P-CEHR) describes an EHR that
is accessible to the patient, who can choose which healthcare professionals may
access it.49
50
Elsewhere the term Personally Controlled Health Record
(PCHR) has been used to describe a class of electronic Personal Health
Record (PHR) such as Google Health and Microsoft’s HealthVault in which the
patient controls both record access and content.
•
Population health record:51 refers to “aggregated and usually de-identified data
[which] may be obtained directly from EHRs or created de novo from other
electronic repositories. It is used for public health and other epidemiological
purposes, research, health statistics, policy development, and health service
management.”
•
3
Virtual EHR:52 This has no authoritative definition, but is related to the concept of
an online record.
NB. While this box focuses on records, rather than information systems, the two
concepts are closely related and the term EHR has been used to refer to both.
In England, previous Department of Health (DH) reports have attempted to
differentiate the EHR from the EPR by defining the latter as a record of a
patient's contacts with one healthcare provider, such as a general practice,
and the former as a record of their overall health and healthcare, which
combines information stored in EPRs held by different healthcare providers,
such as general practices, community services, and hospitals.
However,
many reports and studies do not make this distinction.53
174
The term Personal Health Record (PHR) has been used to describe a range
of patient-oriented EHR, ranging from those designed for the individual to
record and manage their own health information for personal use, to those
that that aid the sharing of relevant health information between patient and
provider, through enabling the patient to access their provider-held EPR,c or
to record data from personal health-monitoring devices. Different platforms
and applications have been designed to support PHR, and the online ‘portal’
model, used by organisations such as Kaiser Permanente in the US, is
gaining currency.48 Such portal models also enable patients to undertake a
variety of administrative communications, such as appointment booking and
prescription renewal and to access patient information. There remains wide
variability between PHR models and approaches. For example, in a recent
report on EHR impact for the European Commission’s Information Society
and Media Directorate-General, Groen et al firmly place PHR at the patientcontrolled end of the spectrum, whilst acknowledging that a range of
healthcare providers may contribute towards it. According to these authors,
“The personal health record (PHR) is different from an electronic health record
(EHR) system maintained by a healthcare provider organisation. The PHR is
maintained by the individual patient. These individuals own and manage the
information in the PHR, which comes from both multiple healthcare providers
and the individuals themselves.”
6 54
Box 6.3 illustrates some of the personal
health management functions that such a system may provide.
The US National Committee on Vital and Health Statistics, amongst others,
has noted on the difficulties presented by the variable conceptual and
technical landscape of PHR existing today “This lack of consensus makes
collaboration, coordination and policymaking difficult. It is quite possible now
for people to talk about PHRs without realizing that their respective notions of
them may be quite different”.55
c
In the UK, this has generally been considered under the banner of ‘Record Access’,
since more complex PHR technologies remain common.
175
Box 6.3: Different Features of Personal Health Records
Potential Functions of electronic
Personal Health Records
•
•
•
•
•
•
•
•
Access to provider’s electronic
clinical record
Personal health organiser or
diary
Self management support
Secure patient-provider
communication
Links to information about
illness, treatments, or self care
Links to sources of support
Capture of symptom or health
behaviour data
Patient empowerment and
education
Information that may be contained in
the Personal Health Record
•
•
•
•
•
•
Fitness data, data from routine
medical checkups
Automatic interpretations and/or
suggestions based on data
entered
Data imported from other health
devices
Details of medical appointments
with automatic reminders
Over the counter medication
record
Health and wellness information,
including
patient-oriented
summaries and links to research
Adapted from Pagliari et al (2007), Ross et al (2003) and Groen et al (2007),
46 56 6 52
with permission
In the UK, patient awareness and use of PHRs remains low. While lack of
exposure to such technologies is clearly a contributing factor, the explanation
may also lie in aspects of the UK healthcare system and culture which differ
from those in other parts of the world where PHR are available. For example,
a key benefit of PHR in disaggregated health systems is that they can act as a
point of data integration mediated by the patient. In the UK, general
practitioners effectively have this integrating role and the benefits of PHR may
be more around addressing the ‘Information Society’ agenda of enabling
citizens to access their personal information and supporting accessible online
services, as well as accommodating the growing use of personal heath
monitoring devices, which provide relevant data streams outwith the NHS.
This may help to explain why the fledgling NHS system HealthSpace (which
provides patient access to the NHS Summary Care Record (SCR), booking
functions and patient information) was not widely taken up in early pilots.45 In
contrast, ‘home grown’ systems designed in general practice to enable
patients to view their detailed GP record via kiosk or online portal appear have
met with more favourable responses, particularly in patients with complex or
long term health needs.57
176
The relationships and differences between the terms are illustrated in Figure
6.1 below. Although it is complex to capture the meanings these varied terms
denote and the contexts in which they are used, an attempt is made in Figure
6.2 to provide an overview of this terminology.
177
Fig 6.1: Map of key terms used to describe electronic health records
Population Health Record
Electronic Health Record (Generic Term)
Integrated &
longitudinal
Usually provider
or setting-specific
Aggregated & anonymised
•
Electronic Health Record (EHR)
•
Integrated Health Record (IHR)
•
Electronic Health Care Record (EHCR)
•
Interoperable Electronic Health Record (IEHR)
•
Integrated Care Record (ICR)
•
Longitudinal Health Record (LHR)
Subset of data,
for emergency
or multispecialty care.
•
Summary Care Record (SCR) or
•
Electronic Patient Record (EPR)
•
Emergency Care Summary (ECS)
•
Electronic Medical Record (EMR)
•
Continuity of Care record (CCR)
•
Computerised Medical Record (CMR)
•
Computerised Patient Record (CPR)
•
Computer-Based Medical Record (CBMR)
•
Digital Medical Record (DMR, sometimes
reserved for web-based records)
•
Electronic Client Record (ECR)
•
Automated Health Record (AHR)
•
Automated Medical Record (AMR)
Older term,
usually a digital
version of a
paper record
•
Personal Health Record (PHR)
•
Electronic Patient-Carried Medical Record (EPCMR)
•
Electronic Personal Health Record (ePHR)
•
Personally Controlled Patient Record (PCPR)
Patient held,
accessible,
or controlled
NB. Personal Health Record may also have elements of the integrated EHR
178
Data
providers
and users
GP
GP
clinic
EMR
Time
Clinician
Hospital
EMR
Lab specialist
PHR
Radiology
EMR
(PACS /
RIS)
Pharmacy
EMR
LHCR
Pharmacist
Primary use
of data
Radiologist
IHCR
Laboratory
EMR (LIS)
Registries
Data & Services
Health Information
Data Warehouse
EMR
EPR
Figure 6.2 Infostructure of EHR, with index below
Electronic Social
Care Record
Community
Care EMR
Multiple
health care
professionals
Interoperability
Middleware
Multiple
patients
Non-Shareable
EHR
Confidential
Data
179
Researchers,
Policy
makers etc
Secondary
use
Shareable
EHR
Anonymised
Data
EHR - electronic medical record of multiple episodes of care of a population by several healthcare professionals over a long period of time
the patient over the web
180
PHR - electronic medical record of multiple episodes of care of a patient by several healthcare professionals over a long period of time accessible by
LHCR - electronic medical record of multiple episodes of care of a patient by several healthcare professionals over a long period of time
IHCR - electronic medical record of multiple episodes of care of a patient by several healthcare professionals over a short or long period of time
EPR - electronic medical record of multiple episodes of care of a patient by one healthcare professional over a period of time
EMR - electronic medical record of a single episode of care of a patient by one healthcare professional
Definitions used:
Abbreviations used: Laboratory information system LIS, Picture Archiving and Communication Systems PACS, Radiology Information Systems RIS
Health Informatics –Electronic Health Record –Definition, scope and context, ISO, 2005.3
EHRS Blueprint-an interoperable EHR framework, Canada Health Infoway, April 2006;58
EHR Impact study, European Commission Information Society and Media, April 2008;6
Source: In part based on:
6.2.2 Settings and usage of EHR
EHR may be used in most healthcare settings, including primary care,
secondary care, tertiary care and increasingly remote and home care. In
practice, however, there are wide variations between settings and sectors in
terms of EHR availability and usage. In the UK, for example, computerisation
is almost universal in primary care, while progress in secondary care has
lagged behind and for this reason has been a strategic priority for the NHS IT
implementation programme. In contrast, the US healthcare IT strategy is
geared towards changing the historically low rates of EHR use in primary
care.65
The purposes for which EHR are used also tend to vary by user group, which
is reflected in the different functionalities adopted. Box 6.4 provides examples
different EHR components commonly used by healthcare staff (doctors,
nurses, radiologists, pharmacists, laboratory technicians and radiographers),
and by patients and their carers.12 As discussed in due course, other users
are more concerned with the ‘secondary uses’ of aggregated or linked and
anonymised data for health surveillance, planning, audit or research.
181
Box 6.4: Components of EHRs used by different clinical stakeholders
Nurses: daily charting, medication administration, physical assessment, admission
nursing notes and nursing care plans.
Doctors: referral, present complaint, e.g. symptoms, past medical history, life style,
physical examination, diagnoses, tests, procedures, treatment, medication and
discharge.
Pharmacists: medication.
Secretarial staff: procedures, problems, diagnoses, findings and immunisation
records.
Multi-professionals (nurse, doctor, laboratory staff, radiology staff, clerk or
administrative staff, pharmacy personnel, healthcare professionals): referral,
present complaint, e.g. symptoms, past medical history, lifestyle, physical
examination, diagnoses, tests, procedures, treatment, medication, discharge,
administration of medication, admission nursing note and daily charting.
Patients: history, diaries and test results.
Parents/carer: history.
Source: Adapted from Häyrinen K, et al, 200812, with permission
Patient information can be stored in different formats in different systems.
Systems that require data to be coded using agreed taxonomic schemes prior
to entry can facilitate Health Information Exchange and interoperability. Box
6.5 illustrates the main international terminologies applicable to different data
types. Data may also be entered into EHR as free text. This may be important
for describing complex cases or issues for patient handover, however
‘secondary uses’ of this type of data are far more difficult to realise. Future
developments in natural language processing may help to bridge this gap.
Box 6.5: Examples of international terminologies used for coding data in EHRs
Data component
International terminology
Diagnoses
International Classification of Diseases (ICD)
Read codes
International Classification of Primary Care (ICPC)
Oxford Medical Information Systems (OXMIS)
International Classification of Primary Care (ICPC)
182
International Classification of Health Problems in
Primary Care (ICHPPC)
Procedures
Current Procedural Terminology (CPT)
Medication
Anatomical Therapeutic Chemical Classification Index
(ATC)
Pathological findings
Systematized Nomenclature of Medicine (SNOMED)
Nursing problems
North American Nursing Diagnoses (NANDA)
International Classification of Nursing Practice (ICNP)
Nursing interventions
Iowa Nursing Intervention Classification (NIC)
International Classification of Nursing Practice (ICNP)
Nursing outcomes
Iowa Nursing Outcome Classification (NOC)
Source: Adapted from Häyrinen K, et al, 200812 and Thiru K, et al, 200323; with
permission
The long-term vision for a fully integrated (cross-sectoral), longitudinal (“cradle
to grave”) patient record, supported by universal data and interoperability
standards and high-level communication and decision support technologies is
shared by England’s DH,59 in common with agencies such as the US Institute
of Medicine, although this is still some way from being realised.60 The current
DH commitment to paving the path towards this “information revolution” is
summarised in Box 6.6 below.
183
Box 6.6: England’s Department of Health has committed to: “move
•
away from information belonging to the system, to information enabling patients
and service users to be in clear control of their care;
•
away from patients and service users merely receiving care, to patients and
service users being active participants in their care;
•
away from information based on administrative and technical needs, to
information which is based on the patient or service user consultation and on
good clinical and professional practice;
•
away from top-down information collection, to a focus on meeting the needs of
individuals and local communities;
•
away from a culture in which information has been held close and recorded in
forms that are difficult to compare, to one characterised by openness,
transparency and comparability;
•
away from the Government being the main provider of information about the
quality of services to a range of organisations being able to offer service
information to a variety of audiences; and
•
in relation to digital technologies, away from an approach where we expect every
organisation to use the same system, to one where we connect and join up
systems.”
Source: Department of Health, Liberating the NHS: An Information Revolution,
executive summary;59reproduced with permission
As already noted, electronic record systems are well-embedded in some parts
of the NHS, particularly in general practice. Almost all general practices in the
UK are now computerised, with many practices now ‘paperless’ or ‘paper-lite’.
Such systems are available from many vendors and although there is some
variation in functionality between these, all primary care systems generally
cover seven broad areas (see Box 6.7).
184
Box 6.7: Classes of data captured within EHR systems in UK primary care
•
Administrative and demographic information, such as name, age, date of birth,
sex, NHS number, address with post code, telephone number, and increasingly
ethnicity and end-of-life care preferences.
•
Information on prevention of disease and screening, such as immunisations,
cervical screening, smoking status and alcohol intake.
•
Allergies and adverse reactions.
•
Free text entered during consultations.
•
Clinical and diagnostic information in the form of READ and OXMIS coded data.
•
Measurements such as height, weight, Body Mass Index (BMI), and blood
pressure.
•
Issuing of prescriptions.
•
Results of laboratory and radiological investigations.
•
Image files of documents scanned after being received from external agencies
such as hospitals and social services.
Although most primary care data in the UK are held by practice-based
systems, certain patient information may also be extracted and stored in
larger databases, such as the NHS ‘Spine’ on which the Summary Care
Record is stored, or in research-focused repositories such as the General
Practice Research Database (GPRD), QRESEARCH and the Doctors
Independent Network (DIN) database in the UK, and the MediPlus database
in Germany, etc. In addition to general practice computing systems, detailed
clinical information (Detailed Care Records) may also be held in specialist
clinical databases and disease registers based in secondary and tertiary care,
with more basic information being held in hospital patient administration
systems (PAS). Such records have historically been maintained separately,
due a combination of incompatible systems (which the interoperability
addenda is seeking to address), different data formats and taxonomies
(addressed by the coding standards strategy) and the absence of a single,
reliable means of recording and sharing current patient demographic
information.9 The latter is being addressed through the introduction of the
NHS Number—a unique patient identifier, which will be applied to all
healthcare transactions in order to facilitate data linkage and thus integrated
185
care records. Scotland has had this facility for some time, enabling the
production of a national integrated care record for diabetes.
6.2.3 Scope of EHR
At the heart of the EHR are individual patient records–often referred to as
EPR. These vary on multiple dimensions, including level of detail (from
summary to detailed care records), data source (single- or multiple-provider)
and timeframe (e.g. episodic or longitudinal). As well as patient histories and
details of recent care (which may include free text and diagnostic codes)
these records may incorporate digital images and scanned documents. The
broader EHR also includes non-medical data relevant to healthcare
administration and/or planning.
Increasingly, EHR are incorporated within complex systems that integrate a
range of other functions for supporting communication, decision-making and
task management, in addition to documentation (e.g. order entry, clinical
decision support, clinical messaging, results reporting, scheduling and
referrals). In some countries functions for billing patients or insurers are a key
part of EHR systems. Although this is not the case in the NHS, relevant
functionality addresses financial requirements around local commissioning of
services and national incentive schemes linking payments to preventive
activities and evidence-based care processes.
In the UK, single-provider EPR are still the most common format of clinical
records, these being generated, for example, in general practice and hospital
clinics . However, the advent of the NHS Care Records Service (NHS CRS)61
as part of England’s National Programme for Information Technology (NPfIT)
(see Chapters 3 and 18) promises increasingly greater integration through, for
example, the SCR.33
These EHRs can also aid patients more directly, by making information about
their health much more readily accessible, thus allowing them to become
more active in the provision of their own healthcare and the improvement of
their health.62
186
The NHS in England envisages a patient-centric, integrated, record, designed
to improve access by health and social care professionals where and when
they are needed, provided they have legitimate relationships. As noted
previously, the English strategy for EHR also involves giving individuals
secure access to a summary of their own health record via the Internet, using
the HealthSpace personal health organiser application, akin to the Personal
Health Record concept illustrated earlier in this chapter. HealthSpace is set
within a complex architecture of centralised (summary) and localised
(detailed) care records, as described in more detail in Chapter 3.33
47
By May
2010, 1.5 million SCR records were created, out of 29.8 million people (73%
of population aged >16
years) who were informed of their SCR creation
unless they opted out, in 113 of 152 primary care trusts (PCTs) in England.63
In contrast, in the US, Kaiser Permanente has reported much greater usage
of its customisable portal “My Health Manager”, which allows members to
access part of their records, including medications, past health visits, main
diagnoses, allergies and immunisations and also facilitates the sending of
email communication with their doctors.64 65
As already noted, while the EHR lies at the heart of IT implementation plans in
healthcare systems around the world, EHR implementation varies across
countries depending on priority areas and the expected benefits, identified by
the respective national governments.66 While some countries have focused on
primary care (e.g. Denmark, New Zealand and Spain), others have
concentrated on secondary care (e.g. England and China) and others on
emergency care (e.g. Scotland and The Netherlands).61 National approaches
for healthcare information exchange also vary across countries, whereby
system standardisation is promoted in England but in other countries systems
interoperability standards are encouraged e.g. Canada, Hong Kong and the
US.
Box 6.8 summarises examples of international EHR strategies
considered in a House of Commons Report on Electronic Health Records.67
187
Box 6.8: Examples of countries embarking on important EHR projects
•
Canada: A Private Lifetime Record for each citizen is being created and is being
co-ordinated by Canada Health Infoway: “By 2010, every province and territory
and the populations they serve will benefit from new health information systems
that will help transform their health care delivery system. Further, by 2010, the
electronic health records of 50 per cent of Canadians and by 2016, those of 100
per cent of Canadians, will be available to their authorized health care
practitioner.”68
•
France: A Dossier Médicale Personnel69 was passed in June 2004, which will
include a collection of health information to be viewed online. Patients will have
their own access and will legally own their record. Over a million Dossier
Médicale Personnel are created and internet based (beta national version) Rollout to citizens aged over 16 years and covered by the State Health Insurance
System will take place 2011. The record will include diagnoses and treatment
details provided by public and private health professionals. Clinical images are to
be added in future.
•
United States: Several integrated electronic records systems already exist and
in 2004, President Bush established that EHRs would be available for “most” US
citizens by 2014. President Obama’s administration provided $19 billion as a
stimulus fund to hospitals and healthcare facilities to digitise patient records and
for the “meaningful use” of IT.
Source: Adapted from House of Commons Report (2007) 67; permitted under open
license agreement for non-commercial research
The primary use of EHR is to support direct patient care through the effective
management and timely provision of information about individual patients’
health history, conditions and treatments. EHR that integrate clinical decision
support also facilitate evidence-based practice, while integrated and
interoperable EHR systems support HIE and interdisciplinary care. Primary
uses also include the support of administrative and financial processes, as
well as uses by patients for self-management.
In addition to supporting healthcare delivery, electronic data and databases
derived from EHR have many secondary uses; for example they can improve
the administration, management and quality control of healthcare through
facilitating clinical audit and the monitoring of service use; contribute to public
health by identifying inequalities and trends; support patient safety through
pharmaco-vigilance and disease surveillance; aid population-based research
into the impacts of new health technologies or policies and provide a source
of data for scientific studies of disease aetiology, genetic influences or drug
effects (see also Chapter 8).70 71 72 73 74
188
6.3 Theoretical benefits and risks of EHR
6.3.1 Theorised benefits
The substantial international investment in EHR and related eHealth
applications is predicated on the assumption that the adoption of accurate and
readily accessible electronic records will improve the quality, safety and
efficiency of healthcare (Table 6.1).
Table 6.1: Key theoretical benefits of electronic health records
Attribute
Benefit
Immediate and universal access to Increased efficiency (e.g. reduced time
the patient record
spent pulling charts, and duplicate
history-taking etc).
Increased quality
(better information at the point of care)
Easier and quicker navigation More efficient point of care assessment
through the patient record
and data abstraction
Increased
legibility
and Better quality information to aid clinical
comprehensiveness,
through decision-making and shared care; fewer
computer-aided
history
taking errors in patient management (e.g. missystems and better formatting (e.g. prescribing)
templates)
Secure record keeping
No lost records, fewer unnecessary waits
or missed appointments, aiding informed
patient care. Patient satisfaction.
Standardisation of care among Through better recording and sharing of
providers within the organisation
information and linkage to computerised
decision support systems (CDSS)
Reduction
of
paperwork, Removes
duplication,
reduces
documentation errors, filing activities
processing time, decreases personnel
costs
Coding efficiency and accuracy
Improved data quality
Alerts for medication errors, drug Safer patient care
interactions, patient allergies
Ability to electronically transmit Fewer delays, more efficient and
information
to
other
providers integrated patient care.
Enhanced
(assessments, history, treatments patient satisfaction.
ordered, prescriptions, etc.)
Availability of non-clinical data
Easier
management
of
costs,
performance and workflow
Availability of data for research
With downstream benefits for patient
care
Source: Healthcare Information and Management Systems Society (HIMSS) 2003;75
reproduced with permission
The fundamental premise is that clinicians require comprehensive and
accurate data on patients at the point of care if they are to provide high quality
health services. Electronic health records can aid this objective by dispensing
189
with the need to use difficult to access, and often illegible, paper-based
records. The improved record keeping, legibility and access afforded by EHR
can help to avoid the inconvenience of lost records, data recapture and reentry and mitigate the risk of recording and prescribing errors, thus improving
the appropriateness and safety of patient management. By facilitating the
secure exchange and sharing of patient information, an EHR can support
shared and continuous care both within single healthcare episodes (e.g.
hospital admission and supported discharge) and over time (as in the
management of long-term illnesses). An EHR with integrated decision support
can help providers improve the quality of clinical decision-making, guideline
compliance, aid preventive care and potentially reduce utilisation of services
and costs of care (see also Chapters 8 and 10).76 77 78
The success of EHR can be assessed by the following criteria, as proposed
by Hayrinen et al,12 using the framework originally developed by DeLone and
MacLean.79
“System quality assesses the information processing system itself, and its
attributes ... include ease of use, ease of learning or usefulness of system.
Information quality measures both the output and input of the information
system; attributes…here include completeness, accuracy, legibility, reliability
and format.
Information use measures end-users’ consumption of the output of an
information system, with attributes…including amount of use and number of
queries.
User satisfaction measures the end-users’ response to the use of the output
of an information system, and attributes … include overall satisfaction and
decision-making satisfaction.
Individual impact measures the effect of information on the behaviour of the
end-user, and attributes …include improved individual productivity and
information understanding.
Organizational impact measures the effect of information on organizational
performance, and its attributes….include return on investment and increased
work volume...”12
190
6.3.2 Theoretical risks of EHR
It is important to recognise that the introduction of an EHR may also have
unintended consequences, some of them negative. As illustrated in Chapter
4, the EHR and other eHealth applications have the potential to introduce
error and risk at many levels, including error associated with the user (e.g. via
inappropriate system operation and interpretation of outputs), the technology
(e.g. poor interface usability leading to missed readings or alerts; system or
network unreliability leading to missed or delayed exchanges; cyber-attacks),
and the underlying knowledge bases (e.g. faulty decision support algorithms).
One of the major sources of risk is related to the quality of the data held within
an EHR. As previously indicated, EHR lie at the heart of eHealth systems,
with related functionality such as ePrescribing, decision support, and
computerised order entry, utilising the data in the EHR. The correctness and
completeness of the data stored in the EHR is thus a key factor in ensuring
that it will be used effectively. There are currently large variations in the
accuracy and completeness of information recorded in EHR by clinicians and
associated administrative staff. For example, in the UK, disease prevalence
and specialist referral rates, based on information derived from an EHR, vary
widely between general practices. The scale of the variation is so large that a
significant proportion is likely to be due to differences in the accuracy of the
data recorded.80 A second issue with data quality arises with the information
stored in ‘legacy’ paper records. Unless this information is summarised and
entered onto the EHR system, the EHR will not give an accurate and
complete record of a patient’s medical history. Patients’ access to their EHR
has the potential to mitigate some of these risks by enabling individuals to
identify errors in their record. In the UK record access systems which allows
patients to view their GP record online, NHS HealthSpace, which provides a
portal to the SCR and disease-centred applications offer potentially important
benefits in this direction.
As with the use of computers in general, the use of EHRs in the context of the
clinical consultation also as potential to influence patient perceptions of care
quality and the clinician-patient relationship,
191
6.4 Empirically demonstrated benefits and risks of EHR: a summary of
the evidence
We have identified and appraised 21 reports of 20 systematic reviews (SRs)
addressing the associations between EHR and healthcare quality and safety
(see Appendices 4 and 5) and these comprise the knowledge base for this
section,11-23 27 81-86 of which Clamp 2007 is a second publication,14 in a journal,
of the same piece of work that had been published as a report in 2005.87
These SRs cover a broad range of outcomes including prevalence of adverse
drug events, compliance with guidelines, consultation time, documentation
time of nurses, doctor-patient relationship, quality of morbidity coding in
primary care, data quality, exchange of information, costs, process change
and a synthesis of all the disparate findings. They also include a SR of
barriers perceived by doctors to adopting EHR, which can be intervened with
IT.11 16
6.4.1 Impact on clinically-relevant outcomes
Clinical endpoints
Relatively few SRs have looked at the impact of EHRs on clinical endpoints,
reflecting the paucity of primary studies which have done so. Delpierre et al.’
included studies of patient outcomes in their 2004 SR of the impact of
computerised patient record systems on medical care quality.17 They found
that, overall, EHR use did not improve measures of cardiovascular risk or
clinical outcomes of acute respiratory distress syndrome, although one study
included in their review reported a significant decrease in diastolic blood
pressure,17 When outcomes were assessed by patients themselves, no clear
benefit was found in relation to the management of asthma, angina or
depression. It is worth noting, however, that all of the relevant studies had
included the EPR as part of a more complex intervention involving clinical
decision support and it is therefore not possible to attribute the effects to the
EHR alone.
Given the paucity of SRs evaluating the impact of EHRs on morbidity or
mortality, any claims about their clinical effectiveness need to be viewed with
caution. While a more detailed review of primary studies is warranted to shed
192
further light on this important question, there are likely to be few studies which
have robustly evaluated these impacts in light of the distance in time between
when EHR implementation takes place and the emergence of clinically or
statistically significant changes in measures of hypothesised clinical
endpoints. In other words, the size and length of study required to
convincingly demonstrate impacts of EHR on mortality and, to a lesser extent
morbidity, are far larger than a typical academic research grant would support.
Instead it is likely that the most convincing evidence of this kind will be
revealed if large-scale implementations are set up with evaluation as a key
objective, so as to enable longer-term changes to be captured. Most of the
evidence captured by the systematic reviews on which this report draws is
therefore concerned with intermediate or surrogate outcomes associated with
care quality or processes or user perceptions of benefit. These are
summarised next.
Surrogate endpoints: preventive care
A comprehensive SR by Chaudhry et al. revealed benefits of EHRs in the
form of increased adherence to guidelines or protocol-based care, especially
in the domain of preventive health.13 The effectiveness of EHRs was
demonstrated in terms of improved monitoring and surveillance, reduction of
medication errors and decreased utilisation of inappropriate or potentially
redundant care. However the authors cautioned against generalising from
these results, as the majority of studies demonstrating such benefits had
taken place in centres of excellence using ‘home-grown’ systems that had
been extensively customised to local needs and practices.13 In contrast, the
majority of institutions procure commercial systems with considerably less
opportunity for local tailoring.
A SR by Jerrant et al., focusing mainly on primary care studies published
between 1996-1999, found evidence that using EHR in the context of
screening interventions improved compliance with certain tasks by both
patients and providers.18
193
Delpeirre et al, in a SR mainly based in secondary care, found EHR to be
associated with increased pneumococcal vaccination (from 1% to 36%),
influenza vaccination (from 1% to 51%), prophylactic heparin (from 19% to
32%) and aspirin prescription (from 28% to 36%).17
6.4.1 Impact on process indicators
Satisfaction
Shachak and Reis’ SR on the effects of EHR on doctor-patient communication
draws primarily on data from primary care.19 The authors identified four key
themes in the literature, including the use of a computer for EHR by the doctor
during consultation; the impact of EHR use on doctor-patient communication;
factors affecting doctor-patient communication in EHR settings; and
behavioural styles of doctors that affect their interaction with patients.19 In the
studies covered, computers were used by doctors to look up patients’ general
medical records, to check patients’ medications, to retrieve test results, to
enter information and write prescriptions and letters, and at times to
demonstrate to patients changes in clinical data over time and to provide
printouts to inform patients. Compared with paper-based systems, patients in
clinics with EHR received more communication with respect to their
medication, in both written and oral format, but received less information on
medication indication, which led to more questions to doctors about
medicines.
Electronic health records have been shown to affect patient’s interaction with
their doctor. In the studies reviewed by Boonstra et al., doctors spent on
average 25-50% of the consultation time looking at the computer screen.16
Computers have also in some cases been perceived by doctors as a physical
barrier to effective doctor-patient interaction.17 Different approaches to
managing data during the consultation were also reported; for example some
doctors entered data during the consultation, others did so after the patient
had left and some did not type into the computer at all, but recorded notes on
dictating machines which were then transcribed into the record by clerical
support staff. It was found that some doctors sometimes used their computer
to send non-verbal cues to patients, by looking up at the screen, holding the
194
mouse or typing, to change the topic or flow of conversation and to indicate an
end to the meeting. Doctors who looked less at the screen and more at the
patient, were more prone to ask psychosocial questions and to respond
emotionally to their patients.19 However some doctors using computers had
improved verbal and non-verbal communication with patients, and could
furthermore manage visits and data better, despite having more to do in terms
of computer use and patient management.19
The layout of the doctor’s clinic can also have an impact on the doctor’s use
of EHR. For example, the SR by Shachak et al. found that when computers
were fixed in one place, doctors had to turn their body or move their chair and
thus eye contact with patients was hampered.19 It was in some cases possible
to obviate some of these adverse effects by repositioning the screen or using
flat monitors on mobile arms. According to Shachak et al, once successfully
implemented, patients in primary care report being satisfied with the corollary
gains associated with EHR use such as improved exchange of information
between healthcare professionals;19
Others have noted indifferent patient responses to EHRs.86 A SR of
predominantly secondary care-based studies found that patients were usually
satisfied and unlike doctors, did not perceive the computer to be a physical
barrier.17 Whilst interesting, these data should be interpreted with caution, as
satisfaction levels can be difficult to compare across studies conducted in a
range of settings and studying an array of computing devices; furthermore,
the fact that patients in general tend to rate doctors highly is an important
potential confounding variable.86
Overall, the evidence suggests that doctors tend to have a mixed response to
EHR use, but generally patients are satisfied. It is difficult from this body of
evidence to assess the extent to which any clinician sense of dissatisfaction is
temporary, reflecting the problem of poor early integration that is widely
recognised in studies assessing health informatics innovations in the early
post-implementation
phase.
Improved
training
in
consultation
skills,
particularly in relation to the use of computers in an inclusive way, and more
195
attention to screen sizes and room layout could help to obviate some of the
adverse effects associated with use of EHR.
Impact on professionals’ time
In a SR that synthesised the findings from 23 studies, predominantly from
secondary care, Poissant et al. found that documentation time was decreased
for nurses, but not doctors. This was thought to be related to differences in the
types of information the two groups record and their different approaches to
data recording. Importantly these differences did not attenuate with increased
or more established use.15 Furthermore, the time taken for doctors to retrieve
information was found to increase.15
A more recent SR by Thompson et al, but this time focusing on nursing
efficiency, found a more mixed picture in relation to nurses’ documentation
time.20 It appeared that any overall reductions in documentation time were
negated by increased patient specific documentation time.20
Somewhat unexpectedly, studies that have looked at the impact of EHR on
process times shortly after system implementation, have tended to find a
reduction in documentation time, whilst studies with a longer time period
between implementation and evaluation have tended to find the reverse.15
Whilst the reasons for this are not entirely clear, this could reflect the
introduction of greater functionality of the EHR system over time, and possibly
users’ greater willingness to engage more meaningfully with the more
integrated record systems that develop over time.
Overall, it seems that time savings are by no means guaranteed and it may
lead to disappointment if such a promise is used as a means of encouraging
EHR implementation.15 The impact on time may also vary between
professional groups, with evidence suggesting that the time trade-offs may be
particularly unfavourable for doctors.
196
6.4.3 Impact on data quality
A high standard of data quality is essential to deliver the promises made
relating to EHRs. The most commonly used measures of validity found in the
SR by Chan et al. were correctness and completeness, which were defined
thus:
“Data completeness is defined as the level of missing data for a data
element. Data accuracy is conceptualized as the extent to which data
captured though the EHR system accurately reflects an underlying
state of interest (e.g., whether a medication list accurately reflects the
number, dose, and specific drugs a patient is currently taking). This
includes the timeliness or “currency” … of the data as well as the
clinical specificity or “granularity” in how individual clinicians or clinical
staff document within an EHR system. A high level of missing data
and/or data errors will substantially reduce measure reliability and
validity.”83
In a review of a substantial body of international studies (n=89) covering the
period 1982 to 2004 and conducted in mixed settings, mostly in secondary
and tertiary care, Hayrinen et al found that the information systems studied
helped clinicians to record information more completely.12 Although
documentation varied across record fields, these EHR systems enabled more
detailed recording. Documentation was judged to be accurate, even when
data were entered by patients. Although EHR also enabled more
comprehensive recording, there were still shortcomings when records were
judged against professional regulations and guidelines.12 Overall, their SR
concluded that EHR can improve the comprehensiveness and accuracy of
data recording, but in some studies the overall quality of data recording
remained sub-optimal.
As discussed above, time trade-offs may also be
encountered with use of EHR systems resulting in an increase in recording
time.
Thiru et al, systematically reviewed studies examining the scope and quality
of data recording in primary care, published between 1980 and 2001 (31 out
of 52 were from the UK). This revealed that data quality has been measured
197
in different ways, most commonly using comparisons of rates derived from the
EHR with an external standard.23 Data validity was examined using a range of
terms
(e.g.
completeness,
correctness,
accuracy,
consistency,
and
appropriateness), but the precise terms were often undefined in many
studies.23 The most commonly used measure of validity was sensitivity or the
proportion of total cases of a disease or events such as prescriptions that
were recorded by the EHR system (i.e. completeness of recording).
Prescriptions had the highest rate of recording (sensitivity), probably because
prescribing is a core function of many EHR systems and is used even in
circumstances when EHRs are not used for other functions such as morbidity
recording.23 Positive predictive values were generally high for all recorded
elements.
As in most studies the component of EHR (numerator) and the component of
the reference standard (denominator) were not clearly stated and, where
these were specified, there were inconsistencies between studies leading the
authors to recommend that in future data quality studies the numerator,
denominator and confidence intervals should be clearly stated.23 Thiru et al
further stated that as verifying data would not be possible with disappearance
of paper notes in EHR practices, the triangulation method (EHR data, clinical
diagnoses and/or notes) could help to ascertain the true quality of data. This
SR also suggested that terms used to measure data quality, such as accuracy
and completeness, should be clearly defined to minimise the risk of ambiguity.
This issue has also been highlighted in a SR by Jordan et al, where data
quality in EHRs in UK primary care was assessed from studies published up
to 2002, using the following four criteria:
(1) completeness–if there is a
morbidity code every time the patient visited the GP; (2) correctness–if
appropriate codes were given; (3) completeness of a morbidity register–if
relevant patients are included in the register; and (4) correctness of the
morbidity register–if patients on that register should be there. While Thiru et
al. also included non-UK studies using any gold standard for data quality,
Jordan et al. looked at UK primary care data quality studies only referring to
specific gold standards.22 As with the findings from Thiru et al’s SR, recording
198
of consultations was generally high (typically greater than 90 per cent) in the
24 studies, but assigning a morbidity code during each consultation was more
variable (66-99% complete). That the results from Jordan et al. are similar to
those reported by Thiru et al. is not altogether unsurprising given that these
SRs largely reviewed the same evidence-base.
Jordan et al. however also pointed to the observation that quality of recording
was higher for conditions with clear diagnostic features such as diabetes
when compared to conditions with possibly more subjective diagnostic criteria
such as asthma.22 Coronary heart disease was the most commonly assessed
condition recorded in disease registers in previous studies and completeness
of recording was generally moderate (typically around 70%).22 The positive
predictive value of coronary heart disease registers was generally high
(typically around 83–100 %).22 Other diseases that were examined in the 24
studies (such as asthma and epilepsy) showed similar patterns of
completeness of recording and positive predictive value of recorded
diagnoses, but rates were generally lower than for coronary heart disease.22
Herrett et al. examined the methods used to validate data and the quality of
the validations between 1987-2008 in the GPRD, which is one of the largest
primary care databases in the world. They found 212 publications that
reported 357 validations to verify 183 distinct diagnoses. Eighty-five percent
(n=303) of validations were done outwith GPRD, by sending questionnaires to
GPs and requesting anonymous patient records from GPs or by comparing
rates of disease with a non-GPRD UK-based data source. The remaining
validations were done using GPRD by comparing diagnoses with codes,
reviewing manually the free texts on anonymised records or by sensitivity
analysis. These validation exercises revealed that in most studies a high
proportion of cases were confirmed as having the diagnoses, with an overall
median of 89% (range 24-100%). The estimates of incidence and prevalence
of diseases in GPRD were comparable to other UK population-based sources,
except in a few cases; for example, the incidence rate of rheumatoid arthritis
from GPRD was 50% higher than previous estimates88 while prevalence
estimates of musculoskeletal diseases in GPRD were lower in comparison to
199
other GP databases.89 This could be due to presence of clearer case
identification in the former than in the latter, as Jordan postulated. There were
however a number of potential limitations to these attempts at validating
recorded data, these including the small sizes of the samples used for
validation, missing data because of transfer and/or death, and in some cases
difficulties with ascribing codes to some diagnoses.
Completeness and correctness of coding has also been shown to be
dependant on the level of enthusiasm of the GPs and practices and GP’s
preference for certain codes.22 Like Thiru et al., Jordan et al. noted the lack of
well-defined data quality standards and the need to correct this if better
measurement of data quality in primary care EHRs was to be established.
An older SR by Hogan and Wagner, covering the period 1978 to 1996, mostly
reviewing studies undertaken in US secondary care (in contrast to the Thiru et
al., Jordan et al. and Herrett et al. SRs, which were primary care-focused),
drew similar conclusions and noted the wide variability in the accuracy of
EHRs and the lack of standard measures of data accuracy, as was also
reported by Hayrinen et al. Errors in EHR data did not just concern data entry;
rather, the use of EHRs at various points in the care pathway made it prone to
multiple errors and it was thus insufficient to implement a single intervention in
an attempt to improve data accuracy.27
Hoerbst and Ammenwerth published a systematic review of quality
requirements for EHR in 2010, based on both published literature and expert
interviews. They considered that both functional and non-functional
requirements of EHRs could potentially create problems before, during or
after EHR implementation.84 A repository of 1,191 possible requirements were
assigned to 59 categories and sub categories.84 Functional categories were
regarded as having the highest quality requirements, followed by data security
and content (22.8%, 22.7%, 19.0%, 12.6% respectively), whilst non-functional
categories such as portability, performance, maintainability, etc, were
regarded as having the lowest requirements.84
200
Overall, the evidence in relation to data quality considerations indicates that
with EHR use there tends to be greater recording of information, but the
accuracy of what is recorded can vary quite considerably. This partly depends
on the ease with which it is possible to detect and classify diseases but also
on doctors’ preferences for using certain codes. Organisations such as
EuroRec in Europe and the Certification Commission for Health Information
Technology (CCHIT) in the USA are working towards establishing EHR quality
certification, to promote and support data quality improvement.
6.4.4 Economic impacts
Successful EHR implementations require considerable capital investment and
can result in major changes to workflows if these are to be adopted. Boonstra
et al. found that worries about the costs of adopting and maintaining EHRs
can be a major barrier to their use.16 The SR by Chaudhry et al. on US studies
between 1995 to 2005, found little information on basic cost or Return on
Investment, and estimating cost-effectiveness remains largely reliant on
statistical modelling, which is not usually available beyond the research
arena.13 Despite reviewing a comprehensive body of studies, the authors
concluded that there was insufficient evidence available to guide stakeholders
on the financial effects of moving to EHR systems.
In a recent update to this influential report, which covered the period 20042007, the authors reported that although patient focussed applications are
rapidly growing in the healthcare service, little economic evaluation has been
undertaken.90 They suggested that greater partnership between public-private
organisations, policies on financial incentives and a robust evidence base for
EHR implementation are all important prerequisites to speed up the adoption
process.
In an earlier SR published in 2007, Clamp and Keen examined the value of
EHR reported in studies between 1998 and 2005. They also found limited
clear evidence on the economic outcomes of EHR systems.14 They urged
caution when interpreting pronouncements in relation to the health economic
benefits of EHR. Even the studies that they regarded as being of reasonable
201
quality, and had presented cost data, had not included health economists in
the research team. There was technically no sound evidence about cost
changes associated with the EHR, apart from a paper by Bryan et al. on
PACS.91 Most papers failed to include all reasonable costs, irrespective of
whether they would tend to increase or decrease overall cost changes
associated with EHR use.14 Most papers made linear projections of costsavings in future years, typically up to five years ahead.14 However, there was
often no explicit basis for these projections.14 There were no studies which
actually sought to capture all of the costs and benefits associated with an
EHR at the level of a process or within a single hospital setting.14 Although
there is some suggestive evidence that information networks lead to greater
economies of scale and scope in non-health care fields and interoperability
can enhance EHR efficiency, no good studies were found in this SR that
reported on this aspect.14
In contrast, the 2008 SR by Uslu et al, which reviewed mainly US studies
between 1991 to 2003, concluded that there were economic benefits from
reductions in time spent in administrative
work, some savings
in
documentation costs or savings in nursing costs or savings in non-personnel
costs.85 Bearing in mind the various caveats stated above, particularly in
relation to the difficulties in ascribing costs, the claims made in this somewhat
uncritical review however need to be interpreted with caution.
With recent policy developments, more indirect costs and economic benefits
also need to be considered. For example, in UK primary care, financial
incentives, driven by the Quality and Outcomes Framework (QOF), have
encouraged the recording of clinical processes and intermediate outcomes
using EHR. While this is acknowledged to have improved certain measures of
quality, there has been some evidence of the use of gaming strategies to
maximise payments and attention has been paid to methods of ameliorating
the problem of ‘perverse incentives’.92 Of more direct relevance to EHR is the
introduction of the “Meaningful Use” criteria in the US, developed as part of
the Health Information Technology for Economic and Clinical Health Act
(HITECH). Under this scheme, stimulus funds are being employed to
202
incentivise the ‘Meaningful Use’ of ‘certified EHRs.93 This releases payments
to Medicare and Medicaid providers for capturing electronically coded clinical
data that will “enable tracking of key clinical conditions, communicating that
information for care coordination purposes, and initiating the reporting of
clinical quality measures and public health information.”94
While the
Meaningful Use incentives are expected to have many benefits, the US
Department of Health and Human Services has taken the precaution of
commissioning an expert panel to consider their possible unintended
consequences (negative and positive) and methods of addressing these, in
the context of the HITECH act.95
Overall, there remain very few primary studies and SRs that have formally
studied the cost-effectiveness of EHR. The AHRQ, and its update, remains
the most comprehensive attempt to study this question. The authors found
that the current evidence-base has emerged from a limited number of
benchmark institutions that have evaluated “home-grown” EHR systems,
which were developed over a period of time iteratively to suit the purpose.
This limits the generalisability of the results, as many institutions do not have
the means to develop their own EHR and so will be implementing commercial
systems. Even when studying these benchmark institutions, it has not been
possible to demonstrate clear return on investments. The AHRQ has the
following caveats as in Box 6.9.
203
Box 6.9: Important caveats outlined by the Agency of Health Related Quality
•
All studies are predictive analyses that are based on many analytical
assumptions and limited empirical data. Hence, the strength of evidence is often
weak.
•
In all studies, the EHR system was assumed to have multiple functionalities and
the functional capability of the EHR system is critical to the benefit that may occur
from using it.
•
The organisations that were the subjects of four studies were all large. The
literature review did not identify cost benefit studies for EHR implementation in
small organisations.
•
The costs of implementing an EHR system may be underestimated. Only one of
the five cost benefit analyses included the cost of the implementation process,
and found that this cost was 1.5 times the cost of the EHR system.
•
Implementing an EHR system requires extensive changes in the organisational
processes, individual behaviours, and the interactions between the two. These
resulting costs are often omitted or not reported from studies but can be
substantial.
•
The financial benefits depend on the financing system.
As shown in the
sensitivity analysis of one study, the benefit estimates are most sensitive to the
assumption of the proportion of capitated patients.
Realising all quantifiable
benefits of EHR implementation would require changes to the current healthcare
financing system.
•
Both the cost and the benefit of attaining interoperability among EHR systems are
directly proportional to the level of data exchange achieved.
For example,
although the cost of achieving machine-organisable or machine-interpretable
interoperability is greatest, it offers the most potential for increased efficiency,
improved healthcare utilisation, and reduced costs.
Adapted from: AHRQ (2006)78 (permission to reproduce applied for)
6.4.5 Human and organisational factors affecting the adoption of EHR
Realising the potential efficiency and quality gains of EHR is dependent on
their effective implementation and adoption, which can prove challenging (see
also Chapters 17 and 18).13
16 96
A SR by Boonstra and colleagues, mainly
covering studies conducted in secondary care settings, identified eight
barriers, namely: 1) financial, 2) technical, 3) time, 4) psychological, 5) social,
204
6) legal, 7) organisational type and 8) change process, of which 1-3 were
more often identified than the rest.16 Financial, technical and time barriers
were classified as primary barriers while the remainder were categorised as
secondary barriers.16 According to the authors, organisational and change
process barriers are the “mediating factors in the success of EMR project”, as
the former determine the importance of other barriers (1 to 6 above) before
EHR adoption, while the latter can mediate the effect of the other (1 to 6
above) barriers during the EHR implementation process.16
The technical barriers faced by doctors were further explored in a SR by
Castillo et al, with some suggested interventions.11 Technical barriers were
typically a lack of data standards that permit effective, accurate and timely
exchange of electronic clinical data between healthcare providers.11 The six
enablers of EHR adoption identified from the studies reviewed were: user
attitude towards information systems, workflow impact, interoperability
(compatibility), technical support, communication among users and expert
support.11 These factors were influential irrespective of practice/clinic size,
type of setting (inpatient or outpatient), primary or secondary care, or national
policy.11 The authors suggest that if EHR usage was logged and interpreted
with reference to the above factors, tailored assistance could be provided; for
example by providing access to frequently asked questions, notification
messages,
electronic
bulletins,
electronic
workflow
and
process
documentation, documents on standards, support and knowledge directories,
social networking, etc.11
Strategies for addressing such barriers, suggested by Boonstra et al in 2010,
are summarised in Table 6.2.
Other barrier-reduction strategies include the certification and standardisation
of applications and systems to permit exchange of electronic clinical data;
removal of legal barriers; and greater security of the medical data stored in
EHRs.11
205
Table 6.2 Perceived barriers to EHR adoption and related interventions
Perceived barriers
A Finance
B Technical
Possible barrier-related intervention strategies
•
•
•
•
•
•
•
•
C Time
•
•
E Social
•
•
•
•
•
•
F Legal
•
•
G Organization
H Change process
•
•
•
•
•
•
•
•
Provide documentation on return on investment.
Show
profitable
examples
from
other
EMR
implementations.
Provide financial compensation.
Educate doctors and support ongoing training.
Adapt the system to existing practices.
Implement EMR on a module-by-module basis.
Link EMR with existing systems.
Promote and communicate reliability and availability of
the system.
Acquire third party for support during implementation.
Provide support during implementation phase to convert
records and assist
Provide training sessions to familiarize users.
Implement a user friendly help function and help desk.
Redesign workflow to achieve a time gain
Discuss advantages and disadvantages for doctors and
patients.
Information and support from doctors who are already
users.
Ensure support, leadership, and communication from
management.
Develop requirements on safety and security in
cooperation with doctors and patients.
Ensure EMR system meets these requirements before
implementation.
Communicate on safety and security of issues.
Redesign workflow to realize a better organizational fit.
Adapt EMR to organization type.
Adapt EMR to type of medical practice
Select a project champion, preferably an experienced
doctor.
Let doctors (or representatives) participate during the
implementation process.
Communicate the advantages for doctors. Use
incentives.
Ensure support, leadership, and communication from
management.
Source: Boonstra and Broekhuis, 2010 16; reproduced under open third party use
agreement
206
6.4.6 Empirically-demonstrated risks of EHR
As noted earlier, different types of human error in the use or interpretation of
EHR may generate patient safety risks. These may be associated with
“...cognitive overload, loss of overview, errors in data entry and retrieval,
excessive trust in electronically-held data, and the tendency to conflate data
entry with communication within and between care teams”.21 We did not find
a systematic review specifically examining the outcomes of such factors for
patient safety. To a large extent, the theoretical risks of EHR to patient safety
remain unsubstantiated. However one area of risk has commanded greater
attention; namely privacy breaches and their implications for patients.
EHR contain sensitive patient information and carry risks to patient privacy
due to unauthorised or inappropriate access. Such privacy breaches have
potential to harm patients in a number of ways, including the embarrassment
of
having sensitive medical secrets disclosed, financial implications
associated with access by interested third parties such as insurers or
employers, or even health impacts, for example if the record were to be
falsified and inappropriate treatments administered. . Although there have
been few reported cases of patient harm due to data breaches, high profile
cases highlighted in the media and the availability of enforceable sanctions,
including dismissal, have raised concerns amongst some clinicians about their
legal liability. Indeed it has been suggested that some doctors are more
concerned about patient privacy and confidentiality than their patients are.16
Concerns over information security and confidentiality are prominent amongst
reasons for poor engagement with the EHR (see Chapter 4), as has been
evidenced by the strong reaction from the British Medical Association (BMA)
and others in NHS England. Although we did not identify any systematic
reviews specifically concerned with the impacts of privacy breaches on
patients, the topic was explored in a recent SR of HIE in primary care
practices by Fontaine et al .97 The authors report that 76% of Americans are
concerned about the privacy of their health information, particularly in the
unauthorised sharing of ‘super confidential’ information about mental health,
chemical dependency and genetics, medical identify theft and fear of
207
discrimination based on health related conditions. They also note the
historical resistance of providers to relinquishing patient information to a
centralised database and its role in the failure of community health information
networks. Whilst acknowledging that federated models of HIE–where records
remain with the provider and are shared only when needed–offer advantages
for privacy and security, they also note that the greater transparency offered
by HIE can make providers reluctant to use EHR for fear of negative exposure
(e.g. of non-compliance with guidelines or poor patient outcomes).
Although establishing EHR systems is a key priority for NHS England, prior
evaluations have been of variable quality and have shown limited benefits in
terms of increased efficiency, patient safety, and quality of care, and little
evidence of improved patient outcomes. Large, methodologically robust
studies which include sound economic evaluations and well-theorised
qualitative studies are needed to fill this evidence-gap.
The evidence base on EHR is muddied by the variable use of EHR-related
terms and concepts as well as differences in the nature of the interventions
described in published research, many of which are largely driven by the
clinical decision support functionality that EHR can offer. This complexity can
make it difficult to generate theory about the mechanisms through which the
intervention will have influence and the types of outcome that are significant,
or to back-track from effect to cause when interpreting the results of
evaluation studies. As noted earlier, a further source of difficulty is the
timescale between intervention and anticipated outcome, which may be too
large to demonstrate clinical or cost-effectiveness in the course of a typical
research project, although intermediate outcomes provide useful indicators for
projection. Likewise longer-term studies may help to establish whether the
early benefits noted in some studies are maintained once initial enthusiasm
has worn off or where usage of the system has evolved to incorporate
additional tasks to those originally envisaged. Integrating a theoreticallyinformed evaluative research agenda into large-scale projects holds the
greatest promise for revealing conclusive evidence of longer-term effects and
208
will help to address the so-called ‘evaluation paradox’ that can act as a barrier
to progress.4 It also provides greater opportunities to embed contextualised
qualitative and observational studies that may help to advance our
understanding of the complex and dynamic interactions between users,
stakeholders and technologies,21 and of how organisational factors (e.g.
workflow management) can influence the realisation of benefits.
In considering the value of evidence, it is also worth reflecting on the role of
EHR as a marker of a modern healthcare system. As with other aspects of the
digital society, the adoption of EHR is inevitable and necessary if we are to
move forward to a more connected, knowledge-driven and flexible healthcare
environment and there are many associated benefits for providers,
organisations, governments and citizens that will never be captured by
simplistic Return on Investment calculations or clinical impact studies. While
robust scientific evidence is highly desirable and may provide insights into the
types of EHR functionalities that may be most useful for meeting particular
objectives in particular groups or settings, the limitations of the current
evidence-base should not be regarded as an overwhelming barrier to
progress, provided that we can operate under the remit of ‘do no harm, but
modernise’. Given the current financial climate, however, and the vast cost of
implementing such systems across health systems, the need to justify
investment will nevertheless remain a priority.
We agree with the authors of the AHRQ review in advocating more research
that evaluates commercially-procured, or ‘off-the-shelf’ systems, in order to
balance the evidence derived from studies of home-grown systems in centres
of excellence, which may be atypical. Whilst independent academic studies
may be valuable, a requirement for evaluation within vendors’ business plans
may be one way of ensuring the routine capture of relevant metrics. We also
agree that there may be value in establishing standard reporting criteria for
4
This refers to the fact that it can be hard to motivate the adoption of systems
without being able to cite evidence of their impacts, but it can also be hard to
generate this evidence without the systems first being adopted.
209
studies of EHR implementation to aid interpretation of results across
contexts.78
A recurring theme in the literature is the importance of clinicians’ attitudes as
a mediator of effective implementation and benefits realisation, and thus the
value of clinician engagement at a strategic and operational level (see
Chapter 17). This message has clearly been taken into account within the
NHS IT strategy, but it remains worthy of repetition.
As discussed in Chapter 17, there is some evidence (mostly from the US)
that clinicians from larger medical practices have higher rates of adoption of
EHR when compared to smaller practices, possibly reflecting the greater
technical support and training facilities available in institutions with larger IT
budgets.16 Similar observations have been made in relation to larger hospitals
versus smaller specialist care settings16 There are also indications that market
forces may contribute to rates of EHR adoption in different settings. For
example Abdolrasulnia et al attributed the greater adoption of EHR in urban,
as compared to rural ambulatory care practices, to inter-provider competition
and the perceived advantages of EHR for relative quality improvement.98 They
suggest that policy makers direct financial incentives to areas where providers
face fewer competitive pressures to adopt EHR.98 Although such factors will
be less influential in nationally-funded health systems like the UK, where
primary care computing is widespread, market forces nevertheless play a role
in shaping GPs choice of off-the-shelf systems through the perceived
functionality or usability of alternative vendor solutions.99
The organisational and workflow changes required to move from paper-based
records to EHR are often substantial and support from the organisation, in the
form of leadership, encouragement, incentives or training opportunities, can
all facilitate EHR adoption; conversely a lack of these can hinder the change
process.
The introduction of EHR systems may offer indirect benefits for the NHS,
including a reduction in clinical errors and improved patient safety, through the
210
associated introduction of CDSS and ePrescribing systems, but further
evidence of these impacts is required. EHR systems may also make
information more readily accessible by patients, thus enabling them to
become more involved in the management of their own healthcare, and this is
again an area that warrants further study.
The ‘secondary uses’ of EHR data for clinical audit, health surveillance and
research also offer considerable potential to improve the quality and safety of
care
through
highlighting
compliance
with
evidence-based
practice
recommendations, facilitating population-based evaluations of health policies
or interventions, supporting the monitoring of health conditions and
inequalities, maintaining pharmacovigilance and enabling medical innovation.
For example, electronic patient records have been used in recent seminal
research employing a range of methods including randomised controlled
trials, observational, cohort, and case-control studies; surveys; and qualitative
research.70-72 100-107 The impact and economic value of such ‘secondary uses’
is difficult to quantify, but in the case of research leading to biomedical
innovation the benefits to the scientific community and the national economy
are likely to transcend those for the health system. In addition to ensuring that
requests to use EHR for ‘secondary uses’ are appropriate, it will also be
important to ensure that access to data for such purposes is minimally
clinically disruptive..
Linked to this is the need for further studies to establish the appropriate
balance between providing secure EHR to healthcare providers, consumers,
primary and secondary users of EHR (see also the safety map in Chapter 4)
and the risks of breaches of confidential information. A range of authorisation
mechanisms, including role-based access, incorporating cryptographic and
biometric techniques are being considered.108 Research to evaluate the
impacts and outcomes of these and other privacy-enhancing solutions is
warranted. Value would also be gained from further work to link legislative and
regulatory frameworks, governing primary and secondary uses of electronic
patient records, with evidence on public attitudes towards data sharing,
particularly in light of the opportunities and risks presented by cross-sectoral
211
data linkage and market drivers for third party (including) commercial uses. It
will also be important to ensure that conventions and procedures for data
sharing are built around ethical principles relating to trust, equity and benefits
sharing.21 83
Key areas for further work include the development and evaluation of data
quality standards for use in EHR, and the evaluation of methods for improving
data quality. Accurate and complete data is an essential prerequisite for
effective EHR use and this should therefore be seen as a priority for NHS
England. For EHR-based healthcare quality assessment, data correctness,
completeness, granularity and comparability have to be taken into
consideration.83
Further research would be worthwhile to help build an understanding of the
contexts in which EHR are used, and to examine which aspects of EHR are
perceived favourably or unfavourably by healthcare professionals and why.
This should include studies of contextual factors affecting the implementation
of large-scale EHR initiatives and rigorous, longitudinal, qualitative studies to
examine how EHR is assimilated into routine practice over time.21
109
Understanding how EHR are used to support HIE between within
organisations is also important.21 Research exploring the soft and hard
impacts of “home-grown” versus “off the shelf” systems and examining under
what conditions small EPR systems are fit for purpose, would also be useful.13
21
Additional work to derive methods of incorporating patients’ experiences of
healthcare and patient reported outcomes into EHR would also be valuable
both in terms of supporting clinician-patient partnership and generating new
types of data for clinical or secondary uses.
As already noted, there is also a real need for more rigorous longitudinal realtime evaluations of national-scale EHR implementations of the kind that are
now either underway or being planned in many parts of the world.61
212
Policymakers worldwide will be interested to observe the outcomes of the
HITECH act in the US; particularly the impact of the ‘Meaningful Use’
incentives in promoting adoption and best practice in EHR and in realising
predicted benefits for care quality and patient outcomes. This raises the more
general need for international-level research on EHR adoption, use and
impacts, bearing in mind national differences in healthcare systems,
population health needs and technological infrastructure, and the demand for
electronic health data and services in a globalised society.
213
References
1.
World
Health
Organization,
Global
eHealth
Observatory.
http://www.who.int/goe/publications/goe_atlas_2010.pdf;
accessed
26
January 2011.
2. Wikipedia. http://en.wikipedia.org/wiki/Electronic_health_record; last accessed 16
January 2011.
3. International Organization for Standards. Health informatics — Electronic health
record — Definition, scope, and context. ISO publication 2005;TR 20514.
4. Department of Health, NHS Executive. Information for health: an information
strategy
for
the
modern
NHS
1998-2005;
http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPo
licyAndGuidance/DH_4007832.
5. US National Institute of Standards and Technology, Information Technology
Laboratory.
Electronic
Health
Record;
http://www.itl.nist.gov/div897/docs/EHR.html; accessed 26 January 2011.
2006.
6. European Commission, Information Society and Media. The conceptual framework
of interoperable electronic health record
http://www.ehrand
ePrescribing
systems.
impact.eu/downloads/documents/EHRI_D1_2_Conceptual_framework_v1_0.
pdf 2008.
7. World Health Organization. Electronic Health Records: Manual for Developing
Countries;
http://whqlibdoc.who.int/wpro/2006/9290612177_eng.pdf;
accessed 26 January 2011. 2006.
8. Healthcare Information and Management Systems Society. Electronic Health
Record; http://www.himss.org/ASP/topics_ehr.asp; accessed 27 January
2011.
9. US Office of Communications. EMR vs EHR – What is the Difference?
http://healthit.hhs.gov/blog/onc/index.php/2011/01/04/electronic-medicalrecord-vs-electronic-health-record-clarifying-the-ehremrdifference/#comments Accessed 25 March 2011. 2011.
10. ACS/Xerox. Electronic Health Records and Health Information Exchange: The Rx
for
an
Ailing
Healthcare
System
http://www.acsinc.com/wp_electronic_health_records_and_health_information_exchange.as
px Accessed 25 March 2011. 2009.
11. Castillo VH, Martinez-Garcia AI, Pulido JR. A knowledge-based taxonomy of
critical factors for adopting electronic health record systems by physicians: a
systematic literature review. BMC Med Inform Decis Mak 2010;10(60):60.
12. Häyrinen K, Saranto K, Nykänen P. Definition, structure, content, use and
impacts of electronic health records: A review of the research literature.
International Journal of Medical Informatics 2008;77(5):291-304.
13. Chaudhry B, Wang J, Wu S, Maglione M, Mojica W, Roth E, et al. Systematic
Review: Impact of Health Information Technology on Quality, Efficiency, and
Costs of Medical Care. Annals of Internal Medicine 2006;144(10):742-752.
14. Clamp S, Keen J. Electronic health records: is the evidence base any use? Med
Inform Internet Med 2007;32(1):5-10.
15. Poissant L, Pereira J, Tamblyn R, Kawasumi Y. The impact of electronic health
records on time efficiency of physicians and nurses: a systematic review. J
Am Med Inform Assoc 2005;12(5):505-16.
16. Boonstra A, Broekhuis M. Barriers to the acceptance of electronic medical
records by physicians from systematic review to taxonomy and interventions.
BMC Health Services Research 2010;10:231.
214
17. Delpierre C, Cuzin L, Fillaux J, Alvarez M, Massip P, Lang T. A systematic review
of computer-based patient record systems and quality of care: more
randomized clinical trials or a broader approach? International Journal for
Quality in Health Care 2004;16(5):407-416.
18. Jerant AF, Hill DB. Does the use of electronic medical records improve surrogate
patient outcomes in outpatient settings? J Fam Pract 2000;49(4):349-57.
19. Shachak A, Reis S. The impact of electronic medical records on patient-doctor
communication during consultation: a narrative literature review. J Eval Clin
Pract 2009;15(4):641-9.
20. Thompson D, Johnston P, Spurr C. The impact of electronic medical records on
nursing efficiency. J Nurs Adm 2009;39(10):444-51.
21. Greenhalgh T, Potts HW, Wong G, Bark P, Swinglehurst D. Tensions and
paradoxes in electronic patient record research: a systematic literature review
using the meta-narrative method. Milbank Q 2009;87(4):729-88.
22. Jordan K, Porcheret M, Croft P. Quality of morbidity coding in general practice
computerized medical records: a systematic review. Family Practice
2004;21(4):396-412.
23. Thiru K, Hassey A, Sullivan F. Systematic review of scope and quality of
electronic patient record data in primary care. BMJ 2003;326(7398):1070.
24. Gray J, Majeed A, Kerry S, Rowlands G. Identifying patients with ischaemic heart
disease in general practice: cross sectional study of paper and computerised
medical records. BMJ 2000;321(7275):1519.
25. Lowrie R, Morrison J, McConnachie A. A cluster randomised controlled trial of
pharmacist led Statin Outreach Support (SOS) in primary care: Design and
baseline characteristics. Contemporary Clinical Trials 2010;31(4):303-311.
26. Bleich HL, Slack WV. Reflections on electronic medical records: When doctors
will use them and when they will not. International Journal of Medical
Informatics 2010;79(1):1-4.
27. Hogan WR, Wagner MM. Accuracy of data in computer-based patient records. J
Am Med Inform Assoc. 1997;4(5):342-55.
28. Quaas J, Lanigan M, Newman D, McOsker J, Babayev R, Mason C. Utility of the
digital rectal examination in the evaluation of undifferentiated abdominal pain.
The American Journal of Emergency Medicine 2009;27(9):1125-1129.
29. Lehmann CU, Kim GR, Johnson KB, Sable C, Reyna M, Holbrook PR.
Telemedicine Applications in Pediatrics. Pediatric Informatics: Springer New
York, 2009a:279-292.
30. Gackstetter GD, Hooper TI, Al Qahtani MS, Smith TC, Memish ZA, Schlangen
KM, et al. Assessing the potential health impact of the 1991 Gulf War on
Saudi Arabian National Guard Soldiers. International Journal of Epidemiology
2005;34(4):801-808.
31. Department of Health and Human ServicesHealth Records and Office of the
Assistant Secretary for Planning and Evaluation, Task Force on Privacy,
Social
Needs
and
Personal
Privacy.
http://aspe.hhs.gov/pic/reports/ahrq/4441.pdf, accessed 26 January 2011.
1993.
32. US General Accounting Office, Report to the Chairman, Commiittee on
Governmental Affairs, US Senate. http://archive.gao.gov/t2pbat8/143217.pdf.
1991.
33. Greenhalgh T, Wood GW, Bratan T, Stramer K, Hinder S. Patients' attitudes to
the summary care record and HealthSpace: qualitative study. Bmj
2008;336(7656):1290-5.
34. Coiera E. Do we need a national electronic summary care record? Med J Aust
2011;194(2):90-2.
35. Morris LM, Cameron J, Brown C, Wyatt JC. Sharing summary care records.
Results from Scottish emergency care summary. Bmj 2010;341(10):c4305.
215
36.
Cresswell KS, Aziz. The NHS Care Record Service (NHS CRS):
recommendations from the literature on successful implementation and
adoption. Informatics in Primary Care 2009;17(3):153-160.
37.
Wikipedia.
Continuity
of
Care
Record;
http://en.wikipedia.org/wiki/Continuity_of_Care_Record, accessed 26 January
2011. 2011.
38. Lehmann CU, Kim GR, Johnson KB, Kim GR, Lehmann CU. Electronic Health
Records and Interoperability for Pediatric Care. Pediatric Informatics:
Springer New York, 2009b:257-264.
39. Gill J, Chearman T, Carey M, Nijjer S, Cross F. Presenting patient data in the
electronic care record: the role of timelines. JRSM Short Reports 2010;1(4).
40. Scottish Government. Partnership for Care: Scotland's Health White Paper;
http://www.scotland.gov.uk/Publications/2003/02/16476/18737. 2011.
41. Satoguina J, Walther B, Drakeley C, Nwakanma D, Oriero EC, Correa S, et al.
Comparison of surveillance methods applied to a situation of low malaria
prevalence at rural sites in The Gambia and Guinea Bissau. Malar J
2009;8(274):274.
42. Ross DA, Hinman AR, Saarlas KN, Lloyd-Puryear MA, Downs SJ. The NearTerm Future for Child Health Information Systems. Journal of Public Health
Management and Practice 2004;10:S99-S104.
43. Höerbst A, Ammenwerth E. Quality and Certification of Electronic Health
Records. ACI. 2010 May;1(2):149-164. Applied Clinical Informatics
2010;1(2):149-164.
44. Lee D, Lau F, Quan H. A method for encoding clinical datasets with SNOMED
CT. BMC Medical Informatics and Decision Making 2010;10(1):53.
45. Halamka JD, Mandl KD, Tang PC. Early Experiences with Personal Health
Records. Journal of the American Medical Informatics Association
2008;15(1):1-7.
46. Gregory KE, Radovinsky L. Research strategies that result in optimal data
collection from the patient medical record. Applied Nursing Research 2010.
47. Greenhalgh T HS, Stramer K, Bratan T, Russell J. Adoption, non-adoption, and
abandonment of a personal electronic health record: case study of
HealthSpace. BMJ 2010b;341:c5814.
48. Pagliari C, Detmer D, Singleton P. Potential of electronic personal health records.
BMJ 2007b;335(7615):330-333.
49. National Health and Hospitals Reform Commission, Australia. Person-controlled
Electronic
Health
Records;
http://www.health.gov.au/internet/nhhrc/publishing.nsf/Content/mediaRelease
300409, accessed 26 January 2011. 2009.
50. Rowlands S, Callen J, Westbrook J. What information do general practitioners
need to care for patients with lung cancer? A survey of general practitioners
perceptions. Health Information Management Journal 2010;39(1):8-16.
51. Cooper C, Steinbuch M, Stevenson R, Miday R, Watts N. The epidemiology of
osteonecrosis: findings from the GPRD and THIN databases in the UK.
Osteoporosis International 2009;21(4):569-577.
52. Austin T, Kalra D, Lea NC, Patterson DL, Ingram D. Analysis of Clinical Record
Data for Anticoagulation Management within an EHR System. Open Med
Inform J 2009;3:54-64.
53. Department of Health. Delivering 21st Century IT support for the NHS: National
specification for Integrated Care Records Service - consultation draft;
http://webarchive.nationalarchives.gov.uk/+/www.dh.gov.uk/en/Publicationsan
dstatistics/Publications/PublicationsPolicyAndGuidance/DH_4108952. 2002.
54. Groen PJ, Goldstein D, Jaime N. Personal Health Record (PHR) Systems: An
Evolving
Challenge
to
EHR
Systems.
http://www.hoise.com/vmw/07/articles/vmw/LV-VM-08-07-26.html 2007.
216
55. National Committee on Vital and Health Statistics, U.S. Department of Health and
Human Services. http://www.ncvhs.hhs.gov/0602nhiirpt.pdf; accessed 25
March 2011. 2006.
56. Ross SE, Lin CT. The effects of promoting patient access to medical records: a
review. J Am Med Inform Assoc 2003;10(2):129-38.
57. Fisher B, Bhavnani V, Winfield M. How patients use access to their full health
records: a qualitative study of patients in general practice. J R Soc Med
2009;102(12):539-44.
58. Canada Health Infoway. Blueprint -an interoperable EHR framework. 2006.
59. Department of Health. Liberating the NHS:An Information Revolution; executive
summary
http://www.dh.gov.uk/prod_consum_dh/groups/dh_digitalassets/@dh/@en/document
s/digitalasset/dh_120601.pdf. 2010.
60. Institute of Medicine. To Err Is Human: Building a Safer Health System National
Academy Press, Washington, DC 1999.
61. Robertson A, Cresswell K, Takian A, Petrakaki D, Crowe S, Cornford T, et al.
Implementation and adoption of nationwide electronic health records in
secondary care in England: qualitative analysis of interim results from a
prospective national evaluation. BMJ 2010;341:c4564.
62. Pagliari C, Detmer D, Singleton P. Electronic Personal Health Records,
Emergence
and
Implications
for
UK,
http://www.nuffieldtrust.org.uk/ecomm/files/Elec%20Personal%20Records%2
0II.pdf. The Nuffield Trust 2007a.
63. Greenhalgh T SK, Bratan T, Byrne E, Russell J, Potts HW.Greenhalgh T,
Stramer K, Bratan T, Byrne E, Russell J, Potts HW. Adoption and nonadoption of a shared electronic summary record in England: a mixed-method
case study. BMJ 2010a;341.
64. Silvestre AL, Sue VM, Allen JY. If you build it, will they come? The Kaiser
Permanente model of online health care. Health Aff (Millwood)
2009;28(2):334-44.
65. Chen C, Garrido T, Chock D, Okawa G, Liang L. The Kaiser Permanente
Electronic Health Record: Transforming And Streamlining Modalities Of Care.
Health Affairs 2009;28(2):323-333.
66. Morrison Z, Robertson A, Cresswell K, Crowe Sa, Sheikh A. Understanding
contrasting approaches to nationwide implementations of electronic health
record systems: England, the USA and Australia. Journal of Healthcare
Engineering 2010.
67. House of Commons Health Committee, House of Commons London. The
Electronic Patient Record Sixth Report of
Session 2006-07
http://www.publications.parliament.uk/pa/cm200607/cmselect/cmhealth/422/4
22.pdf. The Stationery Office Limited 2006-07;1(HC 422-II).
68. Canada Health Infoway, Electronic Health Information and Communication
Technology Projects, March 2010 Performance Evaluation. Final Report,
http://www2.infowayinforoute.ca/Documents/Performance%20Evaluation%20Report_English.pdf,
accessed 26 January 2011.
69.
Dossier
Médicale
Personnel,
France.
http://www.ehealtheurope.net/Features/item.cfm?docId=194, accessed 26
January 2011.
70. Simpson CR, Sheikh A. Trends in the epidemiology of asthma in England: a
national study of 333,294 patients. J R Soc Med 2010;103(3):98-106.
71. Gray L, Batty GD, Craig P, Stewart C, Whyte B, Finlayson A, et al. Cohort Profile:
The Scottish Health Surveys Cohort: linkage of study participants to routinely
collected records for mortality, hospital discharge, cancer and offspring birth
217
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
characteristics in three nationwide studies. International Journal of
Epidemiology 2010;39(2):345-350.
Ford I, Murray H, Packard CJ, Shepherd J, Macfarlane PW, Cobbe SM. LongTerm Follow-up of the West of Scotland Coronary Prevention Study. New
England Journal of Medicine 2007;357(15):1477-1486.
Gnani S, Majeed A. A user's guide to data collected in primary care in England;
http://www1.ic.ac.uk/resources/579D8B09-C1C1-4026-A7BEC3E936EE9567/, accessed 26 January 2011. Eastern Region Public Health
Observatory, Cambridge 2005.
Centres for Disease Control and Prevention, Atlanta, USA. Assessing the
potential of national strategies for electronic health records for population
health
monitoring
and
research;
http://www.cdc.gov/nchs/data/series/sr_02/sr02_143.pdf;
accessed
26
January 2011. 2006.
Healthcare Information and Management Systems Society. HIMSS Electronic
Health
Record,
Definitional
Model,
Version
1.0;
http://www.providersedge.com/ehdocs/ehr_articles/HIMSS_EMR_Definition_
Model_v1-0.pdf; accessed 9th March 2011. 2003.
Evans RS, Pestotnik SL, Classen DC, Clemmer TP, Weaver LK, Orme JF, et al.
A Computer-Assisted Management Program for Antibiotics and Other
Antiinfective Agents. New England Journal of Medicine 1998;338(4):232-238.
Jones DL, Kroenke K, Landry FJ, Tomich DJ, Ferrel RJ. Cost savings using a
stepped-care prescribing protocol for nonsteroidal anti-inflammatory drugs.
JAMA 1996;275(12):926-30.
AHRQ. Costs and Benefits of Health Information Technology, Evidence
Report/Technology Assessment Number 132. 2006;6(E006).
DeLone WH, McLean ER. Information Systems Success: The Quest for the
Dependent Variable. INFORMATION SYSTEMS RESEARCH 1992;3(1):6095.
Majeed A. Sources, uses, strengths and limitations of data collected in primary
care in England. Health Stat Q 2004;21:5-14.
Herrett E, Thomas SL, Schoonen WM, Smeeth L, Hall AJ. Validation and validity
of diagnoses in the General Practice Research Database: a systematic
review. Br J Clin Pharmacol 2010;69(1):4-14.
Jamal A, McKenzie K, Clark M. The impact of health information technology on
the quality of medical and health care: a systematic review. Him J
2009;38(3):26-37.
Chan KS, Fowles JB, Weiner JP. Review: Electronic Health Records and the
Reliability and Validity of Quality Measures: A Review of the Literature.
Medical Care Research and Review 2010;67(5):503-527.
Hoerbst A, Ammenwerth E. Electronic health records. A systematic review on
quality requirements. Methods Inf Med 2010;49(4):320-36.
Uslu AM, Stausberg J. Value of the electronic patient record: An analysis of the
literature. Journal of Biomedical Informatics 2008;41(4):675-682.
Irani JS, Middleton JL, Marfatia R, Omana ET, D'Amico F. The Use of Electronic
Health Records in the Exam Room and Patient Satisfaction: A Systematic
Review. J Am Board Fam Med 2009;22(5):553-562.
Clamp S, Keen J. The Value of Electronic Health Records: A Literature Review.
2005.
Watson DJ, Rhodes T, Guess HA. All-cause mortality and vascular events
among patients with rheumatoid arthritis, osteoarthritis, or no arthritis in the
UK General Practice Research Database. J Rheumatol 2003;30(6):1196-202.
Jordan K, Clarke AM, Symmons DP, Fleming D, Porcheret M, Kadam UT, et al.
Measuring disease prevalence: a comparison of musculoskeletal disease
218
using four general practice consultation databases. Br J Gen Pract
2007;57(534):7-14.
90. Goldzweig CL, Towfigh A, Maglione M, Shekelle PG. Costs And Benefits Of
Health Information Technology: New Trends From The Literature. Health
Affairs 2009;28(2):w282-w293.
91. Wang SJ, Middleton B, Prosser LA, Bardon CG, Spurr CD, Carchidi PJ, et al. A
cost-benefit analysis of electronic medical records in primary care. Am J Med
2003;114(5):397-403.
92. McDonald R, Cheraghi-Sohi S, Tickle M, Roland M, Doran T. The impact of
incentives on the behaviour and performance of primary health care
professionals. NHS Service Development and Organisation Programme.
http://www.sdo.nihr.ac.uk/files/project/158-final-report.pdf; accessed 25 March
2011. 2010(
).
93. Centers for Medicare & Medicaid Services. CMS finalizes definition of meaningful
use of certified electronic health records (EHR) technology;
http://www.cms.gov/apps/media/press/factsheet.asp?Counter=3794&intNum
PerPage=10&checkDate=&checkKey=&srchType=1&numDays=3500&srchO
pt=0&srchData=&keywordType=All&chkNewsType=6&intPage=&showAll=&p
Year=&year=&desc=&cboOrder=date; accessed 28th February 2011. 2010.
94. HITECH Law Blog. HHS Releases Proposed EHR “Meaningful Use” Standards
http://healthitlawblog.wordpress.com/2010/01/07/hhs-releases-proposed-ehrmeaningful-use-standards/; accessed 26 March 2011. 2010.
95. Department of Health and Human Services. Anticipating Unintended
Consequences of Health Information Technology and Health Information
Exchange. Presolicitation notice;
https://www.fbo.gov/index?s=opportunity&mode=form&id=22633354ce0c9b300b832f
f16c2658a7&tab=core&_cview=0; Accessed 25 March 2011. 2010.
96. Coiera E. Building a National Health IT System from the Middle Out. Journal of
the American Medical Informatics Association 2009;16(3):271-273.
97. Fontaine P, Ross SE, Zink T, LM. S. Systematic review of health information
exchange in primary care practices. The Journal of the American Board of
Family Medicine 2010;23(5):655-670.
98. Abdolrasulnia M, Menachemi N, Shewchuk RM, Ginter PM, Duncan WJ, Brooks
RG. Market effects on electronic health record adoption by physicians. Health
Care Manage Rev 2008;33(3):243-252.
99. NHS Connecting for Health. GP Systems of Choice (GPSoC)
http://www.connectingforhealth.nhs.uk/systemsandservices/gpsupport/gpsoc;
accessed 26 March 2011.
100. Buckley BS, Simpson CR, McLernon DJ, Murphy AW, Hannaford PC. Five year
prognosis in patients with angina identified in primary care: incident cohort
study. BMJ 2009;339.
101. Sanderson H, Adams T, Budden M, Hoare C. Lessons from the central
Hampshire electronic health record pilot project:evaluation of the electronic
health record for supporting patient care and secondary analysis. BMJ
2004;328(7444):875-878.
102. Black C, Kaye JA, Jick H. Relation of childhood gastrointestinal disorders to
autism: nested case-control study using data from the UK General Practice
Research Database. BMJ 2002;325(7361):419-421.
103. Smith AG, Fear NT, Law GR, Roman E. Representativeness of samples from
general practice lists in epidemiological studies: case-control study. BMJ
2004;328(7445):932.
104. Mackay D, Haw S, Ayres JG, Fischbacher C, Pell JP. Smoke-free Legislation
and Hospitalizations for Childhood Asthma. New England Journal of Medicine
2010;363(12):1139-1145.
219
105. Hippisley-Cox J, Coupland C. Unintended effects of statins in men and women
in England and Wales: population based cohort study using the QResearch
database. BMJ 2010;340:2197.
106. Helms PJ, Ekins Daukes S, Taylor MW, Simpson CR, McLay JS. Utility of
routinely acquired primary care data for paediatric disease epidemiology and
pharmacoepidemiology. Br J Clin Pharmacol 2005;59(6):684-90.
107. Gray J, Orr D, Majeed A. Use of Read codes in diabetes management in a
south London primary care group: implications for establishing disease
registers. BMJ 2003;326(7399):1130.
108. Win KT. A review of security of electronic health records. Him J 2005;34(1):138.
109. Robert G, Greenhalgh T, MacFarlane F, Peacock R. Adopting and assimilating
new non-pharmaceutical technologies into health care: a systematic review. J
Health Serv Res Policy 2010;15(4):243-50.
220
Chapter 7
Summary
•
Most computer-assisted history taking systems are designed for use
by professionals, often as templates incorporated into the electronic
health record; these are however increasingly becoming available
for completion by patients and carers, either on-site in healthcare
settings or remotely before or after the clinical assessment.
•
Computer-assisted history taking systems can be used in a range of
contexts, these include, primary prevention (e.g. population
screening programmes), to support the extraction of potentially
sensitive information (e.g. a detailed sexual or alcohol intake history)
and to support disease management of those living with long-term
conditions (e.g. collection of dietary intake information for diabetes
care).
•
Most of the applications are at present supported only by face
validity and modest or weak empirical evidence. The clearest
evidence of benefit at present relates to the elicitation of sensitive
information.
•
There is however inconclusive evidence that computer-assisted
history taking systems translate into time savings for professionals
or improved important patient outcomes.
•
Key areas in which to consider the wider use of computer-assisted
history taking systems include situations in which patients and/or
carers
might
be
apprehensive
about
disclosing
potentially
embarrassing information and situations in which a fixed repertoire
of questions need to be asked of large numbers of patients such as
in the context of screening programmes.
221
7.1 Introduction
Medical history taking lies at the centre of clinical diagnosis and decisionmaking. As described by Pringle over a decade ago, computer-assisted
history taking systems (CAHTS) are tools that aim to aid clinicians in
gathering data from patients to inform a diagnosis and/or formulating a
treatment plan.1 Of relevance is the increasing potential for patients to
complete aspects of their history, prior to or post the consultation, whilst onsite in surgeries and clinics, or remotely. Online systems can also help with
data collection for screening programmes, such as England’s national
vascular risk assessment and management strategy.2 Despite the many
possible applications and even though CAHTS have been available for nearly
three decades,3 these are however not very commonly used in routine clinical
practice.
7.2.1 Definition
The medical history refers to the information obtained by a clinician or other
healthcare professional by listening to the patient/carer’s history and then
asking questions which can provide further insight and clarification about the
condition. A history taking system is a tool that aids clinicians in gathering
such information from patients to help formulate the diagnosis or inform a
treatment plan. If computerised, these tools are referred to as CAHTS.
7.2.2 Description
Computer-assisted history taking systems can be used by healthcare
professionals, or directly by patients, as in the case of, for example, preconsultation interviews.4-6 They can be used remotely, for example via the
Internet, telephone or on-site. They draw on a range of technologies such as
personal computers (PCs), personal digital assistants (PDAs), mobile phones
and electronic kiosks;
data input can be mediated via, amongst others,
keyboards, touch screens and voice-recognition software (see Table 7.1).
222
Data Collection
Modes
On-site
Telephone
Administration
The
patient
Modes
present on-site with
Internet
is
a health practitioner The patient responds
who facilitates the to
process
completed
technological
with
a practitioner
who N/A
records
the
(audio responses
system
(audio
oral, presentation;
presentation;
patient
is The patient responds The patient is given
present on-site and to
an
automated access to an online
conducts the history telephone
by
oral,
keyed input)
keyed input)
The
telephone
interview by a health
history-taking
Professional-
a
himself/herself that
through
system survey to complete,
records
the which can then be
a responses for access accessed
Patient-
technological
completed
system,
by
such
laptop,
a
as practitioner
desktop presentation;
health health
a
practitioner
(audio and linked to other
keyed similar patient
computer, or PDA input)
records
(audio
audio
&
by
visual
presentation; keyed
(visual
&
presentation;
keyed input)
input)
Table 7.1 Modes of data collection for computer-assisted history taking
systems
7.2.3 Scope
As noted above, there are two main approaches to using these tools to
facilitate history taking, these being either for use by healthcare professionals
or where these involve obtaining data directly from patients.
223
Health professional completed
This is where health professionals input patient details into computers,
typically via templates; this can include remote telephone interview and onsite
interview. Examples include:
•
Emergency services: Where mobile systems are used in ambulatory
services and/or in accident and emergency departments. Sometimes
computer-assisted history taking and computer decision support
systems are used in combination. For example, nurses in the out-ofhours, emergency service type in patient responses to questions
generated by the computer programme and then direct patients to
answer further relevant questions leading ultimately to a diagnosis
and/or a management plan.7;8
•
General practice systems: Where CAHTS record aspects of the
patient’s history into a computerised template, which informs part of the
patient record.9
Patient/carer completed (self-administered)
This refers to instances in which the computer software prompts or enables
the patient to input relevant information themselves; this can include on-site
self-interview, remote, automated telephone self-interview and remote, online
self-interview. Examples of these include:
•
Pre-consultation questionnaire: This enables patients to complete
the history either on a computer in the waiting room or at home.4-6;10;11
This has, for example, been shown to help professionals rapidly to
appraise psychiatric referrals and prioritise patients.12;13
•
Electronic and on-line health records: Where the patient types their
own history into sections of their electronic health record (EHR), which
can then link to other similar patient-held records.
With the addition of diagnostic and reminder functionalities, CAHTS may
influence all stages of the patient care pathway before, during and after the
consultation. For example, with the addition of a diagnostic platform such as
224
probabilistic
advice
and
question-prompting,
CAHTS
may
become
instrumental in the decision-making process.
7.3Theoretical benefits and risks
7.3.1 Potential benefits
Saving professionals’ time on documentation
Professionals have limited time in consultations and, in a traditional face-toface clinical encounter, it is not always possible to obtain a complete or, in
some cases, even relevant medical history.14 In an evaluation of ambulatory
practices, clinicians were, for example, found to spend 20% of their day
writing.15;16 In an Ohio family practice, dictation and charting outside of
examination rooms occupied 56 minutes of an eight-hour working day.17 In an
antenatal clinic, it was estimated that two-thirds of the working day was spent
recording information.18 These studies demonstrate the potential importance
of CAHTS as a time-saving device for clinicians, particularly if these devices
can also sift and present relevant data in an easily accessible format.
Collection of more comprehensive and valid information
Clinicians need to remember many questions relating to the management of
each condition. Omitting an important question can have considerable
implications for diagnosis and treatment. For example, studies show that 50%
of psychosocial and psychiatric problems are missed
19
and that 54% of
patient problems and 45% of patient concerns are neither elicited by the
clinician nor disclosed by the patient.20 Furthermore, studies have
demonstrated fewer errors in PDA data records than in paper diaries and that
PDA datasets were correctly completed in 93% of cases21 while one study
reported patient compliance of 100%.22
Patient level impacts
Computer-assisted history taking systems can be used in several clinical
settings and are particularly useful in eliciting potentially sensitive information,
for example, on alcohol consumption,23; 24psychiatric care,25-28 sexual health29
and gynaecological health.30 Using CAHTS before the consultation would
225
also, in principle, allow more time for the patient to discuss their actual health
problem rather than routine aspects of medical history with their physician.
There are important potential benefits of CAHTS in the care of people with
special needs. For example, computers can allow questions to be asked in a
number of different languages. They can also provide multi-media forms for
patients who cannot read and write through making computers voice
questions and digitally record spoken responses.4; 31
7.3.2 Potential risks
A central limitation of CAHTS may be that not every question can be
meaningfully answered in a questionnaire; such issues are likely to be
particularly relevant when using general questionnaires which may include
questions irrelevant to individual patient concerns and needs.32 Computerassisted history taking systems are inherently limited with respect to nonverbal communication. Computers are currently unable to detect non-verbal
behaviour since a computer cannot, for example, sense a patient’s mood that
might easily be picked up in a consultation.4 Another related concern is that
computers may depersonalise the doctor-patient relationship, although this
has not actually been demonstrated in the clinical setting.6
7.4.1 Benefits
We found five completed published systematic reviews (SRs) on this subject.
32; 33; 34; 35; 36
We have furthermore initiated three additional related Cochrane
reviews in key areas.37-39 We synthesise this body of evidence together with
additional selected evidence to illustrate key findings in relation to CAHTSs.
Saving professionals’ time on documentation
The length of time associated with documenting information was noted almost
four decades ago and was one of the reasons that Mayne et al.40 initiated
their research into the scope for patient interviews being aided by computers.
Our review of the literature found that evidence on the effectiveness of
CAHTS on saving professionals’ time on documentation is non-conclusive.
226
A study comparing PDA use by patients for documenting acute pain
assessment time and completeness and comparing this to pen-and-paper
showed that PDA usage decreased the overall consultation time (such as
chart review, assessment and documentation) by 22 per cent.35 However, as
Poissant et al. point out, although over 90 patient encounters were considered
in this study, it was a single clinician study so the results may not be
generalisable to other settings.35 Further, Poissant et al. demonstrated that
the use of central station desktops for CAHTS is slightly less time-consuming
(with a weighted average reduction of 8 per cent) for clinicians, when
compared to bedside or point-of-care computer systems.35 Dale and Hagen
point out that PDA use for CAHTS benefit healthcare professionals as there is
a ‘reduction in the data management and processing time’.20-22; 41; 42 However,
barriers still exist as all five (out of a total of nine) of the studies in this SR that
reported data concerning feasibility of collecting data using PDAs identified
persistent ‘technical difficulties with the PDA technology’.20;
21
‘Power
problems’ and ‘PDA malfunction’ were the most common problems.
Collection of more comprehensive and valid information
Kerkenbush et al. found increased compliance with data entry45 and noted
that Invivodata™ (a company specialising in electronic diary technology)
showed that patients responded timely to 93% of all the electronic data
gathering prompts.46 Also, Kamarack et al. found 99% compliance with
assessments that needed to be completed every 45 minutes during waking
hours over a six-day period.45 Studies have demonstrated fewer errors in PDA
data records than in paper diaries and that PDA datasets were correctly
completed in 93% of cases22 while one study reported patient compliance of
100%.21
A recent randomsied controlled trial (RCT) by Richens et al46 has found that
the odds of disclosure for seven key sensitive behaviours were found to be
40% greater when computer-assisted interviewing was compared with penand-paper interviewing. De Jarlais47 reported that CAHTS enable substantially
more complete reporting of HIV risk behaviour than face-to-face interviewing.
Furthermore, if patients use a PDA for a CAHTS, they are less likely to falsify
227
data
when
compared
with
those
generated
by
pen-and-paper,
as
demonstrated, by four RCTs.21; 41; 42; 48
Similarly, Tiderman et al,49 found that women undertaking the CAHTS
interviewing reported a significantly higher median number of male partners
for the preceding 12 months than those undertaking face-to-face interviewing.
However the overall responses to sexual history related questions using
CAHTS and face-to-face interviewing were similar.
Patient level impacts
There is good evidence, that patients prefer to record aspects of their history
using PDA technology when compared with recording using pen-and-paper.21;
41; 42
Furthermore, if patients use a PDA for CAHTS, they are less likely to
falsify data when compared with those generated by pen-and-paper. 50
One RCT for example, reported that ‘young people find computerised
questionnaires equally or more acceptable than the usual clinical interview or
a written questionnaire’.23 In another study, parents originally assumed that an
interview using a computer was not as ‘friendly and personal’, but became
more optimistic after the interview was completed.25
7.4.2 Risks
We should not expect CAHTS to pose risks to patients’ wellbeing. However,
technical problems, lack of training and poorly integrated systems can result
in frustration for patients and professionals. A SR assessed the use of PDAbased templates for CAHTS by medical trainees.34 The review highlighted that
while most medical trainees who use handheld computers for patient history
taking appear comfortable and are generally satisfied with them, barriers still
exist, these including: 1) lack of technical experience; 2) a preference for
pen-and-paper; 3) difficulty handling the small device; and 4) concerns about
data loss and security.34 Also, what may work in some clinical settings (such
as an infertility, antenatal or allergy clinic) may not work as well in other areas
where patients more typically present with undifferentiated problems (such as
general practice or a general medical or gynaecology clinic, for example).33; 35
228
7.5.1 Implications for policy
The progressive change in focus from hospital care to community-based care
means that staff will become more mobile and therefore need to access and
input data at the point-of-care. There is also a drive within health systems
globally to focus on health promotion, disease prevention, and early detection
of disease. For example, in England, a vascular risk assessment programme
for people aged 40-74 is now starting; CAHTS can be helpful in collecting risk
factor and behavioural data for this.51;
52
Gathered histories can be stored
locally or in a user-friendly online space that can be accessed by patients
supporting tailored advice and self-management.6
7.5.2 Implications for practice
Several studies have found that self-administered computer-assisted
interviewing is perceived favourably by patients because computer systems
cannot be judgemental towards sensitive behavioural data such as sexual
history taking and domestic violence.53 Clinician and patient-operated CAHTS
data are potentially important additions to the EHR as they can help to
improve data quality through: data entry forms with data validation checks;
encoding of data; legibility; easier access to past records; attribution of
entries; easier availability; and facilitating patient checks of their own data.
Consideration also needs to be given to incorporating patient completed
diaries online, thereby allowing information on key complaints and selfgenerated data (for example, blood pressure or peak expiratory flow) to be
made available to clinicians before the actual consultation. This may result to
more complete history taking and more time available to spend on the actual
consultation.
However Greenhalgh et al.,54 found that patients perceived HealthSpace (a
personal electronic health record accessed via the Internet) as neither useful
nor easy to use and its functionality aligned poorly with their expectations and
self-management practices. The same authors conclude that ‘unless personal
EHRs align closely with people's attitudes, self management practices,
229
identified information needs, and the wider care package (including
organisational routines and incentive structures for clinicians), the risk that
they will be abandoned or not adopted at all is substantial. Conceptualising
such records dynamically (as components of a socio-technical network) rather
than statically (as containers for data) and employing user centred design
techniques might improve their chances of adoption and use’.
In general, the adoption of healthcare informatics is complex, multidimensional, and influenced by a variety of factors that relate to individuals
and organisations. A recent systematic review of mixed methods studies55
found that successful implementation strategies should include:
‘a)
involving
users
at
different
development
and
implementation phases, b) using project champions or other
key staff, c) providing adequate training and support, and d)
monitoring
system
use
in
the
early
stages
of
implementation’.55
7.5.3 Implications for research
Seminal reports on quality and safety of healthcare56 invariably recognise
information technology as one of the main vehicles for making radical
improvements in delivery of healthcare. Although CAHTS have been available
for around 30 years, successful use in routine healthcare is still variable.
A recent SR36 found that results of CAHTS effectiveness depend on the
sensitivity of the question, the tool used as well as the population assessed.
Certainly, any interpretation of results is based on two sets of assumptions.
Firstly, no self-report tool can be proven to be more accurate than another,
given that no gold standard exists with which to compare alternative methods.
Secondly, many assumptions are made about people’s social desires when
answering questions that may not always go in the direction expected. In
further developing methods for collecting patient histories multiple methods
assessing self-reported behavioural risks should be complemented with
appropriate biological markers as appropriate, depending on the type of
study.36
230
Important changes should take place in how we conceive, plan and conduct
primary and secondary research on CAHTS so that we provide the framework
for a comprehensive evaluation that will lead to an evidence base to inform
policy and practice. Findings of such comprehensive evaluation should also
inform a sustainability model that views CAHTS as an integral component and
working platform for EHRs (see Chapter 6).
Relatively few studies reported on safety outcomes57 when evaluating
CAHTS, whilst others sometimes failed to assess the most salient dimensions
of quality such as access, accessibility and equity.58 Although CAHTS are
frequently promoted as being “cost-saving”,
59-62
this attribute was rarely
evaluated rigorously.35 Most of the technologies are at present supported only
by face validity and modest or weak empirical evidence. Unless these
systems are adequately studied, they may not “mature” to the extent that is
needed to realise their full potential when deployed in everyday clinical
settings.58
Additional research on the reliability of different methods of data collection
would also be useful in assessing the value of collecting data using CAHTS in
healthcare settings.56 Moreover, even after successful interventions, accuracy
may not be maintained over time. Medical processes are complex and
changing, and data error, as well as procedural changes may occur due to the
high turnover of personnel. Hence, there is a pressing need for regular
evaluations of CAHTS, analogous to techniques used in continuous quality
improvement.35; 51; 63
References
1. Pringle M. Using computers to take patient histories. BMJ 1998; 297: 698-698.
2. Department of Health: Putting prevention first. Vascular checks: risk assessment
and
management.
http://www.dh.gov.uk/prod_consum_dh/groups/dh_digitalassets/@dh/
@en/
documents/digitalasset/dh_083823.pdf
3. Slack WV, Hicks GP, Reed CE, Van Cura LJ. A computer-based medical-history
system. New England Journal of Medicine 1966; 274: 194-198.
4. Healthspace. https://www.healthspace.nhs.uk/ (last accessed 20/09/2009). 2009.
231
5. RelayHealth.
https://www.relayhealth.com/rh/general/news/newsArchive/news54.aspx
(last
accessed 20/09/2009). 2009.
6. Bachman JW. The patient-computer interview: a neglected tool that can aid the
clinician. Mayo Clin Proc 2003; 78: 67-78.
7. Parkin A. Computers in clinical practice: applying experience from child
psychiatry. BMJ 2000; 321: 615-618.
8. Lilford RJ, Guthrie K, Kelly M. History-taking by computer. Baillieres Clin Obstet
Gynaecol 1990; 4: 723-742.
9. NHS Direct in England. Report by the controller and auditor general. HC 505
Session 2001-2002: 25 January 2002. 25-1-2002.
10. http://www.nhsdirect.nhs.uk/ (last accessed 20/09/2009). 2007.
11. Millman A, Lee N, Brooke A. ABC of medical computing. Computers in general
practice--I. BMJ 1995; 311: 800-802.
12. Jones RB, Atkinson JM, Coia DA, Paterson L, Morton AR, McKenna K et al.
Randomised trial of personalised computer based information for patients with
schizophrenia. BMJ 2001; 322: 835-840.
13. Morrison-Beedy D, Carey MP, Tu X. Accuracy of audio computer-assisted selfinterviewing (ACASI) and self-administered questionnaires for the assessment of
sexual behavior. AIDS Behav 2006; 10: 541-552.
14. Taylor J. http://www.e-health-insider.com/news/item.cfm?ID=1162. Last accessed
28/05/07 2007.
15. Tang PC, Jaworski MA, Fellencer CA, LaRosa MP, Lassa JM, Lipsey P et al.
Methods for assessing information needs of clinicians in ambulatory care. Proc
Annu Symp Comput Appl Med Care 1995; 630-634.
16. Tang PC, Jaworski MA, Fellencer CA, Kreider N, LaRosa MP, Marquardt WC.
Clinician information activities in diverse ambulatory care practices. Proc AMIA
Annu Fall Symp 1996; 12-16.
17. Jancin B. Two hours a day of unreimbursed time. Fam Pract News . 2001.
18. Wong WS, Lee KH, Chang MZ. A microcomputer based interview system for
antenatal clinic. Comput Biol Med 1986; 16: 453-463.
19. Davenport S, Goldberg D, Millar T. How psychiatric disorders are missed during
medical consultations. Lancet 1987; 2: 439-441.
20. Palermo TM, Valenzuela D, Stork PP. A randomized trial of electronic versus
paper pain diaries in children: impact on compliance, accuracy, and acceptability.
Pain 2004; 107: 213-219.
21. Tiplady B, Crompton GK, Dewar M.H, Bollert FGE, Matusiewicz SP, Campbell
LM. The use of electronic diaries in respiratory studies. Drug Inf J 1997; 31: 759764.
22. Dale O, Hagen KB. Despite technical problems personal digital assistants
outperform pen and paper when collecting patient diary data. J Clin Epidemiol
2007; 60: 8-17.
23. Supple AJ, Aquilino WS, Wright DL. Collecting sensitive self-report data with
laptop computers: Impact on the response tendencies of adolescents in a home
interview. Journal of Research on Adolescence 1999; 9: 467-488.
24. Skinner HA, Allen BA. Does the computer make a difference? Computerized
versus face-to-face versus self-report assessment of alcohol, drug, and tobacco
use. J Consult Clin Psychol 1983; 51: 267-275.
25. Carr AC, Ghosh A, Ancill RJ. Can a computer take a psychiatric history? Psychol
Med 1983; 13: 151-158.
26. Sawyer MG, Sarris A, Baghurst P. The use of a computer-assisted interview to
administer the Child Behavior Checklist in a child psychiatry service. J Am Acad
Child Adolesc Psychiatry 1991; 30: 674-681.
27. Sawyer MG, Sarris A, Baghurst P. The effect of computer-assisted interviewing
on the clinical assessment of children. Aust N Z J Psychiatry 1992; 26: 223-231.
232
28. Erdman HP, Klein MH, Greist JH. Direct patient computer interviewing. J Consult
Clin Psychol 1985; 53: 760-773.
29. Jaya, Hindin, M. J., & Ahmed, S. 2008, "Differences in young people's reports of
sexual behaviors according to interview methodology: a randomized trial in India",
Am.J Public Health, vol. 98, no. 1, pp. 169-174.
30. Brownbridge G, Lilford RJ, Tindale-Biscoe S. Use of a computer to take booking
histories in a hospital antenatal clinic. Acceptability to midwives and patients and
effects on the midwife-patient interaction. Med Care 1988; 26: 474-487.
31. Slack WV, Slack CW. Patient-computer dialogue. N Engl J Med 1972; 286: 13041309.
32. Wu RC, Straus SE. Evidence for handheld electronic medical records in
improving care: a systematic review. BMC Med Inform Decis Mak 2006; 6: 26.
33. Garritty C, El EK. Who’s using PDAs? Estimates of PDA use by health care
providers: a systematic review of surveys. J Med Internet Res 2006;8:e7.
34. Kho A, Henderson LE, Dressler DD, Kripalani S. Use of handheld computers in
medical education. A systematic review. J Gen Intern Med 2006; 21: 531-537.
35. Poissant L, Pereira J, Tamblyn R, Kawasumi Y. The impact of electronic health
records on time efficiency of physicians and nurses: a systematic review. J Am
36. Phillips AE, Gomez GB, Boily MC, Garnett GP.A systematic review and metaanalysis of quantitative interviewing tools to investigate self-reported HIV and STI
associated behaviours in low- and middle-income countries. Int J Epidemiol. 2010
Dec;39(6):1541-55.
37. Whiteford A, Allen A, Pappas Y, Car J, Minozzi S, Pagliari C, Sheikh A, Majeed A
(Protocol). Computer-assisted versus oral-and-written alcohol intake history
taking for diagnosing alcohol misuse or abuse in screened participants. Cochrane
Database of Systematic Reviews.
38. Pappas Y, Wei I, Car J, Majeed A, Sheikh A. Computer-assisted versus oral-andwritten family history taking for identifying people with elevated risk of type 2
diabetes mellitus. Cochrane Database of Systematic Reviews 2010, 4. Art. No.:
CD008489. DOI: 10.1002/14651858.CD008489.
39. Wei I, Pappas Y, Car J, Majeed A, Sheikh A. Computer-assisted versus oral-andwritten dietary history taking for diabetes mellitus. Cochrane Database of
Systematic
Reviews
2010,
4.
Art.
No.:
CD008488.
DOI:
10.1002/14651858.CD008488
40. Mayne JG, Martin MJ, Morrow GW Jr, Turner RM, Hisey BL. A health
questionnaire based on paper-and-pencil medium individualized and produced by
computer, I: technique. JAMA. 1969;208: 2060-2063.
41. Gaertner J, Elsner F, Pollmann-Dahmen K, Radbruch L, Sabatowski R. Electronic
pain diary: a randomized crossover study. J Pain Symptom Manage 2004; 28:
259-267.
42. Lauritsen K, Degl' IA, Hendel L, Praest J, Lytje MF, Clemmensen-Rotne K et al.
Symptom recording in a randomised clinical trial: paper diaries vs. electronic or
telephone data capture. Control Clin Trials 2004; 25: 585-597.
43. Kerkenbush NL, Lasome CE. The emerging role of electronic diaries in the
management of diabetes mellitus. AACN Clin Issues 2003; 14: 371-378.
44. Shiffman S, Hufford M. Methods of measuring patient experience: paper versus
electronic patient diaries. White Paper. Scotts Valley, Calif: invivodata, Inc; 2001.
2001.
45. Kamarack TW, Shiffman S, Smithline L. Effects of task strain, social conflict, and
emotional activation on ambulatory cardiovascular activity: daily life
consequences of recurring stress in a multiethnic adult sample. Health
Psychology 17, 17-29. 1998.
46. Richens J, Copas A, Sadiq ST, Kingori P, McCarthy O, Jones V, Hay P, Miles K,
Gilson R, Imrie J, Pakianathan M. A randomised controlled trial of computer-
233
assisted interviewing in sexual health clinics. Sex Transm Infect. 2010
Aug;86(4):310-4.
47. Des Jarlais DC, Paone D, Milliken J, Turner CF, Miller H, Gribble J, Shi Q, Hagan
H, Friedman SR. Audio-computer interviewing to measure risk behaviour for HIV
among injecting drug users: a quasi-randomised trial. Lancet. 1999 May
15;353(9165):1657-61.
48. Bulpitt CJ, Beilin LJ, Coles EC. Randomised controlled trial of computer-held
medical records in hypertensive patients. BMJ 1976; 677-679.
49. Tideman RL, Chen MY, Pitts MK, Ginige S, Slaney M, Fairley CK. A randomised
controlled trial comparing computer-assisted with face-to-face sexual history
taking in a clinical setting. Sex Transm Infect. 2007 Feb;83(1):52-6
Am Med Inform Assoc 1997;4:342–355.
51. Friedman RH. Automated telephone conversations to assess health behavior and
deliver behavioral interventions. J Med Syst 1998; 22(2): 95-102.
52. Revere D, Dunbar PJ. Review of Computer-generated Outpatient Health
Behavior Interventions: Clinical Encounters “in Absentia””. J Am Med Inform
Assoc 2001; 8(1): 62-79.
53. Farzanfar R. When computers should remain computers: a qualitative look at the
humanization of health care technology. Health Informatics J 2006; 12(3): 239254.
54. Greenhalgh T, Hinder S, Stramer K, Bratan T, Russell J. Adoption, non-adoption,
HealthSpace. BMJ. 2010 Nov 16;341:c5814. doi: 10.1136/bmj.c5814
55. Gagnon MP, Desmartis M, Labrecque M, Car J, Pagliari C, Pluye P, Frémont P,
Gagnon J, Tremblay N & Légaré F. Systematic Review of Factors Influencing the
Adoption of Information and Communication Technologies by Healthcare
Professionals. J Med Syst doi 10.1007/s10916-010-9473-4
56. Auerbach AD, Landefeld CS, Shojania KG. The tension between needing to
improve care and knowing how to do it. N Engl J Med 2007; 357: 608–613.
57. Rigby M. Essential prerequisites to the safe and effective widespread roll-out of
e-working in healthcare. Int J Med Inform 2006; 75: 138–147.
58. Stevens A, Milne R, Lilford R, Gabbay J. Keeping pace with new technologies:
systems needed to identify and evaluate them. BMJ 1999; 319: 1291.
59. Stroetmann K, Jones T, Dobrev A, Stroetmann V. eHealth is worth it: The
economic benefits of implemented eHealth solutions at ten European sites. 2006.
Luxembourg, Office for Official Publications of the European Communities.
Manage 2003; 24: 29–38.
61. Eisenstein EL, Ortiz M, Anstrom KJ, Crosslin DR, Lobach DF. Economic
evaluation in medical information technology: why the numbers don’t add up.
AMIA Annu Symp Proc 2006; 914.
62. First Consulting Group. Costs, benefits and challenges of CPOE. 2003.
63. Sidorov J. It ain’t necessarily so: the electronic health record and the unlikely
prospect of reducing health care costs. Health Aff (Millwood) 2006; 25: 1079–
1085.
234
Chapter 8
Computerised decision support systems
Summary
•
Improved
access
to
relevant
clinical
information
for
healthcare
professionals can translate into improvements in practitioner performance,
and possibly also the quality, safety and efficiency of care.
•
Computerised decision support systems are software applications that use
patient data, a repository of clinical information (the knowledge-base) and
an inference mechanism to generate patient specific recommendations for
care. These applications are highly variable in sophistication and in the
extent to which they integrate with other clinical information systems.
•
There is particular interest in using computerised decision support systems
in the context of delivering preventive care, ordering and interpreting
investigations, diagnostics, monitoring of patients, and guideline-based
and prescribing-related decisions on care provision.
•
Although numerous evaluations of these applications have taken place,
very little consistent and generalisable evidence exists on their ability to
improve patient outcomes; there is however emerging evidence that they
can, particularly if they involve provision of point-of-care recommendations
that integrate well with clinician workflows, result in modest improvements
in practitioner performance.
•
The use of computerised reminders for preventive care has been
empirically demonstrated to be of the most benefit in impacting on relevant
proxy measures. However, there are limited data on patient outcomes;
this to a large extent reflects the prohibitive size and/or duration of study
needed to demonstrate an effect on clinically relevant outcome measures.
•
These applications are largely unregulated in the US and UK, as they fall
outside the remit of the Federal Drug Administration and Medicine and
Healthcare Products Regulatory Agency, respectively.
235
•
Without formal quality and safety assurances for these applications, the
potential risks to patient safety need to be considered as they may in
certain situations inadvertently introduce new risks and errors.
•
As evidence of benefit is clearest and risk of harm is lowest in relation to
support for preventive healthcare, the NHS should consider introducing a
range of computerised health promotion and prevention-related tools into
primary care and, in the context of the ongoing roll-out of the National
Health Service Care Records Service, also into secondary care.
•
Future research should focus on understanding the contexts in which
these applications are most likely to prove effective and cost-effective in
improving patient outcomes and this should be a priority consideration for
NHS Connecting for Health as it introduces new eHealth applications with
built-in decision support functionality.
8.1 Introduction
Computerised decision support systems (CDSSs) are software applications
that integrate patient data with a knowledge-base and an inference
mechanism to produce patient specific output in the form of care
recommendations, assessments, alerts and reminders to actively support
practitioners in clinical decision-making.1 These systems can take a number
of forms; this depends on the particular task that they seek to support
clinicians with, the approach to drawing on patient data (which may have been
input by professionals or which may involve automated interrogation of
existing patient records), the type of knowledge-base that is drawn upon and
the inference mechanism that is employed, the types of outputs that are
generated and the ways in which these are communicated to healthcare
providers. Finally, whilst patient or consumer orientated CDSSs now exist
(and may well have a lot of potential for health gains) these will not be
considered in this chapter as these fall outside the scope of this report.
236
This chapter aims to provide an overview of CDSSs aiming to support
healthcare providers with delivery of care. It should be read in conjunction
with Chapter 9, which focuses in detail on the specific case of CDSS support
for image interpretation in the context of
mammography screening, and
Chapters 10, 11 and 12, which are concerned with CDSS for prescribing
decisions and medicines management considerations.
8.2 Definition, description and scope for use
8.2.1 Definition
There is no universally agreed definition of CDSS, this in part at least
reflecting the continuing evolution of the scope, reasoning approach and
application of artificial intelligence (AI) based computerised clinical support
systems.
Wyatt and Spiegelhalter’s definition of CDSS as ‘…active
knowledge systems which use two or more items of patient data to generate
case-specific advice’2
is widely used and is useful, but as this definition
currently stands it excludes simple memory aids to clinicians (such as the
more basic computer-assisted history taking systems discussed in the
previous chapter). It further also excludes the more novel patient support
tools (sometimes known as decision support systems), systems that make
clinical decisions on population level data and systems that provide additional
information to a clinician (such as prognosis) without explicitly supporting the
decision-making process.
A related, but more inclusive definition has more
recently been proposed, defining these tools as providing
‘…access to
knowledge stored electronically to aid patients, carers and service providers in
making decisions on health care’.3 There is thus, as with most other eHealth
applications, need for greater clarity in what exactly is being referred to when
using the term CDSS. In the context of this chapter, we will be focusing on
studies investigating computerised systems that draw on specific patient data
to support the management of individual patients by professionals.
8.2.2 Description of use
Computerised decision support systems vary in design and function.4
Applications can be used by any healthcare worker involved in the provision
of healthcare. They can be “stand-alone” or, as is increasingly the case,
237
“integrated” with other clinical information systems, in particular electronic
health records (EHRs).
Patient data can be inputted by manual entry,
transferred from other clinical information systems or transmitted from medical
devices. Engagement by professionals with a CDSS may be “active” and/or
“passive”, this referring to whether the information is “pushed” to the user–
often in real-time so as to be able to support care decisions at the point-ofcare (POC)–or whether professionals need to retrieve or “pull” the advice from
the CDSS.
In generating the recommendations, the patient data (input) are compared
against a knowledge-base (the collection of clinical knowledge) and made
sense of by an inference mechanism (the logic). These knowledge-bases
may have been developed in-house or, as is increasingly the case, procured
from a commercial provider. The underlying reasoning approach or inference
mechanism can be highly variable in sophistication and approach, these
ranging from the relatively simple clinical algorithms and Bayesian-based
approaches to the more advanced rules-based expert systems and machine
learning-based approaches.1
The output can also take different forms and
can be delivered to a number of destinations occurring at any time before,
during or after the patient-professional interaction.
Several helpful
classification approaches to CDSSs have been proposed, but none appears
as yet to be able to do justice to the complex multi-dimensional nature of
CDSSs.3-7 Perhaps most interesting in this respect is Sim et al.’s framework,
in which they distinguish between evidence-adaptive applications (i.e., those
in which the underlying knowledge-base of the CDSS is based on evidence
that is continually updated) and those that draw on more static knowledge
sources.7
8.2.3 Scope for use
As discussed in Chapters 1 and 4, there are concerns now being expressed
internationally about major deficiencies in the quality, safety and efficiency of
healthcare.8 This is at least in part due to the difficulties that clinicians have in
accessing and processing the vast amounts of research now being published,
the challenges in applying this to decisions relating to the care of individual
238
patients, and the increasing data that can now be generated in relation to
individual patients (e.g. a patient in an intensive care unit who is being
continually monitored), which can rapidly lead to cognitive overload.9
The
general premise underlying the use of CDSSs is that they support clinicians in
making more informed decisions. As such, knowledge-bases can fill the gaps
in clinicians’ knowledge and inference mechanisms can aid in the
interpretation of patient data to improve clinical decision-making.
By
improving clinical decision-making, CDSSs have the potential to impact on the
quality—the appropriateness, timeliness and effectiveness—and safety of
healthcare; they are also widely seen as having the potential to remove some
of the inefficiencies of care through rationalising expenditure on diagnostic
tests and treatment decisions.
CDSSs can be used for a variety of clinical activities such as preventive care,
diagnostics, therapeutics, disease management, prognostics and public
health surveillance. Theoretically, CDSSs can be used for any speciality of
clinical care and in any setting where the requisite knowledge-base and
technological infrastructure exists.
They can potentially also be used to
support research through, for example, identifying eligible patients and
supporting data collection according to pre-defined protocols in enrolled
patients. The plasticity of this application thus suggests a diverse range of
opportunities to impact on care provision.
8.3.1 Benefits
Whilst the range of ways in which CDSSs can be used is potentially endless,
the greatest interest currently relates to supporting management decisions
rather than the more difficult to model processes of diagnostics and
prognostication. The immediate impacts of these tools are likely to be the
understanding, confidence and competencies of clinicians, which will it is
hoped translate into more appropriate, evidence-based decisions and
improved outcomes for patients. Through promoting best practice, supporting
monitoring of patients and alerting professionals to potential risks and errors,
these clinician-support tools may also result in a reduction of adverse events.
239
It is based on the presumed improvements in both the quality and safety of
care that claims about the cost-effectiveness of decision support tend to be
made.
8.3.2 Risks
There is a risk that the introduction of CDSSs may have a detrimental effect
on care processes and outcomes, this resulting either from faulty algorithms
resulting in erroneous advice, degradation in practitioner knowledge and
resulting over-reliance on the system and the disruption resulting from a
change in workflow processes resulting from poorly integrated tools.
The
known data inaccuracies in patient EHRs are also potentially important
considerations in the context of the introduction of tools that draw on such
information.10 11 There are also potential legal considerations associated with
clinicians ignoring appropriate advice and also institutions investing in poor
quality CDSSs that also need to be considered.12
The costs of investing in
CDSSs also need to be considered, as these can be substantial. It is
therefore important that these potential adverse outcomes, whether at the
level of the practitioner, patient or the organisation are considered and, where
relevant, studied.13
We found 24 relevant systematic reviews (SRs) investigating various aspects
of the effectiveness and safety of CDSSs (see Appendix 5 for details of these
reviews).14-37 These reviews had a range of foci, these including: providing an
overview of the impact of eHealth and/or CDSSs; investigating their impact in
specific clinical settings (e.g., nursing, primary care, intensive care settings
and
pathology services) and
disease
contexts
(e.g.,
diabetes
and
hypertension); test ordering; and the promotion of preventive and guidelinebased care. In a few cases, these represented substantive updates of earlier
reviews,38-41 in which case, we focused on the most recent reports. We also
identified two meta-analyses investigating the impact of CDSSs on aspects of
preventive care (vaccination and screening) that reported before 1997, but
which nonetheless provided some relevant insights.42
43
For the sake of
brevity, we do not detail reviews on the use of CDSS for image interpretation.
240
We do, however, undertake a detailed case study of the use of CDSS in the
context of mammography screening in the following chapter, which illustrates
the most pertinent issues. Reviews specifically focusing on the use of CDSSs
in prescribing decisions (ePrescribing) are discussed in Chapter 10, with other
aspects of medicine management consideration explored in the case studies
presented in Chapters 11 and 12.
Given the range of CDSS interventions, study designs, settings, contexts,
outcomes and time periods studied, synthesising this body of evidence was in
many respects challenging. Also of relevance was that these SRs tended not
to report on the full range of practitioner, patient and organisational outcomes
discussed above; in particular there was a deficiency in reporting on possible
adverse events.
8.4.1 Benefit
Practitioner performance
A number of SRs have demonstrated that CDSSs can impact positively on
practitioner performance, such as the promotion of preventive care and
guideline adherence. The SR by Garg et al. is the most important in this
respect (and in general in that it evaluates a substantial body of evidence from
100 trials), revealing that CDSSs had positive impacts in 16/21 (76%) of the
studies investigating the issuing of reminders for cervical cancer screening,
mammography and vaccination.19
Similar findings have also been
demonstrated in other systematic reviews for aspects of practitioner
performance about relatively simple prevention decisions. Interestingly, a
broadly similar finding was observed in much earlier systematic reviews.42 43
For example, Austin et al. found in their meta-analysis of three cervical
screening trials that reminders resulted in an increased number of
procedures: OR=1.18 (95%CI 1.02, 1.40); similarly, the pooled estimate from
the three tetanus vaccination trials demonstrated an increase in vaccination
rate: OR=2.82 (95%CI 2.66, 2.98).42
Practitioner performance is more usually improved in relation to guideline
adherence and specific tasks in relation to chronic disease management, this
241
having been demonstrated across a range of care settings and disease
contexts. Garg et al. for example found improvements in practitioner
performance in 23/37 (62%) of the trials studying this outcome.19 The review
by Heslemans is interesting in that it focused on studies investigating the
impact of CDSSs on adherence with more complex guidelines (which they
characterise as “multidimensionality”), but this found improvements in
measures of practitioner performance in 7/17 (41%) studies, this suggesting
that CDSSs might be more effective with the relatively simple tasks that can
easily be modelled.21
This may also be why these tools are less effective in
the context of supporting clinical decision making (positive impacts
demonstrated in 4/40 (10%) of the diagnostic studies.19
Patient outcomes
A recurring theme in this body of work is that patient outcomes tend not to be
studied by investigators and when they are the results are on the whole
disappointing. For example, fifty-two of the 100 trials retrieved by Garg et al
reported on patient outcomes and of these only seven (13%) showed positive
patient level outcomes (two of which were in the context of medicines
management interventions, which will be considered further in Chapter 10).19
Similar disappointing patient-level outcomes have also been reported by,
amongst others, Georgiou et al.,20 Heselmans et al.,21 Jamal et al.,22 Mador et
al.27 and Main et al.28
Cost-effectiveness
Another important concern highlighted in the literature relates to the issue of
cost-effectiveness. Garg et al were unable to uncover much data to estimate
the cost-effectiveness of CDSSs, and most of the other SRs we found are
silent in this respect, this perhaps also reflecting the lack of available data.19
The recently reported review by Main et al.,28 which set out to compare the
effectiveness of a CDSS-based approach when compared to simpler order
communications support (also know as computerised physician order entry) in
the context of test requests only found two studies yielding economic data:
one showed a cost reduction with CDSS, whilst the other found an increase in
costs for the tests requested.
242
8.4.2 Risks
The SRs reviewed have, as noted above, tended not to report adverse
outcomes. This may reflect the fact that these interventions are largely safe
or, alternatively, it may represent a failure adequately to consider such risks,
whether in the primary studies or the secondary reviews. Looking more
widely, concerns have been expressed about the effects of poor data
quality—accuracy and completeness—on CDSS functioning. For instance,
Berner et al. explored the effect of incomplete patient data on CDSS accuracy
and found that when the available data were input into the CDSS, the missing
data elements resulted in inappropriate and unsafe recommendations in
almost 77% of encounters.44
8.5.1 Policy
Computerised decision support systems are currently outwith the remit of the
Medicines Health Regulation Authority (MHRA) in the UK and the Federal
Drugs Authority (FDA) in the USA. However, in view of the potential for the
introduction of new risks, it is important that consideration be given to the
need for regulation of CDSSs.12 As discussed in Chapter 11, such regulation
could also address the issue of “defensive” practices by software developers
(e.g. the factoring into the algorithms of any conceivable possibility), which
may through “over-alerting” inadvertently increase the risk of adverse events.
The policy challenge here however is not to over-regulate so it stifles
innovation in this field. 12
Given the considerable costs of developing and maintaining CDSSs, Wright et
al have made an interesting suggestion in relation to the possible benefits
from sharing the content of CDSSs both within and, more contentiously in the
US context, between organisations.45
This has considerable potential
applicability to the NHS context where competing partners could be brought
together in the national interest to work on collaborative projects. We discuss
this issue further in Chapter 11 when considering ePrescribing. It also seems
243
prudent in this respect for the NHS to develop an inventory–that is periodically
updated–of decision support tools that are available for and in use in the
NHS.28 46
8.5.2 Practice
Based on the evidence reviewed, it seems likely that most CDSSs will not
result in measurable patient-level benefits; they may however translate into
practitioner or process level improvements, particularly in the context of
CDSSs aiming to improve aspects of preventive care, guideline adherence
and disease management; achieving even this level of benefits seems more
challenging in the context of tools aiming to improve diagnosis. It is therefore
important that claims made about these decision support tools are carefully
scrutinised before investing in these technologies. The research to date has
also offered several other pointers as to which tools may prove particularly
effective in improving these practitioner/process end-points, which it may also
prove helpful to consider when investing in CDSSs, these including: from
Garg et al.,19 automatic prompting to use the system (adjusted OR 3.0, 95%CI
1.2, 7.1) and designers of the systems also being involved with the evaluation
(adjusted OR 6.6, 95%CI 1.7, 26.7); from Kawamoto et al.25: automatic
provision of decision support as part of clinician workflow (adjusted OR 105.0,
95%CI 10.4, ∞); and from Damiani et al.,16 automatic provision of
recommendations that are integrated with workflow: (adjusted OR 17.5,
95%CI 1.6, 193.7) and publication year (adjusted OR 6.7, 95%CI 1.3, 34.3).
More fundamentally, there is the need for clinical engagement to identify the
areas in which CDSSs are most likely to prove useful to clinicians, rather than
serving as a technology that is introduced for technology’s sake.47 48
8.5.3 Research
There are a number of important unanswered questions in relation to CDSSs
that warrant further investigation. Arguably, the most important is to need to
establish which contexts, if any, these decision support tools can translate into
patient level benefits.
Neither the technology or its implementation or
maintenance is cheap–it is therefore also crucially important that we obtain a
244
better idea of the cost-effectiveness of CDSSs, particularly in contrast to
potentially cheaper and less disruptive approaches.
Such economic
considerations are particularly important in these times of economic difficulties
facing the UK and many other countries.
It seems more likely however that the impact of CDSSs aimed at
professionals will, at best, be on those professionals, rather than translating
into demonstrable downstream improvements in patient outcomes. This is
because the causal chain of events to impact on patients is (relatively) long
i.e. technology → practitioner action → patient actions → patient outcomes,
and also because many of the clinical outcomes of interest are uncommon
events (e.g. cervical cancer), which would necessitate studies involving large
numbers of people who may need to be followed up over many years in order
to have adequately powered studies.
There is therefore the need for
academic consensus building work on the context in which proxy measures
are suitable surrogate markers to study; clinical endpoints should however
continue to be studied in order to facilitate future meta-analysis of the types
undertaken by Austin et al.42 and Shea et al.43
It is also important that
researchers study and report on the full range of potentially relevant outcomes
associated with CDSSs, whether showing evidence of benefit or harm, using
an appropriate range of experimental and naturalistic designs.
7 13 24
Finally, it
is important that future work focuses on the effectiveness of commercial “offthe-shelf” rather than “home-grown” solutions.1 19 33 38
References
1. Coiera E. Guide to Health Informatics. 2nd ed. London: Hodder Arnold, 2003.
2. Wyatt J, Spiegelhalter D. Field trials of medical decision-aids: potential problems
and solutions. Proc Annu Symp Comput Appl Med Care 1991; 3-7.
3. National Electronic Decision Support Taskforce (NEDST). Electronic Decision
Support for Australia’s Health Sector. Canberra: Australian Government
Department of Health and Ageing, 2003.
4. Berlin A, Sorani M, Sim I. A taxonomic description of computer-based clinical
decision support systems. J Biomed Inform 2006; 39: 656-67.
5. Perreault L, Metzger J. A pragmatic framework for understanding clinical decision
support. J Healthc Inf Manag 1999; 13: 5-21.
6. Broverman CA. Standards for clinical decision support systems. J Healthc Inf
Manag 1999; 13: 23-31.
245
7. Sim I, Gorman P, Greenes RA, Haynes RB, Kaplan B, Lehmann H et al. Clinical
decision support systems for the practice of evidence-based medicine. J Am Med
Inform Assoc 2001; 8: 527-34.
8. Institute of Medicine. Crossing the Quality Chasm: A New Health System for the
21st Century. Washington, D.C., National Academy Press, 2001.
9. Smith R. Strategies for coping with information overload. BMJ 2010; 341: c7126.
11. Jordan K, Porcheret M, Croft P. Quality of morbidity coding in general practice
computerized medical records: a systematic review. Fam Pract 2004; 21: 396412.
12. Fox J, Thomson R. Clinical decision support systems: a discussion of quality,
safety and legal liability issues. Proc AMIA Symp 2002; 265-269.
13. Ash JS, Sittig DF, Campbell EM, Guappone KP, Dykstra RH. Some unintended
consequences of clinical decision support systems. AMIA Annu Symp Proc 2007;
26-30.
14. Balas EA, Krishna S, Kretschmer RA, Cheek TR, Lobach DF, Boren SA.
Computerized knowledge management in diabetes care. Med Care 2004; 42:
610-21.
15. Bryan C, Boren SA. The use and effectiveness of electronic clinical decision
support tools in the ambulatory/primary care setting: a systematic review of the
literature. Inform Prim Care 2008; 16: 79-91.
16. Damiani G, Pinnarelli L, Colosimo SC, Almiento R, Sicuro L, Galasso R, et al.
The effectiveness of computerized clinical guidelines in the process of care: a
systematic review. BMC Health Serv Res 2010; 10: 2.
17. Delpierre C, Cuzin L, Fillaux J, Alvarez M, Massip P, Lang T. A systematic review
of computer-based patient record systems and quality of care: more randomized
clinical trials or a broader approach? Int J Qual Health Care 2004; 16: 407-16.
18. Dexheimer JW, Talbot TR, Sanders DL, Rosenbloom ST, Aronsky D. Prompting
clinicians about preventive care measures: a systematic review of randomized
controlled trials. J Am Med Inform Assoc 2008; 15: 311-20.
19. Garg AX, Adhikari NK, McDonald H, Rosas-Arellano MP, Devereaux PJ, Beyene
J, et al. Effects of computerized clinical decision support systems on practitioner
performance and patient outcomes: a systematic review. JAMA 2005; 293: 122338.
20. Georgiou A, Williamson M, Westbrook JI, Ray S. The impact of computerised
physician order entry systems on pathology services: a systematic review. Int J
Med Inform 2007; 76: 514-29.
21. Heselmans A, Van de Velde S, Donceel P, Aertgeerts B, Ramaekers D.
Effectiveness of electronic guideline-based implementation systems in
ambulatory care settings - a systematic review. Implement Sci 2009; 4: 82.
22. Jamal A, McKenzie K, Clark M. The impact of health information technology on
the quality of medical and health care: a systematic review. HIM J 2009; 38: 2637.
23. Jerant AF, Hill DB. Does the use of electronic medical records improve surrogate
patient outcomes in outpatient settings? J Fam Pract 2000; 49: 349-57.
24. Kaplan B. Evaluating informatics applications--clinical decision support systems
literature review. Int J Med Inform 2001; 64: 15-37.
25. Kawamoto K, Houlihan CA, Balas EA, Lobach DF. Improving clinical practice
using clinical decision support systems: a systematic review of trials to identify
features critical to success. BMJ 2005; 330: 765.
26. Lisboa PJ, Taktak AF. The use of artificial neural networks in decision support in
cancer: a systematic review. Neural Netw 2006; 19: 408-15.
246
27. Mador RL, Shaw NT. The impact of a Critical Care Information System (CCIS) on
time spent charting and in direct patient care by staff in the ICU: a review of the
literature. Int J Med Inform 2009; 78: 435-45.
28. Main C, Moxham T, Wyatt JC, Kay J, Anderson R, Stein K. Computerised
decision support systems in order communication for diagnostic, screening or
monitoring test ordering: systematic reviews of the effects and cost-effectiveness
of systems. Health Technol Assess 2010; 14: 1-227.
29. Montgomery AA, Fahey T. A systematic review of the use of computers in the
management of hypertension. J Epidemiol Community Hlth 1998; 52: 520-25.
30. Randell R, Mitchell N, Dowding D, Cullum N, Thompson C. Effects of
computerized decision support systems on nursing performance and patient
outcomes: a systematic review. J Health Serv Res Policy 2007; 12: 242-49.
31. Shiffman RN, Liaw Y, Brandt CA, Corb GJ. Computer-based guideline
implementation systems: a systematic review of functionality and effectiveness. J
32. Rosado B, Menzies S, Harbauer A, Pehamberger H, Wolff K, Binder M et al.
Accuracy of computer diagnosis of melanoma: a quantitative meta-analysis. Arch
Dermatol 2003; 139: 361-67.
33. Shekelle PG, Goldzweig CL. Costs and Benefits of Health Information
Technology: an updated systematic review. London: The Health Foundation,
2009.
34. Shojania KG, Jennings A, Mayhew A, Ramsay C, Eccles M, Grimshaw J. Effect
of point-of-care computer reminders on physician behaviour: a systematic review.
CMAJ 2010; 182: E216-25.
35. Sintchenko V, Magrabi F, Tipper S. Are we measuring the right end-points?
Variables that affect the impact of computerised decision support on patient
outcomes: a systematic review. Med Inform Internet Med 2007; 32: 225-40.
36. Smith MY, DePue JD, Rini C. Computerized decision-support systems for
chronic pain management in primary care. Pain Med 2007; 8: S155-66.
37. Tan K, Dear PR, Newell SJ. Clinical decision support systems for neonatal care.
Cochrane Database Syst Rev 2005; 2: CD004211.
38. Chaudhry B, Wang J, Wu S, Maglione M, Mojica W, Roth E, et al. Systematic
review: impact of health information technology on quality, efficiency, and costs of
medical care. Ann Intern Med 2006; 144: 742-52.
39. Shekelle PG, Morton SC, Keeler EB. Costs and Benefits of Health Information
Technology. Rockville: AHRQ, 2006.
40. Johnston ME, Langton KB, Haynes RB, Mathieu A. Effects of computer-based
clinical decision support systems on clinician performance and patient outcome.
A critical appraisal of research. Ann Intern Med 1994; 120: 135-42.
41. Hunt DL, Haynes RB, Hanna SE, Smith K. Effects of computer-based clinical
decision support systems on physician performance and patient outcomes: a
systematic review. JAMA 1998; 280: 1339-46.
42. Austin SM, Balas EA, Mitchell JA, Ewigman BG. Effect of physician reminders on
preventive care: meta-analysis of randomized clinical trials. Proc Annu Symp
Comput Appl Med Care 1994: 121-24.
43. Shea S, DuMouchel W, Bahamonde L. A meta-analysis of 16 randomized
controlled trials to evaluate computer-based clinical reminder systems for
preventive care in the ambulatory setting. J Am Med Inform Assoc 1996; 3: 399409.
44. Berner ES, Kasiraman RK, Yu F, Ray MN, Houston TK. Data quality in the
outpatient setting: impact on clinical decision support systems. AMIA Annu Symp
Proc 2005; 41-45.
45. Wright A, Bates DW, Middleton B, Hongsermeier T, Kashyap V, Thomas SM, et
al. Creating and sharing clinical decision support content with Web 2.0: Issues
and examples. J Biomed Inform 2009; 42: 334-46.
247
46. Sittig DF, Wright A, Simonaitis L, Carpenter JD, Allen GO, Doebbeling BN, et al.
The state of the art in clinical knowledge management: an inventory of tools and
techniques. Int J Med Inform 2010; 79: 44-57.
47. Catwell L, Sheikh A. Information technology (IT) system users must be allowed to
decide on the future direction of major national IT initiatives. But the task of
redistributing power equally amongst stakeholders will not be an easy one. Inform
Prim Care 2009; 17: 1-4.
48. Handler JA, Feied CF, Coonan K, Vozenilek J, Gillam M, Peacock PR Jr, et al.
Computerized physician order entry and online decision support. Acad Emerg
Med 2004; 11: 1135-41.
248
Chapter 9
Case study: Computerised decision support for
mammography screening
Summary
•
Breast cancer is one of the leading causes of cancer death in the
developed world. Screening mammography is recommended by many
major institutions for women in order to reduce mortality from breast
cancer.
•
Routine breast cancer screening begins in women from the age of 50
and continues every three years until the age of 70 years; ongoing
scans are available to women who request this service beyond the age
of 70 years.
•
In England, “gold standard” screening mammography involves four
images: two views per breast, read by a reader trained in
mammography; these readings are then, in most cases, independently
double checked by a second reader. This second reading increases
sensitivity and specificity of cancer detection and is at present usually
performed by a radiologist.
•
Increases in the proportion of women being screened, the frequency
with which they are screened and the lack of trained professionals
places a considerable resource burden on health systems. Although the
gold standard for screening mammography, readings made by
professionals are subject to factors that risk the quality of readings and
therefore patient safety.
•
Computerised decision support systems for image interpretation (most
commonly referred to as computer-aided detection) denotes using the
output of software analysis as a “second read” before making a
diagnosis.
•
Computer-aided detection for screening mammography has the potential
to improve the sensitivity and specificity of screening mammography and
reduce the burden placed on health services.
249
•
Recent studies suggest that computer-aided detection can result in
comparable detection rates of breast cancer to that achieved by double
human readings.
•
Computer-aided detection does however result in an increase in recall
rate with an associated increase in the number of diagnostic procedures
in healthy women.
•
There is as yet a lack of clear evidence from randomised controlled trials
of the impact of computer-aided detection for screening mammography
on improving long-term patient outcomes or on the cost-effectiveness of
this approach when compared with double reading.
•
In the absence of such evidence, the NHS Screening Programme
should continue with the current practice of double-reading of
mammograms by humans.
9.1 Introduction
Computerised decision support systems (CDSSs) have been employed for a
variety of clinical activities including image interpretation, with the intention of
improving the quality and safety of healthcare (see Chapter 8). Although a
type of CDSS, these applications are generally referred to in the literature as
computer-aided detection (CAD). In this case study, we consider CAD for
screening mammography as an example of a CDSS, which illustrates some
important issues typically encountered when considering the implementation
and evaluating the impact of new eHealth applications.
This case study
demonstrates that, despite theoretically and some empirically proven benefits,
novel technologies do not always translate into improvements in everyday
healthcare practice. We use this case study to highlight some of the reasons
underlying the lack of effectiveness of CAD in practice.
The purpose of a screening programme is to identify as many people as
possible with pathological findings in the early stages of disease and therefore
have the opportunity to initiate treatment early. Breast cancer is one of the
250
leading causes of cancer death in women in the developed world.1 2 Detection
of pathology and initiation of treatment in the early stages of breast cancer
leads to reduced mortality, but mammography-based screening also may lead
to over-diagnosis.3
Mammography is the “gold standard” procedure for breast cancer detection in
screening programmes. Mammography is a technique that uses low dose xrays to visualise early malignant changes in breast tissue.
In England, it
involves four images: two views per breast, read by a reader trained in
mammography; these readings are then, in most cases, independently double
checked by a second reader.4
Although considered the gold standard, this procedure is not ideal as up to
one-third of visible cancers are missed by screening programmes.5 There are
several factors recognised as possibly contributing to suboptimal readers’
performances in screening programmes. These include:5-8
•
Difficulties resulting from reading large numbers of normal radiographs
whilst at the same time needing to remain vigilant about the very few
abnormalities;
this
is
because
continuous
reading
of
normal
mammograms is monotonous, which can lead to reading fatigue with
the resulting heightened risk of overlooking rare pathological findings
•
Low image quality
•
Inappropriate viewing conditions
•
Distractions
•
Oversight of pathology in one region of the breast resulting from
focusing on more obvious findings in another region
•
Reliance on prior knowledge of the most probable locations of cancers
•
Insufficient “reading” skills and/or experience.
With increasing numbers of women being screened, an increase in views per
person,
coupled with a lack of radiologists trained in mammography has
created conditions that have in the UK necessitated a number of other
healthcare professionals—such as breast clinicians and radiographers—to
251
become increasingly involved in reading mammograms. Their training and
experience remain critical to the accuracy of diagnosis.
Computer-aided detection in mammography refers to establishing a diagnosis
with the help of specific pattern recognition software that marks suspicious
features on mammograms thereby attracting readers’ attention to possible
pathology.9 10 CAD analysis typically takes place after the initial reading by a
radiologist thus taking the place of a second human reader.11
The widespread use of CAD in many centres in the US12 raises the possibility
of this eHealth application being adopted in other healthcare settings. It is,
however, important that prior to any such decision being taken that the
benefits and risks associated with CAD for mammography be critically
assessed.
9.2 Theoretical considerations
Computer-aided detection has the potential to aid radiologists by marking
suspicious lesions thereby reducing the possibility of missing pathological
findings.7
13
In other words, CAD could increase the sensitivity of screening
mammography, i.e. reduce the number of false negatives.
Increased
sensitivity should in turn translate into fewer missed cancers and hence
improved survival rates.
Approximately 95 per cent of women with abnormalities on screening
mammograms do not have breast cancer and a variable proportion of these
women undergo unnecessary additional diagnostic procedures and/or
treatment.3 Accurate algorithms applied to mammograms could potentially
“recognise” only the true positive lesions with few or no false-positives thus
increasing specificity with the computer-aided analysis. Increased specificity
would then translate into reduced recall rates, reduced number of
unnecessary diagnostic procedures and treatments and consequently,
reduced risks associated with these interventions.
252
If high sensitivity and specificity were achieved this could then obviate the
need for the current dual human reading of mammograms, with the potential
for efficiency gains.14
Furthermore, CAD systems have the potential to
increase the accuracy and consistency of less-skilled readers.15
Flaws in CDSSs’ software design can however inadvertently lead to a
deterioration of patient safety. For example, if CAD results in a decrease in
specificity this will result in an increased number of false positives, resulting in
unnecessary
anxiety,
testing
and
associated
procedure-related
complications.3 16-18
On the other hand, low sensitivity (i.e. the absence of clinically relevant
prompts) might also detrimentally impact on readers’ decision making.
Alberdi et al. emphasise that readers might base their decisions to not recall
patients on the absence of computer prompts.19 Since CAD analysis is
unlikely to be able to identify all lesions, a negative test result might induce a
false sense of security. Theoretically, such reliance on the CAD support might
also result in a deskilling effect for readers—reducing their ability to interpret
images without the support of CAD.
9.3 Empirically demonstrated impact
We found two relevant systematic reviews (SRs),20 21 which have shown that
there is to date no clear evidence that CAD in screening mammography
programmes improves patient outcomes.
The first SR by Taylor et al.,20 based on an analysis of 10 prospective
observational studies, found that a single human reading and CAD resulted in
no increase in cancer detection rate (OR 1.04, 95%CI 0.96, 1.13), but an
increase in recall rate (OR 1.13, 95%CI 1.05, 1.23) when compared to double
human reading. Based on this analysis, they concluded that double human
reading remains the preferred model.
253
The more recent SR by Noble et al.,21 which drew on largely the same body of
evidence, came to a similar conclusion. The authors identified three studies
that allowed sensitivity and specificity of CAD to be calculated, this finding a
pooled sensitivity of 86.0% (95%CI 84.2, 87.6) and a pooled specificity of
88.2% (95%CI 88.1, 88.3).
Five studies yielded data, which allowed the
calculation of rates of correct cancer diagnosis and invasive investigational
procedures in women without breast cancer. This found that CAD resulted in
50 (95%CI 30, 80) additional correct breast cancer diagnoses per 100,000
women screened, but there was an additional 1,190 (95%CI 1090, 1290)
recalls in healthy women and 80 (95%CI 60, 100) biopsies in these women.
They did not however find any data on whether these additional diagnoses
translated into improved patient outcomes, which is important because it has
been suggested that CAD might preferentially detect slow-growing cancers
that might not have resulted in clinical problems during patients’ lifetime.
The recent randomised controlled trial (RCT) from the CADET II Group has
also generated broadly similar findings.22
This English multi-centre
equivalence trial involving 31,057 women compared the effectiveness of
single-reading with CAD and double reading. It found equivalent numbers of
cancers detected by double reading and single reading with CAD (87.7% vs.
87.2%; P=0.89), but higher recall rates with CAD (3.9% vs. 3.4%; P<0.001).
No data were however presented on patient outcomes associated with cancer
detection or indeed on cost-effectiveness considerations. In a follow-on
analysis, the authors have shown that cancers with different radiological
features were detected using these two different approaches.23 The clinical
implications of this observation is at present unclear.
9.4 Implications for practice, policy and research
Overall, the evidence suggests that single reading with CAD is equivalent to
double reading by humans in detecting cancer, but there is an increase in
recall rate and associated diagnostic procedures in those screened using
CAD.
There is as yet no evidence on whether CAD-based screening
translates into improved patient outcomes or indeed whether this is costeffective to health systems. Based on the available evidence, it is we feel
254
appropriate that in developed country settings such as the UK, where double
screening is the routine, screening programmes should continue with dual
human screening.
Given the strong theoretical grounds for benefits associated with CAD, it is
important to consider why this technology is currently failing to realise its
potential. Different developers of CAD systems use different algorithms for
image analysis and this might, in some cases, result in different prompts for
the same lesions.24 Although the output given by various brands and versions
of CAD software is very similar, possible small differences in output have not
yet been properly evaluated. And whilst application invalidity (low specificity
for example) might play a part in its relative lack of diffusion, it appears that
the automation of image interpretation also proves problematic for other
reasons.
Providing reminders concerning preventive care and calculating drug doses
are examples of activities that are readily automatable (see Chapters 8 and
10). The art of diagnosis or image interpretation is just that, and it is therefore
extremely difficult to automate; though not well studied, but probably also
important, is that professional autonomy might be perceived as being
encroached by applications such CAD.
Although there are difficulties
associated with reading mammograms there have been no calls to greatly
improve the accuracy of double-reading and the use of CAD for screening
mammography screening thus seems to be technology driven rather than
fulfilling a genuine clinical need.25
References
1. American Cancer Society. Breast cancer facts and figures. Available from:
http://www.cancer.org/Research/CancerFactsFigures/BreastCancerFactsFigure
s/breast-cancer-facts--figures-2009-2010 (last accessed 8/1/11).
2. Cancer Research UK, Cancer mortality - UK statistics. Available from:
http://info.cancerresearchuk.org/cancerstats/mortality/ (last accessed 8/1/11).
3. Gotzsche PC, Nielsen M. Screening for breast cancer with mammography.
Cochrane Database Syst Rev 2009; 4: CD001877.
255
4. NHS
Breast
Screening
Programme.
Available
from:
http://www.cancerscreening.nhs.uk/breastscreen/index.html (last accessed
8/1/11).
5. Yankaskas BC, Schell MJ, Bird RE, Desrochers DA. Reassessment of breast
cancers missed during routine screening mammography: a community-based
study. AJR Am J Roentgenol 2001; 177: 535-41.
6. Erickson BJ, Bartholmai B. Computer-aided detection and diagnosis at the
start of the third millennium. J Digit Imag 2002; 15: 59-68.
7. Astley SM. Computer-aided detection for screening mammography. Acad
Radiol 2004; 11: 1139-43.
8. Huynh PT, Jarolimek AM, Daye S.
The false-negative mammogram.
Radiographics 1998; 18: 1137-54.
9. Giger ML, Karssemeijer N, Armato SG, III. Computer-aided diagnosis in
medical imaging. IEEE Transactions on Medical Imaging 1920; 1205-08.
10. Astley SM, Gilbert FJ. Computer-aided detection in mammography. Clin
Radiol 2004; 59: 390-99.
11. Bywood P, Newton S, Merlin T, Braunack-Mayer A, Hiller J. Computer Aided
Detection Systems in Mammography. Canberra: Australia and New Zealand
Horizon Scanning Network. Horizon Scanning Report, 2004.
12. Doi K. Current status and future potential of computer-aided diagnosis in
medical imaging. Brit J Radiol 2005: 78 (Spec No 1): S3-S19.
13. Taylor P, Champness J, Given-Wilson R, Johnston K, Potts H. Impact of
computer-aided detection prompts on the sensitivity and specificity of screening
mammography. Health Technol Assess 2005; 9: iii:1-58.
14. Dinnes J, Moss S, Melia J, Blanks R, Song F, Kleijnen J. Effectiveness and
cost-effectiveness of double reading of mammograms in breast cancer
screening: findings of a systematic review. Breast 2001; 10: 455-63.
15. Blue Cross Blue Shield Association.
Technology Evaluation Center
Assessment: Computer-aided Detection with Full-field Digital Mammography,
2006.
16. Stoate HG. Can health screening damage your health? J R Coll Gen Pract
1989; 39: 193-95.
17. Thornton H, Baum M. Should a mammographic screening programme carry the
warning: screening can damage your health!? Br J Cancer 1999; 79: 691-92.
18. Brewer NT, Salz T, Lillie SE. Systematic review: the long-term effects of falsepositive mammograms. Ann Intern Med 2007; 146: 502-10
19. Alberdi E, Povyakalo AA, Strigini L, Ayton P, Hartswood M, Procter R et al. Use
of computer-aided detection (CAD) tools in screening mammography: a
multidisciplinary investigation. Brit J Radiol 2005; 78(Spec No 1): S31-40.
20. Taylor P, Potts HW. Computer aids and human second reading as interventions
in screening mammography: two reviews to compare effects on cancer
detection and recall rate. Eur J Cancer 2008; 44:798–807.
21. Noble M, Bruening W, Uhl S, Schoelles K. Computer-aided detection
mammography for breast cancer screening: systematic review and metaanalysis. Arch Gynecol Obstet 2009; 279: 881-90.
22. Gilbert FJ, Astley SM, Gillan MG, Agbaje OF, Wallis MG, James J, Boggis CR,
Duffy SW; CADET II Group. Single reading with computer-aided detection for
screening mammography. N Engl J Med 2008; 359: 1675-84
23. James JJ, Gilbert FJ, Wallis MG, Gillan MG, Astley SM, Boggis CR, et al.
Mammographic features of breast cancers at single reading with computeraided detection and at double reading in a large multicenter prospective trial of
computer-aided detection: CADET II. Radiology 2010; 256: 379-86.
24. Kim SJ, Moon WK, Kim SY, Chang JM, Kim SM, Cho N. Comparison of two
software versions of a commercially available computer-aided detection (CAD)
system for detecting breast cancer. Acta Radiol 2010; 51: 482-90.
256
25. Catwell L, Sheikh A. Information technology (IT) system users must be allowed
to decide on the future direction of major national IT initiatives. But the task of
redistributing power equally amongst stakeholders will not be an easy one.
Inform Prim Care 2009; 17: 1-4.
257
Chapter 10
ePrescribing
Summary
•
There is now a growing body of data indicating the high incidence of
medication errors that occur in a range of clinical settings; whilst the
majority of these errors are relatively minor, some will translate into
morbidity and, in a minority of cases, death. Many of these errors are
now believed to be preventable.
•
A variety of ePrescribing systems have been developed as information
technology-enabled responses to minimising of the risk of prescribingrelated harm and/or improving the organisational efficiency of healthcare
practices in relation to prescribing.
•
These ePrescribing initiatives range from support for prescribers on
placing medication orders and prescribing decision to the broader more
visionary
perspectives
of
cross-organisational
integration
often
advocated in key policy documents. As technology has advanced, the
scope of ePrescribing has also expanded and charting the evolution of
definitions shows the emergence of a progressively more complex
picture. In essence, however, this term embraces both the simpler
computerised physician (or provider) order entry systems and the more
sophisticated computerised decision support system functionality.
•
There is evidence that practitioner performance and surrogate
prescribing outcomes can be improved through ePrescribing. Positive
evidence on safer prescribing outcomes has tended to be reported in the
more recent studies. However, overall the evidence showing improved
prescribing safety has not been shown to lead to reduced patient
morbidity and/or death.
•
Evidence of benefits from ePrescribing applications has in the main been
derived from evaluations of “home-grown” applications from a few
centres of excellence in the United States. Most applications in use are,
however, commercially procured and typically lack the sophistication of
the more tailored home-grown systems.
258
•
Poorly
designed
applications
and
a
failure
to
appreciate
the
organisational implications associated with their introduction may
introduce unexpected new risks to patient safety and the evidence from
evaluations of these home-grown systems is therefore not easily
transferable to settings implementing commercial systems.
•
The persistent high rates of over-riding of alerts generated by the more
advanced ePrescribing systems remains a major concern; finding ways
of increasing the sensitivity and perceived relevance of alerts is a major
issue that warrants further investigation.
•
This will, amongst other things, necessitate undertaking detailed
medico-legal work to more accurately quantify the risks to system
developers of changing from the current defensive practice in which they
take a “belts and braces” approach to generating warnings to one in
which there is more selective warning of major risks.
•
There is also a need to investigate–in a few carefully selected contexts–
the role of “hard-stop” restrictions, which prevent the over-riding of
alerts.
•
Given that electronic health record systems are now being introduced
into NHS hospitals in England, there is a need to consider introducing
ePrescribing systems, preferably in an evaluative context that allows the
effectiveness and cost-effectiveness of these new systems to be
established.
10.1 Introduction
There is now a considerable body of evidence demonstrating that prescribing
errors are common and that these are responsible for considerable–
potentially avoidable–patient morbidity and mortality.1 2 For example, recent
UK data indicate that medication-related harm is frequently implicated in
admission to hospitals3 and furthermore that an estimated one in seven
hospitalised patients experience one or more episodes of prescribing-related
259
harm.4 Many studies have now demonstrated that a large proportion of these
adverse drug reactions (ADRs) are potentially preventable.5 6
Given the vast array of drugs now available and the considerable scope for
their interaction with aspects of the patients’ history and/or other coprescribed treatments, it is simply no longer feasible for clinicians to know
about, retain and judiciously draw on all such information from memory.
Electronic prescribing (henceforth referred to as “ePrescribing”) has the
potential to support professionals by helping them to identify and select
potentially appropriate treatments and doses, and also by using patient
specific and other local data to guide treatment decisions. In this chapter, we
review the potential and empirically demonstrated benefits and risks
associated with ePrescribing, building on the more generic discussions on
computerised decision support systems (CDSS) in Chapter 8.
A more
focused critique of the literature on the potential of information technology (IT)
to support prescribing in two exemplary particularly high risk contexts (i.e. the
use of oral anticoagulants in primary care and approaches to minimising risk
of repeat drug allergy in hospitals) is presented in the case studies in the
following two chapters.
10.2.1 Definition
There is no agreed definition of ePrescribing. For example, Dobrev et al. have
defined ePrescribing as “the use of computing devices to enter, modify,
review, and output or communicate drug prescriptions”.7 In contrast, the
definition used by NHS Connecting for Health (NHS CFH) is somewhat
broader, including aspects of the governance of prescribing decisions i.e.
“utilisation of electronic systems to facilitate and enhance the communication
of a prescription or medicine order, aiding the choice, administration and
supply of a medicine through knowledge and decision support and providing a
robust audit trail for the entire medicines use process”.8 This definition
embraces the use of technology to support the whole process of medication
management and it also implies the integration of medication systems with
existing electronic health record (EHR) systems (see Chapter 3). The
260
taxonomy of ePrescribing systems proposed by Teich et al. emphasises the
importance of integration with EHRs (see Figure 10.1). 9
Figure 10.1. The level of sophistication of ePrescribing systems.
Modified based on “Teich et al.” (2004)9 (permission obtained)
Also of relevance is the degree of support that prescribers are offered. Many
of the initially developed systems were, for example, computerised physician
(or provider) order entry (CPOE) systems that provided clinicians with dropdown menus to support the prescribing decisions that were being made.
More recently, however, the focus of developers has been to build on this
basic prescribing support and offer prescribers functionality that takes into
account other relevant contextual information about the patient using in-built
decision support (see Chapter 8 for a more general discussion about
computerised decision support systems (CDSSs). Table 10.1 provides a
framework for considering the degree of decision support offered. It should be
noted that both CPOE and CDSS can be used in other contexts, in particular
the ordering of investigations; these other uses will not however be
considered further in the context of this chapter.
261
Level 1
Standalone electronic prescription writer or CPOE
Level 2
Electronic drug reference manual
Level 3
Electronic
prescription
writing
and
electronic
transmission
of
prescriptions—connectivity to dispensing site
Level 4
Patient specific prescription creation or refilling
Level 5
Basic decision support functionality (integrated or interfaced)— dosage
(default and frequencies) and formulary support
Level 6
Drug management—access to electronic medication administration record
(eMAR) checks for allergies, drug interactions and duplicate therapies
Level 7
Integration with an EHR
Level 8
Integration with EHR and other clinical information systems (radiology,
laboratory and pharmacy information systems)
Level 9
Advanced decision support functionality (integrated or interfaced): adjusting
dosages in light of patient characteristics (e.g. ethnicity), physiologic status
(e.g. uraemia) and co-morbidities; other medications currently being taken;
previous response to the drug, single, daily and life dose limits
Table 10.1 Levels of system sophistication
As is evident from the above discussion, the term “ePrescribing” thus
encompasses a wide range of systems, these including both CPOE and
CDSS and varying degrees of integration with other electronic record
systems. In the absence of any agreed definition, ePrescribing is used in this
chapter as an inclusive term referring to at minimum the electronic generation
of prescriptions, but which may include point-of-care (POC) decision support
and, amongst other things, electronic communication of the prescription
information to other professionals and agencies involved with medicines
management.
Prescribing is a complex organisational practice, including a range of
processes spread across locations and involving diverse actors, so it is
unsurprising that that ePrescribing systems are also organisationally complex;
the choices available in their implementation and dimensions that can be
included in their evaluation are hence also multifarious. Figure 10.2 depicts
262
the complexity of ePrescribing processes. It shows how ePrescribing can
involve different healthcare professionals at different points of prescribing
procedures and how these may require professionals to have access to
patients’ healthcare or medical records to prescribe appropriately. Medication
errors can occur at any point in the prescribing processes i.e. prescribing,
transmitting the order, dispensing, administration and monitoring, and
ePrescribing systems can therefore potentially support any of these functions.
10
Figure 10.2 High-level ePrescribing architecture
Source: Bell et al.11 (permission obtained)
ePrescribing systems have been developed for use in a range of healthcare
settings, these ranging from primary care to hospital-based care. This issue
of setting of use is emphasised in the taxonomy developed by Cornford et
al.12 i.e.:
•
Pharmacy-based systems
•
Clinical specialty-based systems (e.g. those used in oncology, renal
medicine and intensive care)
•
Components or modules of larger hospital information system
packages
•
Home-grown software.
263
This framework also points to the importance of the genealogy of the system
i.e. whether it is home-grown or commercially procured, this also having been
highlighted by several other experts (discussed below).
The term ePrescribing may therefore include systems with a range of
functions and implemented in a wide range or organisational contexts. When
critically reviewing the evidence on the impacts of ePrescribing systems it is
important to be sensitive to this variation, as the benefits and risks may be
influenced by the system functionality and implementation context. By treating
all implementations as being commensurate and aggregating evidence across
heterogeneous systems there is a risk that significant insights into system
design and implementation will be overlooked.
It follows from the discussion above that there are four dimensions that
particularly need to be considered:
1. Interoperability (stand-alone vs. integrated with other health information
systems)
2. Functionality (CPOE vs. CDSS-based)
3. The degree of customisation (home grown vs. commercially procured)
4. Setting of use (e.g. primary care vs. secondary care).
These four dimensions are depicted in Figure 10.3.
264
1. Interoperability:
Stand-alone vs. integrated
3. The degree of customisation:
Home-grown vs. commercial
2. Functionality:
Basic support vs. advanced
support/ alerts
4. Setting of use:
Primary care (including in-general,
intensive care units, pediatric care)
vs. Secondary care (including
ambulatory and pharmacy settings)
Figure 10.3 Four dimensional taxonomy of technological aspects of
ePrescribing systems to understand local implementation landscape
Dimension 1: Degree of inter-operability and integration
This refers to the degree of integration of ePrescribing systems into other
healthcare information technologies such as EHRs. ePrescribing systems
may be modules within integrated IT systems, linking them to other functional
systems, including patient records, accounting systems or inventory systems,
or they may be stand-alone systems, with little or no integration with the data
held on other systems. Some studies suggest that increasing integration with
other systems is likely to be associated with the realisation of greater
operational and other benefits.13 15
Dimension 2: Degree of decision support
This relates to the extent of decision-making support embedded in the
systems. An ePrescribing system may simply automate aspects of the preexisting paper-based system, but ePrescribing systems can also alert
prescribers and pharmacists to prescribing decisions that break rules
embodied in the systems. The categorisation of tiers of alerts fired, whether
the alerts warn or block decisions, the rule-setting for alert messages, the
extent to which decision support rules draw on individual patient records and
the creation of knowledge-bases for prescription, may vary across countries,
healthcare organisations, specialities and indeed clinical teams. Contextual
265
factors, such as patients’ health care records, drug of medical history,
ethnicity, sex, age, weight and local agreements (as embodied for example in
practice or hospital-based formularies) add further complexity to the decision
making of clinicians. The growing functional complexity of ePrescribing
systems is therefore likely to be correlated with their increasing integration
with EHR and other local information systems.
Dimension 3: System development
This refers to the genealogy of the ePrescribing platform–i.e. whether it is a
bespoke locally developed (or “home-grown”) system or a standardised
commercial package. Locally developed systems may, once mature,
demonstrate increased benefits because the systems are developed to fit with
idiosyncratic working practices and also because clinicians tend to more
tolerant of the shortcomings of a system that they have, even in a small way,
contributed to the development of. However, due to the cost of maintaining
local IT development resources and the cost of developing bespoke solutions,
the trend across the IT sector has move away from bespoke development and
become more dependent on standardised packages from major commercial
suppliers. This axis is a continuum because systems may initially have been
developed for local implementation, but their kernel then forms the basis for a
commercial package that is sold onto users elsewhere. This is often
witnessed in the major software development processes15 (see Box 10.1 as
an example of this). Furthermore, this is so because commercial systems
allow differing degrees of configurability, ranging from none or very little to
substantial. Rothschild points out the limitation this creates in the current
ePrescribing literature16 as there is possible limited generalisability of findings
from studies focusing on locally developed ePrescribing systems, rather than
“off-the-shelf” commercial projects that are more commonly found in nonresearch settings.17 Most early adopted ePrescribing systems were homegrown while commercial systems tend to be seen as more rigid and lacking
the adoptability to meet individual organisational needs.
266
Box 10.1 Biography of a hospital ePrescribing system: evolution from
bespoke system to commercial package
The Prescribing Information and Communication System (PICS) is a portable rulesbased CDSS developed by staff at the University Hospitals Birmingham NHS
Foundation Trust and is now available on the market following an implementation
partnership agreement with CSE Healthcare systems. It is currently used for inpatients, but is being developed for outpatient and ambulatory care settings.
PICS provides ePrescribing and medication functions supported by various patient
management services, including laboratory/ radiology ordering and results
reporting. For prescriptions, it can provide advanced decision support to check for
drug interactions, contra-indications, dose limits and other clinical needs linked
with prescribing. Advanced functions include an infection control module, druglaboratory warnings/ prescription-driven test requests, and thrombosis risk
assessment and management. The decision support was developed in partnership
with the British National Formulary (BNF) so that the drug dictionary is modifiable
into a resource with widespread applicability. A number of system interfaces have
been developed for users, from simple “pop-up” windows to bi-directional mobile
email communications. The system uses standards to ensure interoperability with
wider record systems.
Dimension 4: Setting
The setting of deployment can have obvious consequences in relation to the
types and frequencies of errors that might be avoided.
Systems may, for
example, be implemented in small-scale organisational settings such as a
single general practice or a ward or intensive care unit or may be
implemented across a group of care providers in, for example, a geographical
region. The more grand-scale plans for ultimate national coverage across
both primary and secondary care settings envisioned by the National
Programme for Information Technology (NPfIT) (see Chapter 3) is in many
ways unique.
These four factors–i.e. “stand-alone vs. integrated”, “basic vs. advanced”,
“home-grown vs. commercial” and “setting”–mutually shape the ways in which
267
ePrescribing systems are implemented and appropriated, and this in turn may
affect the quality, safety and efficiency of care. As will be clear from the
discussion above, these dimensions are not necessarily mutually exclusive
(see, for example, the description in Box 10.1), nonetheless examining the
systematic review papers through the lens of this four-axis typology is
potentially useful as it can help to interpret the at-times conflicting body of
evidence.
10.2.3 Scope for use
Accurately estimating the incidence of prescribing errors is complicated by the
various definitions used and also be range of approaches to detect and
measure errors. More importantly, however, there is a need to appreciate the
degree of harm that these result in, this being particularly relevant given the
repeated observation that many errors are relatively minor and do not
necessarily translate into patient harm. A key question therefore is whether
ePrescribing systems can improve the safety of care by reducing risk of errors
that are particularly associated with risk of avoidable harm. If so, there are
then related important follow-on questions of whether any particular type of
ePrescribing system (see discussion above) is any more effective or costeffective than the other types.
Medication errors are now known to occur at any point in the medicines
management process. As discussed above, depending on the type of
ePrescribing system used, any of these prescribing functions may to varying
degrees be supported by the technology.
Medication prescribing and
administration are however the two areas of the delivery process with the
highest incidence of error and ePrescribing systems are thus potentially
particularly effective in supporting the task of generating and issuing
prescriptions (see also Chapter 8 and Bates et al.18).
Prescribing of medication is a high volume and high cost activity, with costs of
medication in the same group sometimes varying several-fold. There are
hence considerable cost savings to be achieved by equipping prescribers with
relevant information about the effectiveness, costs and relative cost268
effectiveness of different medications, hence this is another important
potential use of ePrescribing systems.
In the UK context, ePrescribing systems of varying degrees of sophistication
are now routinely used throughout primary care. A key question therefore is
to establish whether their use should be extended into hospital in- and outpatient settings.
10.3.1 Benefits
Quality of care
The two generic domains of eHealth that are in theory supported by
ePrescribing are the storing and managing of data, this support being
provided irrespective of the level of functionality of the ePrescribing system,
and the informing and supporting of decisions when applications have
decision support capabilities (see Chapter 4). Box 10.3 details the range of
potential benefits associated with ePrescribing systems. A number of more
specific claims are made by NHS CFH on their website.19
269
Box 10.3 Main potential benefits of ePrescribing applications on healthcare
quality
Potential benefits to patients/carers:
•
Reduced under- and over-prescribing
Professionals:
•
Standardisation of prescribing practices via the provision of guidelines
•
Improved communication amongst prescribers and dispensers (e.g. call
back queries, instant reporting that item is out of stock, alerts for unfilled,
prescriptions or those that have not been renewed)
•
Instant provision of information about formulary-based drug coverage
including on-formulary alternatives and co-pay information
•
Data available for immediate analysis including post-marketing reporting,
drug utilisation review, etc
Healthcare systems/organisations:
•
Reduction in lost orders
•
Shorter process turn-around time such as the transit time to dispensing
site, time until first dose, prescription renewal or refill
•
Generation of economic savings by linking to algorithms emphasising
(offering as a first choice when a drug is selected) cost-effective drugs
270
Patient safety
Although healthcare quality and patient safety are inextricably inter-linked
(see Chapter 4), much of the premise underpinning the use of ePrescribing
relates in particular to improving the safety of medicines management. Errors
related to medicines management are probably the most prevalent type of
medical error in both primary and secondary care within the UK. Of all types of
medicines management errors—prescribing, dispensing, administration,
monitoring, repeat prescribing—errors in prescribing decisions are typically
the most serious.18
ePrescribing applications should facilitate improved communication between
healthcare providers, patient identification, and improved decision and safety
support.
Improved communication is an inherent benefit of ePrescribing.
Improved identification is on the other hand dependent on whether the system
is integrated with other clinical information systems such as an EHR (see
Chapter 6).
Improved decision and safety support is in turn dependent on
how alerts are configured and whether decision support is integrated (see
Chapter 8); again, the degree to which this is improved is also likely to be
dependent on integration with other clinical information systems.
Most notably, ePrescribing has the potential to improve patient safety by
decreasing errors in prescribing, monitoring and repeat prescribing.
The
reduction in these types of errors is clearly potentially dependent on the level
of system sophistication, i.e. the degree to which the system is integrated with
patient data and decision tools such as drug ontologies and the degree to
which it is configured (customised) to the needs of individual prescribers.
Table 10.1 (above) provides a schematic framework of the extent to which
different applications are likely to improve prescribing safety. The types of
drug errors potentially mitigated relative to the level of ePrescribing
applications’ sophistication include:
•
Miscommunication of drug orders: due to poor handwriting,
confusion between drugs with similar names, misuse of zeroes and
271
decimal points, confusion of metric and other dosing units and
inappropriate abbreviations (Levels 1, 2, 3 and 7)
•
Inappropriate drug(s) selection: due to incomplete patient data such
as contraindications, drug interactions, known allergies, current and
previous diagnoses, current and previous therapies, test results etc
(Levels 4, 5, 6, 8 and 9)
•
Miscalculation of drug dosage: incorrect selection of route of
administration; mistakes with frequency or infusion rate (Levels 2 and
5)
•
Out-of-date drug information: for example, in references to alerts,
warnings etc or information on newly approved drugs (Levels 2 and 6)
•
Monitoring failures: results of laboratory test monitoring and drug
administration monitoring not being taken into account (Levels 6, 7 and
9)
•
Inappropriate drug(s) selection: due to clinical incompetence (Level
9).
The use of ePrescribing facilitates identification of the prescribing clinician and
the date of prescription thereby allowing quality control measures to be
targeted at specific clinicians. It is also possible to configure a system so that
it will not process certain orders that are considered dangerous, for instance
the accidental prescribing of oral methotrexate for daily use when the
intended prescription is for weekly use.20 Additionally, the applications are
capable of linking to other clinical information systems for ADE monitoring and
reporting21 and electronic-based representations of prescriptions can form the
basis for additional safety measures related to dispensing and administration
errors (e.g. automatic dispensing machines and bar-coding of drugs to ensure
that patients receive the ordered drug in the correct dose at the specified time;
see Chapter 12 for a fuller discussion of this issue in relation to approaches to
minimise the risk of recurrent drug allergy).
272
Improved efficiency
The more integrated systems should also in theory offer advantages in
relation to the provision of drawing on data from a variety of sources and
hence offer the potential for more advanced decision support functionality;
they may furthermore also increase the efficiency of prescribing by, for
example,
reducing
time
to
dispensing
through
better
end-to-end
communication in hospitals between wards and pharmacy.
10.3.2 Risks
Patient safety
How is the safety of these applications ensured? In the United States (US),
the Food and Drug Administration (FDA) has classified medical software as a
medical device since 1976 and therefore requires proof of software
verification by demonstrating consistency, completeness and correctness of
the software at each stage of the development life cycle.
For the following
three types of medical software, proof of software validation is also
determined by the correctness of the final software product with respect to the
users’ needs and requirements:22
•
Software as an accessory
•
Software as a component or part
•
Stand-alone software.
ePrescribing software is however exempted if it is ‘…intended to involve
competent human intervention before any impact on human health occurs’.22
In the UK, the Medicines and Healthcare Products Regulatory Agency
(MHRA; the UK FDA equivalent) does not consider medical software to be a
medical device and therefore does not undertake quality assurance activities.
In recognition of this regulative deficit, NHS CFH created a mechanism based
on other safety critical software industries’ guidance for medical software
products. This quality assessment and assurance however only applies to
products developed for NHS CFH and no regulatory paradigm exists in either
the US or the UK for commercially available medical software products, these
being excluded by the “competent human intervention” clause.
273
This issue is important because although the use of ePrescribing applications
for the ordering of drugs should in theory reduce the burden of some types of
drug errors, these applications might also introduce new errors. These errors
in system design and oversights in development might lead to:23
•
Incorrect decision support provided → incorrect medicines ordered and
administered → e-Iatrogenesis.
Theoretically, risks to patient safety by ePrescribing applications could occur
at any point in the use of applications due to errors made by the end-user,
such as:
•
Incorrect patient data input → incorrect decision support → incorrect
medicines ordered and administered → e-Iatrogenesis
•
Incorrect orders selected →
incorrect medicines ordered and
administered→ e-Iatrogenesis
•
Incorrect patient selected → inappropriate medicines ordered and
administered → e-Iatrogenesis.
Dependence on the support provided by the application can furthermore put
patients’ safety at risk when support is not available as, for example, when
general practitioners (GPs) prescribe in the context of home visits or hospital
doctors change practices or hospitals. Similarly, not understanding the nature
of the support provided, such as its limitations, can lead prescribers to
misjudge the robustness of the support provided. In contrast, ignoring the
advice generated may also threaten patient safety.
Organisational inefficiency
Although the use of ePrescribing is intended to improve the quality of
healthcare processes by reducing complexity, the complexity of care often
increases as a result of the incorporation of technology into health service
delivery. This is primarily due to the significant process changes associated
with ePrescribing implementation.
Implementing ePrescribing applications
274
may therefore inadvertently impact on the efficiency of care by, for example,
resulting in:24
• New or additional work
• New training needs
• Negative emotions/perceptions
• Unwelcome changes to workflows
• Parallel use of electronic and paper-based systems
• Changes in relationships and/or power dynamics
• Time-inefficiencies
• Costs.
We identified 32 reports13 16 25-52 121 122 of 30 systematic reviews (SR) 13 16 25-52
on the benefits and risks associated with ePrescribing systems. Two articles
121 122
are updates of previous SR studies30 31. A detailed description of these
studies is presented in Appendices 4 and 5; here we consider the over-riding
messages to emerge from this body of work.
As an overview of the SR study, evidence on patient outcomes by
ePrescribing has been reported by some of the SRs,13 16 25 28 30 31 35 39 40 48-51
121 122
but not all of them. Therefore, the evidence on reported impacts on
patient outcomes is limited. For example, no impact on mortality was reported
in any included study, but there are some studies demonstrating a limited
impact on actual ADEs. A greater effect on potential ADEs and/or serious
medication errors (MEs) is reported by many of the SR studies.28 31 32 35 36 39 44
46 49 50 52 122
Some SR studies30 33 34 37 44 46 reported projected cost savings
from ePrescribing by extrapolating from the reductions in MEs, prescription
dosages and patients’ hospital stays through the use of ePrescribing, but no
evidence of direct cost-effectiveness was presented. Furthermore, the
majority of studies used weak experimental designs which expose them to the
risk of bias.
275
10.4.1 Empirically demonstrated benefits
Many of these reviews focused on CPOE and CDSS supporting prescribing.13
16 25 28 30-34 39 44 46 51 121 122
Some of the SRs focused on CDSS for
prescription.35 36 47-49 52 We detail most reviews below, omitting those where
there is duplication of studies (and conclusions) with other reviews discussed
either here or in Chapter 8 on CDSS.
Impact on patient safety: mortality, morbidity and surrogate marker of
medication errors
Sub-standard prescribing practices, such as inappropriate drug selection due
to allergies or contraindications and incorrect dosing, are frequently evaluated
outcomes. Differences in the way errors are defined and measured make
generalising across organisational settings difficult.
and Metzger, citing Nebeker et al., write that:
For example, Classen
53
‘One of the ongoing controversies in medication safety is how to measure
the safety of the medication system reliably and how to assess the effect of
interventions designed to improve the safety of medication use. Clearly,
common nomenclature, definitions, and an overall taxonomy for medication
safety are essential to this undertaking and the lack thereof has significantly
hampered the comparison of various medication safety interventions among
different centres.
54
At the heart of an even more fundamental controversy is
whether the focus of patient safety should be on errors or adverse events as
a means of assessing and improving the safety of the healthcare system.’
Several SR papers
16 25 30 46 51 122
demonstrated limited evidence on the
reduction of ADEs through the use of ePrescribing systems. For example,
Wolfstadt, et al.51 conducted a SR evaluating ten studies to examine the effect
of CPOE with CDSS functionality on a range of ADEs in a range of clinical
settings and found that half of the studies found a significant reduction in
ADEs. None of the studies however employed randomised controlled trial
(RCT) designs, and seven of the 10 studies evaluated home-grown systems.
The weak study designs and heterogeneity of patient settings, outcome
276
measures and system genealogies precluded any definitive overall conclusion
on effectiveness. There was no discernible impact on other important
outcomes such as death.
The SR of 27 studies by Ammenweth et al.25 highlighted the complexity of
interpreting this body of evidence. This SR, which reported at about the same
time, included studies evaluating CPOE systems both without and with CDSS
of varying degrees of complexity. The authors included studies that employed
controlled and before-after designs undertaken in a range of in-patient
settings and evaluating both home-grown and commercial systems. Overall,
this evidence showed that 23/25 studies reported reduced rates of medication
errors, the effect size ranging from 13-99%. Furthermore, six of the eight
studies reporting on potential ADEs found a reduction in the incidence of this
outcome (effect size 35-98%). More importantly, two-thirds of the six studies
reporting an actual ADE also found a significant reduction of similar size
magnitude for potential ADEs. Expressed in another way, however, only four
of these 27 studies–which employed study designs that rendered it difficult to,
in the authors’ words, ‘exclude a major source of bias’ –found a positive
impact on clinical endpoints. To their credit, however, the authors undertook a
range of subgroup analysis on the “ME” surrogate outcome: this revealed that
systems that were home grown and had advanced CDSS functionality were
more likely to be effective than commercial and basic CPOE systems. There
were, however, no detectable differences by population studied, inpatient
setting or study design.
Clamp and Keen30 121 conducted an SR adopting Pawson’s Realist Synthesis
approach. Their two reports examined the same 70 studies of healthcare IT
systems with variable designs in a range of settings including general
medical, surgical, intensive care, paediatric, tertiary, acute and subspecialty
renal care settings. Twenty-seven of these studies were concerned with the
evaluation of CPOE systems. Although there is no strong evidence that CPOE
reduces preventable ADE rate, one study reported decrease in preventable
ADEs of 17%
55
while another study showed one ADE would be prevented
every 64 days by the use of CPOE in a paediatric unit, but there was no
277
reporting of statistical significance.56 They also concluded that all MEs were
reduced significantly by the use of COPE by 40-80% with a significant
decrease in serious MEs by 55% 55 and non-serious MEs by 86%.57
Rothschild also assessed ePrescribing in critical care, general inpatient and
paediatric care settings16. Their review of 18 publications reported
improvements in a range of process and surrogate markers. Three studies,
two of which
55 57
have been already presented in the review by Clamp and
30
Keen , reported that the incidence of serious MEs/ADEs were significantly
reduced by the use of CPOE, but the same effect was not identified in
pediatric ICU study.58
Shekelle et al.46 conducted a SR study on the evaluation of costs and benefits
of health IT mainly in US outpatient and inpatient paediatric settings, and in so
doing identified 30 studies in relation to CPOE. The review showed consistent
evidence that CPOE with CDSS has significant potential to reduce harmful
MEs, particularly in inpatient paediatric and neonatal intensive care settings.
Mullett et al.59 found that ePrescribing with CDSSs decreased pharmacist
interventions for erroneous drug doses by 59%.
CPOE (which is not
combined with CDSS) was found to be effective in reducing medication
dosing errors. Potts et al.56 conducted a prospective cohort study to examine
medication prescribing errors and potential ADEs before and after
implementation of a home-grown CPOE system in a paediatric intensive care
unit. They found that the use of home-grown CPOE significantly reduced both
MEs (30.1 to 90.2 %, p<0.001) and potential ADEs (2.2 to 1.3%, p<0.001).56
There are other SR studies which address the effect of COPE on serious in
medication errors regarding safety, but apart from the above four studies,
most of the studies failed to provide the evidence on ADEs or a few can
provide the effect with low statistical power. For example, in a variety of
settings with different types of patient populations, Shamliyan et al.45
systematically studied 12 studies in in-patient settings ranging from adult
primary care, acute care to paediatric/ newborn intensive care unit settings.
The review shows that prescribing errors amongst the majority of the studies
278
(8/10 studies) while they also showed a significant reduction in doing errors
(3/7 studies) and in ADEs (3/8 studies), compared with handwritten orders. It
also reported that one RCT and 5 uncontrolled interventions and four
observational studies demonstrated that the implementation of CPOE is
associated with the reduction of medication errors in both adults and
paediatric patients without providing quantitative estimation of relative risk.
Three studies also suggested that the implementation of CPOE systems had
a positive impact on reducing ADEs, but without providing clear statistical
evidence in support of this conclusion. For example, one study showed that
the use of “CPOE would prevent 9 ADEs per 1,000 prescriptions in
paediatric”56 populations
while another study showed that “12 ADEs per
1,000 prescriptions in adult population” could be prevented.60 On the
reduction in prescribing errors, the evidence shows that CPOE was
associated with a 66% reduction in adults with odds ratio [OR] =0.34; 95%CI
0.22-0.52 and a positive tendency in children (P for interaction =0.028).45
From their wider review of 30 CPOE studies from 32 reports conducted in outpatient settings, Eslami et al. identified four studies evaluating the effect of
ePrescribing on drug safety;35 all four systems had in-built CDSS functionality,
but none of these four studies demonstrated any significant improvement in
ADEs; the potential reasons for this may have varied across studies including
non-use of systems and a small number of events and associated low power.
Eslami et al. concluded that ‘…in spite of the cited merits of enhancing safety
and reducing costs, published evaluation studies do not provide adequate
evidence that ePrescribing applications provide these benefits in outpatient
settings’. 35
Hider systematically reviewed the effectiveness of ePrescribing to improve
practitioner performance and patient outcomes by covering 52 studies in a
range of settings, including primary care, intensive care, inpatient and
outpatient settings and most studies reported that CDSS could improve
practitioner performance, especially for the prescribing of potentially toxic
drugs and that alerts on prescription can reduce ADEs and improve other
patient health outcomes including the risk of renal impairment.39 The inability
279
of some of the studies to find any improvement in health outcomes may be
due to the small sample sizes, but the results from a meta-analysis of the
studies evaluating electronic dose adjustment found that CDSS can reduce
the frequency of adverse reactions and decrease the length of hospital stays.
Schedlbauer et al. conducted a SR in which they identified 20 studies
evaluate the impact of alerts on prescribing behaviour.44 The authors were
found that the majority of alerts resulted in a reduction of MEs, but only a
minority of studies reported on clinical outcomes.
Apart from the reduction of ADEs and MEs, some SR papers reported that
ePrescribing has shortened the length of patients’ hospital stay through more
appropriate and effective prescription. For example, the SR study conducted
by Durieux et al.31 122 assessed the beneficial effects of CPOE with CDSS on
the process or outcome of health care with the focus on drug dosage in
inpatient and outpatient settings by covering 26 comparisons in 23 articles of
majority were RCTs (23 RCTs, 1 CT and 2CCT).31
122
Computerised advice
on drug dosage had the effect to increase the initial dose of drug and serum
drug concentrations, and this led to a more rapid therapeutic control, the
reduction of the risk of toxic drug levels, as a result, shortening the length of
patients’ hospital stay. The length of time spent in hospital was reported in six
papers, finding a significant reduction in hospital stay overall in the computer
group (SMD-0.35, 95%CI 0.52, 0.17).31
122
In one study a reduction in length
of stay was found (SMD -0.04, 95%CI -0.07, -0.01), but Durieux et al. query
the reliability of the confidence intervals due to a potential unit of analysis
error.31 122
Eslami et al. conducted a SR study addressing the characteristics of CDSSs
for tight glycemic control (TGC) and reviewed their effects on the quality of the
TGC process in critically ill patients.32 Most of the CDSSs included in the
studies were stand-alone. All of the controlled studies in Eslami et al.’s review
reported on at least one quality indicator of the blood, but only one study
reported a reduction in the number of hypoglycaemia events.32
280
Mollon et al. also conducted a SR study of 41 papers to evaluate the features
of ePrescription with CDSS for successful implementation, prescribers’
behaviours and changes in patient outcomes.13 The authors identified five
studies
from
37
(12.2%)
“successfully
implemented”
trials
showing
improvement in patient outcomes. It is interesting to note that all the five
studies showing patient outcome improvements were published after 2005,
implying that systems are becoming more effective in mitigating patient
outcome risk.
Yourman et al. assessed systematically the improvement of medication
prescription in older adults, with the focus on CDSS intervention in outpatient
and inpatient settings, mainly in the US.52 A majority of the studies were of
systems providing direct support at the point of care and were not condition or
disease-specific. Of 10 studies testing CDSS interventions for older adults,
eight showed at least modest improvements (median number needed to treat,
33) in prescribing, as measured by minimising drugs to avoid, optimising drug
dosage, or improving prescribing choices in older adults (according to each
study's intervention protocols). The majority of studies reported medicationrelated process outcomes, for which CDSS generally showed positive effects,
including lower rates of prescribing inappropriate drugs and closer adherence
to better drug choices or dosages for older persons. The studies reviewed
indicate that often straightforward point of care recommendations showed
modestly effective results from a process outcomes perspective.
Improved practitioner performance in prescription
Reviews assessing the impact of ePrescribing on the quality of care include
studies focusing on the ordering of prophylactic prescriptions, adherence to
prescribing guidelines and organisational efficiency. The definitions and
measurement of quality outcomes vary, so generalising across organisational
settings is difficult.
Garg et al. included 29 trials of drug dosing and prescribing, with single-drug
dosing improving practitioner performance in 15 (62%) of 24 studies; another
five applications used electronic order entry for multi-drug prescribing with four
281
of these applications improved practitioner performance.36
Nies et al.61
however, assessed the same studies included in the review by Garg et al.36,
but came to a different conclusion, namely that ‘…drug dosage adjustment
was less frequently observed in positive studies (29 per cent) than in negative
studies (71 per cent).’61 Whilst Nies et al. noted that their conclusions differed
to those made by Garg et al. they did interpret why this contradictory finding
was made.61 This discrepancy may have resulted from differences in the
definition of success, but merits further exploration.
Time efficiency and improved work-flows
One of the important themes for organisational implications of the
implementation of ePrescribing systems highlighted in the literature is time
efficiency and improved working practices.
Tan, et al. systematically examined whether the use of CDSS has an effect on
newborn infants’ mortality and morbidity and on healthcare practitioners
performance by assessing three RCTs.49 One study,62 which investigated the
effects of a database programme in aiding the calculation of neonatal drug
dosages found that the length of time for the calculation was significantly
reduced among resident paediatric staff, paediatricians and, to a lesser
extent, for nurses by the use of the Neodosis spreadsheet program, while the
system eliminated serious errors. However, there were insufficient data from
the randomised trials to determine the patient benefit or harm from CDSS in
neonatal care.
Niazkhani systematically reviewed 51 papers covering 45 studies to evaluate
the impact of CPOE on organisational efficiency and, in particular, clinical
workflow.42 Most of the CPOEs studied were commercial and the majority
were in adult inpatient settings in teaching hospitals but also included
pediatric settings. The evidence shows that the implementation of CPOE
resolved many disadvantages associated with the work-flow in paper-based
processes. For example, 11 studies showed that CPOE systems improved
work-flow efficiency by removing many intermediate and time-consuming
tasks for healthcare professionals, while six before-after studies showed a
282
substantial decrease in the drug turnaround time, varying from 23-92%.
Furthermore, three studies found a significant reduction of 24-69% in the time
interval between clinicians’ radiology requests and the completion of the
procedures pre- and post-implementation. The same three studies also
showed that a shorter turnaround time was found for laboratory orders,
varying from 21-50%.
Clamp and Keen also noted that although there was no evidence of reduction
in pharmacists’ time spent dealing with prescriptions, there were changes in
their working patterns.30 The authors argued that pharmacists have an
important quality control role in checking prescriptions, with one study finding
that pharmacists only spent 5-20% of their time on direct clinical care.63
Prescription monitoring and adaptation was reduced to less than 10% in a UK
hospital using ePrescribing, allowing pharmacists to spend around 70 % of
their time on direct patient care.64 In a US study the pharmacists spent 46%
more time on problem-solving activities and 34% less time filling in
prescriptions.65 The authors noted that three studies—including one RCT—
found that the total time for direct and indirect patient care increased due to
the introduction of the ePrescribing system and there was a reduction in
pharmacist interventions for prescriptions.66-68
Evidence for improved
organisational efficiency was also found by Clamp and Keen in their review of
turn-around times.30 Mekhijan et al. found a statistically significant reduction in
turn-around times following the implementation of ePrescribing (64%
reduction; P<0.001).69 Turn-around time from ordering to dispensing was
found to decrease by up to 2.5 hours in a study by Lehman et al.70
Sintchenko et al. conducted an SR study of 24 papers (RCTs) to assess the
importance of the type of clinical decisions and decision-support systems and
the severity of patient presentation on the effectiveness of CDSS use in US
and Europe.48 The study reported that control trials of CDSS indicated greater
effectiveness in hospital settings than when applied to chronic care and CDSS
improved prescribing practice and outcomes for patients with acute
conditions, although CDSS were effective in changing doctors’ performance
or outcomes in primary care.
283
Guideline compliance
Another recurring theme in the literature is guideline compliance. A number of
SRs have demonstrated the improved level of adherence to guidelines.16 33 34
37 40 47
This body of evidence suggests that achieving guideline compliance would
increase cost effectiveness by reducing unnecessary prescriptions and
laboratory tests. For example, Eslami et al.33
34
examined adherence to
guidelines in outpatient and inpatient settings as part of their systematic
reviews of ePrescribing. For outpatient settings, the authors concluded that
there is evidence of the ability of ePrescribing applications to increase
healthcare professionals adherence to guidelines and hypothesised that cost
reduction can be achieved when guidelines are specifically geared towards
this goal.33 The authors based their conclusions on 11 studies evaluating the
impact of ePrescribing with a CDSS on the adherence to a guideline or
another standard. Among these, four studies showed that there was a
significant positive effect on adherence;71-74 two studies showed a positive
effect without reporting on statistical significance;75 76 and five studies did not
find a significant difference between the control and the intervention groups.
Rothschild16 systematically evaluated the effects of CPOE on clinical and
surrogate outcomes in hospitalised patients in both general and critical care
settings, covering 18 papers. The review found that several process outcomes
improved with CPOE, including increased compliance with evidence-based
practices, reductions in unnecessary laboratory tests and cost savings in
pharmaco-therapeutics. Guideline compliance (corollary orders) increased
from 21.9% to 46.3% (P=0.01), but there was no effect on length of stay.77
10.4.2 Empirically demonstrated risks
A main limitation to studies reporting negative consequences associated with
the use of ePrescribing is that they tend to not indicate which of the many
possible mechanisms might have resulted in the adverse effects.
284
Impact on patients
Eslami et al. noted that recent studies suggested that errors, ADEs and even
mortality may have increased after CPOE implementation.34 Van Rosse et
al.50 also address the increase in mortality rates associated with the
introduction of a CPOE system reported by Han et al..80 This study has been
discussed extensively in the literature.81-83
Han et al. describe the most
serious of risks to patient safety, mortality.80 The authors found that the
unadjusted mortality rate increased from three per cent before ePrescribing
implementation to seven per cent after ePrescribing implementation
(P<0.001).
Observed mortality was consistently better than predicted
mortality before ePrescribing implementation, but this association did not
remain
after
ePrescribing
implementation.80
The
Han
et
al.
study
demonstrated that increased mortality can be associated directly with
modifications in standard clinical processes: With the ePrescribing system,
order entry was postponed until after the processing of patient admission.84
Although accurate patient registration is important to patient safety, delaying
care and treatment of severely ill patients due to the new work practices
embedded in computer systems may adversely affect patient outcomes.85
However, van Rosse et al.50 also refer to the study conducted by Del Beccaro
et al.
81
which evaluated the same CPOE system as Han et al.,80 but did not
find a significant change in mortality rates. Only three hours of training were
conducted during the three months before the implementation day.
Ammenwerth et al. compared these studies and noted that there were
important differences in design and implementation strategies.83 Han et al.80
studied CPOE use with a more critically ill and younger patient population
than Del Beccaro et al.81 Furthermore, Han et al. studied outcomes only five
months after CPOE implementation,80 whereas Del Beccaro et al. extended
their post-implementation study period to 13 months.81 The longer study
period of Del Beccaro et al. may have averaged out a potentially higher error
rate in the first few months after CPOE implementation due to a learning
curve effect.81 Keene et al.82 also studied the effect of CPOE introduction in a
critically ill paediatric population with comparable results to those of Del
285
Beccaro et al.81 The study conducted by Keene et al. suggest that most of the
possible factors which led to the increase of mortality after the implementation
cannot be attributed to the CPOE system itself, but rather resulted from the
implementation process.82
Furthermore, Rosenbloom et al. noted that the implementation process for the
application described by Han et al. did not incorporate steps or elements
known to ensure system dependability and usability.84
Bradley et al. has also noted that total error reports increased postimplementation of ePrescribing, but found that the degree of patient harm
related to these errors decreased.86 Furthermore, Shulman et al.60 noted that
the proportion of drug errors fell significantly from seven per cent before
ePrescribing introduction to five per cent thereafter (P<0.05), but that this
occurred against the backdrop of a strong declining linear trend of the
proportion of drug errors over time (P<0.001).60 These authors, however,
reported three important errors intercepted by ePrescribing which could
otherwise have resulted in permanent harm or death; these errors were
identified and then acted upon by pharmacist or nurse intervention60, i.e.:
‘A potentially fatal intercepted error occurred when diamorphine was
prescribed electronically using the pull down menus at a dose of seven
mg/kg instead of seven mg, which could have lead to a 70-fold overdose. In
a separate case, amphotericin 180 mg once daily was prescribed, when
liposomal amphotericin was intended. The doses of these two products are
not interchangeable and the high dose prescribed would have been
nephrotoxic.
In the third case, vancomycin was prescribed one g
intravenously daily to a patient in renal failure, when the appropriate dose
would have been to give one g and then to repeat when the plasma levels
fell below 10 mg/L.
The dose as prescribed would have lead to
nephrotoxicity.’ 60
Koppel et al. conducted a study on drug errors introduced by ePrescribing.
The authors ‘…identified 22 previously unexplored drug error sources that
286
users reported to be facilitated by ePrescribing through their assessment.’87
The sources were grouped as: (1) information errors generated by
fragmentation of data and failure to integrate the hospital’s several computer
and information systems; and (2) human-machine interface flaws reflecting
machine rules that do not correspond to work organisation or usual
behaviours.87 However, this study has been criticised due to the high risk of
bias with respect to their key findings. In response to this study, Bates,88 for
example, notes that:
‘A main limitation of Koppel et al.’s study was that it did not count errors or
adverse events, but instead measured only perceptions of errors, which
may or may not correlate with actual error rates. Furthermore, it did not
count the errors that were prevented. As such, it offers no insight into
whether the error rate was higher or lower with ePrescribing.
Unfortunately, however, the press interpreted the study as suggesting that
ePrescribing increases the drug error rate. While the authors did not
state this, a press release put out by the journal that published the article
did so.’ 88
Risks to patient safety may arise indirectly from application use. For instance,
a survey of UK GPs found that some respondents erroneously believed that
their computers would warn them about potential contraindications or if an
abnormal dose or frequency had been prescribed, highlighting how lack of
knowledge and training in how ePrescribing systems function can
compromise patient safety.89
Risks to patient safety can arise not only from system use but also from a lack
of actual usage undermining the ability of ePrescribing applications to confer
the envisaged benefits to patient safety.
A sub-section of the review by
Eslami et al. looked at system usage, the authors noted that there was wide
variability in the degree of ePrescribing usage.33 Four studies found that of all
prescriptions, 3–90 per cent were entered electronically.66 90-92
287
The SR by Chaudry29 included one study which used a mixed quantitative–
qualitative approach to investigate the possible role of such a system in
facilitating PEs reported that 22 types of ME risks were found to be facilitated
by ePrescribing, relating to two basic causes: fragmentation of data and flaws
in user-system interface.87
Ammenwerth et al. looked at the relative risk reduction on ME and ADE by
CPOE covering 27 studies on ePrescribing mainly in the US in patient care
settings with study designs including before-after studies/time-series analysis
and two RCTs.25 Twenty-three of these studies showed a significant relative
risk reduction for medication errors of 13-99%, but it is also worth highlighting
one exceptional study which looked at the implementation of a commercial
ePrescribing system with advanced CDSS at two study units between 2002
and 2003 for three months, which reported a significant increase of 26% for
the risk of medication errors.93
Negative impact on professionals’ performance and organisational
efficiency
It should be noted that negative impact of ePrescribing systems on healthcare
professionals’ performance and organisational efficiency can result in risks to
patient
safety.
However,
such
negative
evidence
requires
careful
consideration to identify whether these risks are intrinsic to ePrescribing
systems or are a part of the socio- organisational learning processes in the
implementation.
Eslami et al.33 and Eslami et al.34 conducted SR studies to evaluate studies of
CPOE with/without CDSS on several outcome measures in outpatient and
inpatient settings respectively. In outpatient settings, the authors found three
studies65 67 94 (one RCT and two non-RCTs) that reported an increase in the
total time for direct and indirect patient care due to the implementation of the
CPOE system, while three studies (one RCT and two non-RCTs) also
reported an increase in ordering time with the introduction of CPOE.95-97 Two
of the studies94
96
were also assessed by Shekelle et al. who evaluated 30
studies on CPOE as a part of their evaluation of the costs and benefits of
288
health information technologies in various healthcare settings.46 The authors
noted that two studies reported an increase of the clinicians’ time for order
entry using CPOE compared to paper methods, and both studies
demonstrated that CPOE took up slightly more clinician time.46
Poissant et al. reviewed 23 papers on EHRs to evaluate time efficiency of
physicians and nurses and identify factors that may explain efficiency
differences across studies.43 The authors found that the use of central station
desktops for CPOE was inefficient, increasing the work time from 98.1% to
328.6% of physician’s time per working shift (weighted average of CPOEoriented studies, 238.4%).43
Tierney et al. found that interns in the intervention group spent an average of
33 minutes longer (5.5 minutes per patient) during a 10-hour observation
period writing orders than did interns in the control group (P<0.001).79
Another BWH study published by Bates et al. using work study techniques
found that for both medical and surgical house officers, writing orders on the
computer took about twice as long as using the manual method, these
differences being both clinically and statistically significant (P<0.001).78
However, medical house officers recovered nearly half the lost time due
efficiency improvements in other administrative tasks, for example looking
for charts.78 Additionally, a pilot of ePrescribing standards in the US found that
providers noted that ‘…everything interacts with everything’ making for an
overwhelming amount of alerting and therefore additional work.79 Other than
writing orders, one observational study by Almond in the UK found that the
time to complete the ward drug administration rounds doubled for healthcare
assistants.80
Niazkhani conducted a SR study of 45 studies in 51 papers to evaluate the
impact of CPOE on organisational efficiency, in particular clinical workflow.42
The author noted that the implementation of CPOE led to the creation of
difficulties in work-flow mainly due to changes in the structure of preimplementation work and negative evidence was reported on time efficiency.
Five studies showed time inefficiency due to the implementation of CPOE and
289
four out of the five studies reported a significant increase in time. The
proceduralisation of order entry and the structuring of relationships between
actors were also found to be a source of time inefficiency. Difficulties in
choreographing the various actors and a reduction in team-wide discussions
were also found.
The scope for potential clinician process efficiency gains from the introduction
of CPOE will be dependent at least in part on the inefficiency and
thoroughness of the previous paper-based processes, combining retrieval,
viewing of information, data entry, and in many cases, responses to alerts and
reminders. The work practices of nurses have been found to be
proceduralised, but clinicians may follow idiosyncratic practices.98
10.4.3 Implications of technological taxonomy to benefits
We identified some of the SR papers assessed technical functionalities based
on the three different types of taxonomy: “commercial versus home-grown”
systems,25 29 30 42 45 46 “basic versus advanced” CDSS for prescription25 28 30 3234 38 44 45 47 51 52
and “stand-alone versus integrated” systems.25 29 30 36 40 47 The
evaluation from these perspectives is extremely important to obtain the insight
into what kinds of elements make contribution to successful implementation of
ePrescribing systems for the improvement of patient safety and organisational
efficiency. However, Wier et al.106 point out how often authors provided little
information about the application, technical infrastructure, implementation
process or other descriptive data. Such missing information leads to
difficulties of obtaining accurate picture of the implementation sites and
generalisability of evidence from studies. Bearing these issues in mind, this
section demonstrates key evidence of benefits of ePrescribing from the body
of literature.
Commercial versus home-grown systems
Not all of the systematic reviews papers evaluated relevant studies from a
point of technological taxonomy and the evidence thus tends to be presented
without distinguishing between the findings from in-house built and
commercial systems. Some of the papers do however provide some insights
290
into findings when viewed through this lens. There are some SRs reporting
positive results with home-grown systems25
29 30 46
but there is limited
evidence reported regarding the benefits of commercial systems with highquality of studies. However they demonstrate that customising commercial
systems to tailor them to local hospital environments can also bring benefits25
33 43 45 51
.
Clamp and Keen30
121
found that there was no overall evidence that use of
ePrescribing systems reduces the rate of preventable ADE, but pointed out
that one study showed that the internally developed systems, with CDDS of
menu of medications, default doses, range of potential doses, limited drugallergy
checking,
drug-drug-interaction
significantly reduced serious MEs by 55%
and
55
drug-laboratory
checking
as well as an 86% decrease in
dose, frequency, route, substitution and allergies.57 The study also found an
overall decrease in preventable ADEs of 17%.55
Shekelle et al.46 also cite a prospective cohort study investigating impact of a
“home-grown” CPOE on MEs and ADEs in a paediatric intensive care.56 This
study found a significant reduction of both MPEs (30.1 to 90.2%, P<0.001)
and PADEs (2.2 to 1.3%, P<0.001).56 Another study by Cordero et al. in
neonatal intensive care found that a CPOE system could eliminate gentamicin
prescribing errors.99 Also another study which implemented a home-grown
CPOE system with advanced CDSS (including allergy alerts, dose checking,
drug interaction, clinical pathways, patient and place specific dosage,
interfaces with clinical data repository–order related and laboratory alerts)
reported the potential reduction of ADE, that is, the prevention of one ADE
every 64 days by the use of the ePrescribing system in a paediatric setting but
no statistical evidence was provided to support this estimation.56
After 2007 some SRs started to report the effect of commercial systems for
patient safety in parallel to the studies of home-grown systems.25 33 45
In a recent review, Eslami et al. employed the taxonomy of “basic support”
versus “advanced support/alerts” to evaluate the impact of ePrescribing
291
systems on safety; cost and efficiency; adherence to guideline; alerts; time;
and satisfaction, usage, and usability in the outpatient setting while it also
addressed the “stand-alone” versus “integrated systems” and “commercial”
versus “home-grown” dimensions of ePrescribing systems.33 However, the
evidence drawn from the study in relation to each dimension is not clearly
stated and is obfuscated in the analysis. However, it is worth mentioning one
observational study100 which showed “important weaknesses in generating
alerts in four commonly used commercial systems in Britain’s GP offices”.
Those systems were unable to generate “all 18 predefined established alerts
for contraindicated drugs and hazardous drug-drug combinations”.33
Another important literature on this taxonomy is the study by Ammenwearth et
al. who conducted a systematic review of 27 papers on ePrescribing
implementation mainly in the US outpatient, inpatient and intensive care
settings.25
The study included only two RCTs–most of the other studies
employed before-after and in some cases time-series designs. The ratio of
studies looking at commercial systems vs. home-grown systems was
approximately 1: 1, with one study adopting both designs. Their sub-group
analysis of 25 studies comparing reductions in medication errors between
home-grown and commercial systems highlighted a greater risk reduction in
errors with the home-grown systems.25 In spite of these reported results, the
quality of those studies was not high as many of them did not fully specify the
experimental design, did not describe the cohort or state whether the
comparison and intervention groups’ treatment was commensurate and only
two studies were randomised trials.
Setting
Shamliyan et al. reviewed 12 studies published from 1990 to 2005 evaluating
the impact of ePrescribing systems on prescribing errors in in-patient
settings.45 One study56 which implemented in-house developed system in a
20-bed paediatric intensive care unit setting with prospective, intervention
study found 95.9% of overall errors reduction (P<0.05), 99.4% of total
prescribing errors reduction and 88.8% of wrong drug reduction (P=0.07) as
absolute change in rate while the study also found 7.6% (P=0.69) of wrong
292
dose increase. Another study99 examined the implementation of a commercial
system with a retrospective study in a post-natal intensive care unit setting
and found that medication errors to prescribe gentamicin reduction and to
prescribe the wrong dose of gentamicin were eliminated, but with no statistical
significances provided.
Overall, although definitive evidence from systematic reviews of RCTs
comparing home grown and package systems is lacking, the data suggest
that home grown systems are more effective than commercial systems in
reducing prescribing errors. There is however as yet no clear data available
on whether these differences translate into improvements in important patient
outcomes such as death.
ePrescribing systems with “basic” vs. “advanced” decision support
Overall picture of the taxonomy of “basic” versus “advanced” CDSS is
ePrescribing systems with advanced CDSS showed a higher relative risk
reduction compared to those with limited or no decision support.44
47
In
particular, the evidence in the literature reported that the “patient-specific”
alerts improve the quality of prescribing.25 This taxonomy is closely linked to
the other one, “stand-alone” versus “integrated” systems. The following
paragraphs refer to the key literature with CDSS elements of “basic” versus
“advanced” decision support.
The SR study by Schedlbauer et al. provides important evidence in relation to
the typology of DSS/alerting systems and reminding systems in relation to
ePrescribing systems.44 The study focused on the effects of those alerts and
reminders on prescribing behaviour mainly in the US inpatient settings
covering the relevant papers published between 1994 and 2007.
Twenty
studies, which have employed randomised and quasi-experimental designs
were included.
Categories of drug alerts comprise basic drug alerts,
advanced alerts and complex alert systems (representing a set of CDSSs
containing features of both basic and advanced alerts). Two papers in their
study investigated the effects of four types of basic alerts, of which three
reported statistically significant beneficial effects on prescribing. Drug allergy
293
warnings decreased allergy error events by 56% (P=0.009).55 It also found
that providing default dosing via basic medication order guidance alerts
resulted in reduced dose errors in two studies of 23% (P=0.02)55 and 71%
(P=0.0013).101
Regarding medication errors, the 40% reduction in error
events achieved by drug-drug interaction warnings did not reach statistical
significance (P=0.89).55 The SR study confirms that advanced alert types
tend to provide positive effects across the five categories, saying that all the
20 papers evaluated more advanced alert types and statistically significant
effects were shown in 21 out of 23 types across five categories.
Shiffman et al. looked at the impact of CDSS on practitioner performance,
patient outcomes and satisfaction, with the focus on functionality and the
effectiveness of the systems.47 The authors studied 25 RCTs, CT and TS
which were published between 1992 and 1998. The SR study was conducted
using a technological taxonomy, dividing studies between stand-alone and
integrated systems and between basic and advanced DSS.
All of the systems displayed relevant patient data, a menu of drugs and a
choice of doses. Half of the studies used a system with advanced CDSS
functionality while the other used systems with no or limited/basic CDSS.
Also, in 14 studies with advanced decision support, the risk reduction was
greater than in 11 studies without advanced decision support, but studies
without advanced support were mostly compared to computer-based ordering
whereas those with advanced support were compared to manual procedures.
Stand-alone versus integrated systems
None of the 30 SRs included in this analysis directly address differences
between “stand-alone” and more “integrated” systems. However, some of the
papers do indirectly address this issue offering some insights. We discuss
below the salient findings from these studies.
A study of ADEs found that having ADE detection and reporting capability in
EHR can improve detection of, and potentially reduce ADE because the EHR
system data can be used to identify patients experiencing ADEs.102 A RCT to
294
explore the impact of an EHR with integrated ePrescribing found positive
effects on resource utilisation, provider productivity, and care efficiency.46 95
Eslami et al.32 looked at the characteristics of CDSS for tight glycemic control
(TGC) and the effects on the quality of the TGC process in critically ill patient
by categorising CDSS into the three features: 1) level of support (merely
displaying the protocol chart or suggesting the specific amount of insulin to be
administered); 2) the consultation mode (passive or active); and 3) the
communication style (in the critiquing mode or in the non-critiquing mode).
Most of the CDSS (14 out of 17) were stand-alone and only two papers
studied more integrated system.103
104
One of these studies103 reported a
reduction in the number of hypoglycaemia events, but without assessing
statistical significance.
Interest in the effectiveness of ePrescribing systems continues.105 The
number of systematic review papers on e-Prescribing/CPOE has thus been
growing. However, recent reviews have been inconclusive and shown wide
variations in findings. For example, Wier et al., who conducted a systematic
review of the scientific quality of empirical research on CPOE application,
found that there are areas requiring improvement in research designs and
analyses.106
There is a tendency for empirical studies of CPOE to lack
adequate study designs and blinding, although there are several high-quality
CPOE studies available:
“Current concerns center in the prominent use of pre-post study designs
less
rigorous
measurement
techniques,
failure
to
include
key
information about CPOE and informatics variables, failure to use
blinding and inappropriate statistical analyses. These concerns must be
addressed to allow the field to build a solid foundation of study
generalisability for this area of inquiry in the future.”106
More importantly, implementation strategies significantly varied and this can
lead to confounded results. Apart from the issues of internal validity (e.g.
design type, testing of group differences, instrumentation bias and blinding),
295
construct validity
including types of ordering functions, level of decision
support available, electronic links to other departments, and whether usage is
mandated, or measured, implementation strategies used and length of time
from implementation to measurement of outcomes and statistical validity are
important, but Wier et al. point out how often authors provided little
information about the application, technical infrastructure, implementation
process or other descriptive data.106 Such missing information leads to
difficulties of generalisability of evidence from studies. Bearing these issues in
mind, this section demonstrates key evidence of benefits and risks of
ePrescribing, which are relevant to healthcare quality, patient safety and
organisational issues, obtained from the relevant literature.
10.5.1 System integration
Reviewing this body of work has revealed that ePrescribing systems have
heterogeneous origins, scope and functionality and are furthermore
implemented into diverse organisational settings, all of which may, along with
the context of and approach to implementation, influence the risks and
benefits that result.
As the functionality of these systems extend, these are
more appropriately seen as expert systems rather than data processing tools,
so the integration or “fit” with organisational knowledge is increasingly seen as
important (see Chapter 17). The integration of ePrescribing systems with
EHRs is a technical bridge to allow patient-specific information to be used in
the delivery of patient-centred care and to minimise the risks of MEs and
ADEs as well as improving prescription efficiency.
10.5.2 Knowledge database sharing
In parallel to the integration into EHRs, the integration of CPOE with CDSS is
a logical development, which should be encouraged. The rule bases for
decision support content can either be locally developed or created centrally,
with clear implications for clinician autonomy. However, the open sharing and
consolidation of complex rules on drug and diagnosis interactions across
healthcare communities offers the possibility of further benefits and
296
efficiencies of scale from ePrescribing that may not have been identified in the
more focused studies.107
Clinicians currently benefit from the development of guidelines on specific
conditions, such as the British Thoracic Society’s asthma guidelines108
109
which can be embedded within CPOE. Guidelines gain legitimacy through the
reputation of their sponsoring body, including the National Institute for Health
and Clinical Excellence (NICE) and the UK’s National Service Framework
(NSF).109 Implemented CPOE systems need to be able to be updated to keep
their knowledge bases up-to-date with this evolving body of knowledge and
ideally be able to provide information on the rules being applied to ensure
clinician compliance. Aronson emphasises that it is not guidelines alone that
influence prescribing behaviour, but also the education and financial
incentives to ensure guideline compliance.109
10.5.3 System standards
Interoperability with other healthcare information technology systems is a key
factor for successful implementation of ePrescribing systems, as the systems
are ideally drawing on patient data, updating patient records and integrating
with consumable inventory systems. In practice, interoperability is achieved
through standards, whether de facto local specifications, proprietary
standards of system vendors or conformance with nationally agreed
standards. At the heart of standardisation in ePrescribing, as in most areas of
health informatics, is the EHR, as there is no point in building decision
functionality into a CPOE system that depends upon patient data that are not
accessible. As the market for advanced ePrescribing systems develops, the
functionality offered by vendors will be shaped by the data on standardised
EHRs, pushing vendors to offer functionally similar systems. There is
therefore a need to ensure that EHR standards are extensible to include
future patient data needs to prevent functional lock-in for ePrescribing
systems.
297
10.5.4 Implementation of ePrescribing systems from human factors
CDSS interventions may include alerting and reminder systems. Employing
advanced guidelines is not in itself sufficient to make sure prompts are acted
on, as alerts may be overridden or ignored.110
Consideration of human
factors becomes crucial on this point (see Chapter 16). Human factors can be
categorised into the following four categories: 1) physical and perceptual
factors; 2) cognitive factors; 3) motivational factors; and 4) situational factors.
The reasons repeatedly found for overriding alerts included: alert fatigue,
disagreement; poor presentation; lack of time; knowledge gap.111 112 Procter
et al. have argued that human factor considerations are the key to the
achievement of effective and safe implementation of healthcare systems and
that healthcare professionals’ involvement is crucial in system design and
development.112
A “user-centred perspective can inform system design to ensure that
individual technologies achieve their intended purpose and benefits”113 known
as the human-tech approach.114 In order to achieve robust patient safety, the
micro (user interfaces, ergonomics) the meso (inter-system communication
and integration) and the macro (organisational design) perspectives all need
to be addressed during system design.113 Taking these factors into account
can increase clinicians’ trust and lead to greater system acceptance. By
recognising that ePrescribing systems are fundamentally socio-technical
systems and investing in addressing the human factors during design and
implementation there will be longer-term gains in lower lifecycle staffing and
training costs, reduced risk of errors and greater rule compliance.115
10.5.5 Database for CDSS and data standardisation
As discussed above, to gain the greatest operational gains from ePrescribing
requires the systems to draw on knowledge bases of complex rules which can
be applied to specific patients. To develop parochial rule bases or rule bases
developed by each system vendor is potentially inefficient. Part of the rule
base, for example drug interactions, will be locality-independent, and could be
developed globally. Other aspects, such as rules on recommended
treatments, may be institution specific and would then need local development
298
and ownership. There is therefore a need for processes to maintain the rulebase, carry out assessments of evidence and provide rule legitimacy.
Similarly, there is a policy need for the specification of EHRs to take account
of the needs of current and future ePrescribing systems, which will require
coordination with the emergent rule-base.
10.5.6 Temporal issues regarding the evaluation of ePrescribing
implementation
This overview of the evidence provides very useful insights into the
implementation of ePrescribing systems in a variety of contexts and settings.
However, systematic reviews can also obfuscate in some respects. This is
seen most clearly in the difficulty of addressing the dynamics of the
emergence of a new technology. Almost without exception the reviews are
atemporal, giving all papers within the review period equal weight. However,
Mollon et al. noted that all of the studies they identified showing positive
impacts on patients were from after 2005.13 This suggests, unsurprisingly,
that there may have been a change in the technology through time. It can
tentatively be proposed that there is evidence in the reviews that this is due to
the benefits increasing as the systems take on more advanced decision
support functions and become more integrated with other record systems.
However, it may also partly be that as experience of these systems increases
there is an underlying process of social learning about how these systems can
be implemented and used effectively.
Similarly, the studies generally ignore the short-term dynamics of system
implementation, taking the implicit assumption that the changes observed
shortly after implementation represent the steady-state system performance.
In the summarising of most cases it is unclear how the time period of system
observation related to the period of implementation, but the pressure to carry
out a controlled before-and-after assessment implies that the evidence in
most studies is from a period shortly following implementation. The potential
danger of drawing conclusions about the long-term impacts based on these
snap-shot evaluations is clearest in comparing the coverage in the reviews of
Ammenwerth et al.25 and van Rosse et al.50 of the studies in Han et al.
80
and
299
Del Beccarro et al.
81
which studied the same ePrescribing implementation. It
is reported that where Han et al. in a shorter study found an increase in
mortality, Del Becarro et al. in their longer study did not find the same effect.81
One interpretation of this is that the mortality increase may have been a
transient artefact of the organisational change process and that organisational
learning and adaptation removed the effect. This is an important insight as it
questions the interpretation of the effects identified as significant in many of
the papers covered by the systematic reviews in this synthesis.
10.5.7 Synthesised research methods for the studies of eHealth
It is important to address the limitations of a “systematic review” approach for
the presentation of solid evidence. While the systematic reviews considered
provide valuable insights into the impacts of ePrescribing systems, there is
danger that the importance of factors influencing the impacts of ePrescribing
systems will be underestimated. There is a trend in the area of health care
study that synthesises qualitative and quantitative health evidence.116
117
Greenhalgh’s paper on meta-narrative approach towards systematic literature
review, ‘Tensions and paradoxes in electronic patient record research: a
systematic literature review using the meta-narrative method’ questions the
meaning of ‘rigorous’ research engaging with philosophical debates.118
Lilford’s paper, ‘Evaluating eHealth: How to make evaluation more
methodologically robust’ argues that a mixed research methods approach to
evaluating IT systems in health care is needed, questioning the validity of
evidence obtained by combining formative assessments with summative
ones.119 The central distinction here is between treating the ePrescribing
system as a work-in-progress where it is being recursively shaped by the
studies or whether it is treated as a stable “black-box” with the focus of the
study being to assess its impacts.
Possible areas which are relatively neglected by the use of “systematic
review” and “critical appraisal” methods are:
•
Communication amongst different groups of practitioners (e.g.
clinicians-pharmacists; clinicians-nurses; pharmacists-nurses, patientspharmacists, or multi-groups of people of the above)
300
•
Capture of complex changes in work-organisation /workflows
•
The impact of institutional differences between national healthcare
systems on the shaping of ePrescribing systems and on the outcomes
of ePrescribing trials.
Systematic review tries to treat eHealth technologies as scientific objects not
as social artefacts which are complex and organic in nature and can
potentially lead to unexpected outcomes (see “blackboxing” arguments in the
studies of science and technologies120). The systematic reviews examined
implicitly assume that the results of small-scale trials can be scaled up,
despite evidence that scaling IT systems leads to increasing problems of
accommodating wider practice diversity and less identification of users with
the systems. The use of short-term studies of recent implementations may
overlook the impacts, both positive and negative, on longer term
organisational learning. Finally, while the aggregation of short-term studies
provides evidence on the impact of the systems on operational risks, it is
harder to assess their impact on the risks of rare, but major systemic failure.
10.5.8 Areas for further research
For future research, more sufficiently powered RCTs are needed.16
31 49 122
Such trials are however difficult to mount in this field, and the alternative of
time-series based designs, preferably with contemporaneous control groups,
should also therefore be considered. 106
Furthermore,
research
on
functionality
specific
effects
or
technical
specification effects31 33 34 46 are urgently needed, in particular evaluating the
implementation of commercial systems.46
51
In order to make the evidence
drawn from such studies, standardised reporting for healthcare IT evaluations
is essential.25 32 33 40 46 The evaluation of the risk of MEs and ADEs will also
be more reliable with the use of standardised metrics and reporting.44 45 50 51
Studying healthcare technology is a complicated task and the evaluation of
ePrescribing systems is not an exception. In order to capture a more holistic
301
picture, multi-disciplinary methods are indispensable.25
33 34 42
There are a
number of studies adopting before-after and time-series designs, but more
evaluation of long-term effects is required. Also, evaluations immediately after
implementation need to pay more attention to organisational learning
processes with the focus on learning curve.50 This allows healthcare
professionals to foresee the generalisability of the obtained evidence in their
particular organisational settings. Evaluations in long-term care setting will
also useful to assess the long-term impacts.51
In employing a more multi-disciplinary approach to the evaluation of
ePrescribing systems, the study of how human factors and socio-technical
issues influence the degree of implementation success becomes central.33 34
39 43 44
Also, the analysis of macro effects on collaborative work-flow and
organisational efficiency is important.42 43
Finally, comprehensive economic
evaluation of immediate and long-term effects is also urgently needed.37-40
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Bates DW, Leape LL, Petrycki S. Incidence and preventability of adverse drug
events in hospitalized adults. Journal of General Internal Medicine
1993;8(6):289-94.
Lesar TS, Briceland LL, Delcoure K, Parmalee JC, Mastagornic V, Pohl H.
Medication prescribing errors in a teaching hospital. Jama-Journal of the
American Medical Association 1990;263(17):2329-34.
Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ, et al.
Adverse drug reactions as cause of admission to hospital: prospective analysis
of 18,820 patients. British Medical Journal 2004;329(7456):15-19.
Davies EC, Green CF, Taylor S, Williamson PR, Mottram DR, Pirmohamed M.
Adverse Drug Reactions in Hospital In-Patients: A Prospective Analysis of 3695
Patient-Episodes. Plos One 2009;4(2).
Howard RL, Avery AJ, Slavenburg S, Royal S, Pipe G, Lucassen P, et al.
Which drugs cause preventable admissions to hospital? A systematic review.
British Journal of Clinical Pharmacology 2007;63(2).
Cresswell KM, Fernando B, McKinstry B, Sheikh A. Adverse drug events in the
elderly. British Medical Bulletin 2007;83(1):259-74.
Dobrev A, Stroetmann, K.A., Stroetmann, V.N., Artmann, J.N., Jones, T., and
Hammerschmidt, R. The conceptual framework of interoperable electronic
health record and ePrescribing systems, 2008.
NHS CfH. Patient safety through e-prescribing: E-Health Insider and British
Computer Society Health Informatics Forum, 2008.
Teich JM, Marchibroda JM et. al., Electronic prescribing: toward maximum
value and rapid adoption. Washington: eHealth Initiative, 2004 (170 pp).
Avery AJ, Sheikh A, Hurwitz B, Smeaton L, Chen YF, Howard R, et al. Safer
medicines management in primary care. British Journal of General Practice
2002;52:S17-S22.
302
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Bell DS, Cretin S, Marken RS, Landman AB. A conceptual framework for
evaluating outpatient electronic prescribing systems based on their functional
capabilities. Journal of the American Medical Informatics Association
2004;11(1):60-70.
Cornford T, Fraklin BD, Svage I, Barber N, Jani Y, Electronic Prescribing in
Hospitals: Challenges and Lessons Learning, NHS Connecting for Health (NHS
CFH) 2009;71.
Mollon B, Chong JJR, Holbrook AM, Sung M, Thabane L, Foster G. Features
predicting the success of computerized decision support for prescribing: a
systematic review of randomized controlled trials. Bmc Medical Informatics and
Decision Making 2009;9.
Dobrev A, Jones, T., Stroetmann, V., Stroetmann, K., Vatter, Y., Peng, K.
Interoperable eHealth is Worth it: Securing Benefits from Electronic Health
Records and ePrescribing, 2010.
Pollock N, Williams R, D'Adderio L. Global software and its provenance:
Generification work in the production of organizational software packages.
Social Studies of Science 2007;37(2):254-80.
Rothschild J. Computerized physician order entry in the critical care and
general inpatient setting: A narrative review. Journal of Critical Care
2004;19(4):271-78.
McDonald CJ, Overhage JM, Mamlin BW, Dexter PD, Tierney WM. Physicians,
information technology, and health care systems: A journey, not a destination.
Journal of the American Medical Informatics Association 2004;11(2):121-24.
Bates DW, Cullen DJ, Laird N, Petersen LA, Small SD, Servi D, et al.
Incidence of adverse drug events and potential adverse drug events implications for prevention. Jama-Journal of the American Medical Association
1995;274(1):29-34.
NHS CfH. ePrescribing: Overview and Benefits.
NHS NPSA. IT requirement Specification Safety Alert: Oral Methotrexate 2.5mg
and 10mg Tablets. 2006.
Bates DW, Evans RS, Murff H, Stetson PD, Pizziferri L, Hripcsak G. Detecting
adverse events using information technology. Journal of the American Medical
Informatics Association 2003;10(2):115-28.
Wolinsky H. Resolution due in medical software regulation. Annals of Internal
Medicine 1997;127(10):953-54.
Caudill-Slosberg M, Weeks WB. Case study: Identifying potential problems at
the human/technical interface in complex clinical systems. American Journal of
Medical Quality 2005;20(6):353-57.
Campbell EM, Sittig DF, Ash JS, Guappone KP, Dykstra RH. Types of
unintended consequences related to computerized provider order entry. Journal
of the American Medical Informatics Association 2006;13(5):547-56.
Ammenwerth E, Schnell-Inderst P, Machan C, Siebert U. The effect of
electronic prescribing on medication errors and adverse drug events: A
systematic review. Journal of the American Medical Informatics Association
2008;15(5):585-600.
Ammenwerth E, Schnell-Inderst P, Siebert U. Vision and challenges of
Evidence-Based Health Informatics: A case study of a CPOE meta-analysis.
International Journal of Medical Informatics 2010;79(4):E83-E88.
Bassi J, Lau F, Bardal S. Use of Information Technology in Medication
Reconciliation: A Scoping Review. Annals of Pharmacotherapy 2010;44(5):88597.
Chatellier G, Colombet I, Degoulet P. An overview of the effect of computerassisted management of anticoagulant therapy on the quality of
anticoagulation. International Journal of Medical Informatics 1998;49(3):311-20.
303
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
Chaudhry B, Wang J, Wu SY, Maglione M, Mojica W, Roth E, et al. Systematic
review: Impact of health information technology on quality, efficiency, and costs
of medical care. Annals of Internal Medicine 2006;144(10):742-52.
Clamp S, Keen J. The Value of Electronic Health Records: A Literature Review,
2005.
Durieux P, Trinquart L, Colombet I, Nies J, Walton R, Rajeswaran A, et al.
Computerized advice on drug dosage to improve prescribing practice.
Cochrane Database Syst Rev 2008(3):CD002894.
Eslami S, Abu-Hanna A, de Jonge E, de Keizer NF. Tight glycemic control and
computerized decision-support systems: a systematic review. Intensive Care
Medicine 2009;35(9):1505-17.
Eslami S, Abu-Hanna A, De Keizer NF. Evaluation of outpatient computerized
physician medication order entry systems: A systematic review. Journal of the
American Medical Informatics Association 2007;14(4):400-06.
Eslami S, de Keizer NF, Abu-Hanna A. The impact of computerized physician
medication order entry in hospitalized patients - A systematic review.
Fitzmaurice DA, Hobbs FDR, Delaney BC, Wilson S, McManus R. Review of
computerized decision support systems for oral anticoagulation management.
British Journal of Haematology 1998;102(4):907-09.
Garg AX, Adhikari NKJ, McDonald H, Rosas-Arellano MP, Devereaux PJ,
Beyene J, et al. Effects of computerized clinical decision support systems on
practitioner performance and patient outcomes - A systematic review. JamaJournal of the American Medical Association 2005;293(10):1223-38.
Georgiou A, Williamson M, Westbrook JI, Ray S. The impact of computerised
physician order entry systems on pathology services: A systematic review.
Hayward GL, Parnes AJ, Simon SR. Using health information technology to
improve drug monitoring: a systematic review. Pharmacoepidemiology and
Drug Safety 2009;18(12):1232-37.
Hider P. Electronic Prescribing: A Critical Appraisal of the Literature, 2002.
Jamal A, McKenzie K, Clark M. The impact of health information technology on
the quality of medical and health care: a systematic review. Health Information
Management Journal 2009;38(3):26-37.
Khajouei R, Jaspers MWM. The impact of CPOE medication systems' design
aspects on usability, workflow and medication orders: a systematic review.
Methods of Information in Medicine 2010;49(1):3-19.
Niazkhani Z, Pirnejad H, Berg M, Aarts J. The impact of computerized provider
order entry systems on inpatient clinical workflow: a literature review. Journal of
the American Medical Informmatics Association 2009;16(4):539-49.
Poissant L, Pereira J, Tamblyn R, Kawasumi Y. The impact of electronic health
records on time efficiency of physicians and nurses: A systematic review.
Schedlbauer A, Prasad V, Mulvaney C, Phansalkar S, Stanton W, Bates DW, et
al. What evidence supports the use of computerized alerts and prompts to
improve clinicians' prescribing behavior? Journal of the American Medical
Shamliyan TA, Duval S, Du J, Kane RL. Just what the doctor ordered. Review
of the evidence of the impact of computerized physician order entry system on
medication errors. Health Services Research 2008;43(1):32-53.
Shekelle PG, Morton SC, Keeler EB. Costs and benefits of health information
technology. Evidence Report/ Technology Assessment (Full Report)
2006(132):1-71.
304
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
Shiffman RN, Liaw Y, Brandt CA, Corb GJ. Computer-based guideline
implementation systems: A systematic review of functionality and effectiveness.
Sintchenko V, Magrabi F, Tipper S. Are we measuring the right end-points?
Variables that affect the impact of computerised decision support on patient
outcomes: A systematic review. Medical Informatics and the Internet in
Medicine 2007;32:225-40.
Tan K, Dear PRF, Newell SJ. Clinical decision support systems for neonatal
care. Cochrane Database of Systematic Reviews 2005(2):CD004211.
van Rosse F, Maat B, Rademaker CMA, van Vught AJ, Egberts ACG, Bollen
CW. The Effect of Computerized Physician Order Entry on Medication
Prescription Errors and Clinical Outcome in Pediatric and Intensive Care: A
Systematic Review. Pediatrics 2009;123(4):1184-90.
Wolfstadt JI, Gurwitz JH, Field TS, Lee M, Kalkar S, Wu W, et al. The effect of
computerized physician order entry with clinical decision support on the rates of
adverse drug events: A systematic review. Journal of General Internal Medicine
2008;23(4):451-58.
Yourman L, Concato J, Agostini JV. Use of computer decision support
interventions to improve medication prescribing in older adults: A systematic
review. American Journal of Geriatric Pharmacotherapy 2008;6(2):119-29.
Classen DC, Metzger J. Improving medication safety: the measurement
conundrum and where to start. International Journal for Quality in Health Care
2003;15:I41-I47.
Nebeker JR, Hurdle JF, Hoffman JM, Roth B, Weir CR, Samore MH.
Developing a taxonomy for research in adverse drug events: Potholes and
signposts. Journal of the American Medical Informatics Association
2002;9(6):S80-S85.
Bates DW, Leape LL, Cullen DJ, Laird N, Petersen LA, Teich JM, et al. Effect
of computerized physician order entry and a team intervention on prevention of
serious medication errors. Jama-Journal of the American Medical Association
1998;280(15):1311-16.
Potts AL, Barr FE, Gregory DF, Wright L, Patel NR. Computerized physician
order entry and medication errors in a pediatric critical care unit. Pediatrics
2004;113(1):59-63.
Bates DW, Teich JM, Lee J, Seger D, Kuperman GJ, Ma'Luf N, et al. The
impact of computerized physician order entry on medication error prevention.
Evans RS, Pestotnik SL, Classen DC, Clemmer TP, Weaver LK, Orme JF, et
al. A computer-assisted management program for antibiotics and other
antiinfective agents. New England Journal of Medicine 1998;338(4):232-38.
Mullett CJ, Evans RS, Christenson JC, Dean JM. Development and impact of
a computerized pediatric antiinfective decision support program. Pediatrics
2001;108(4):art. no.-e75.
Shulman R, Singer M, Goldstone J, Bellingan G. Medication errors: a
prospective cohort study of hand-written and computerised physician order
entry in the intensive care unit. Critical Care 2005;9(5):R516-R21.
Nies J, Colombet I, Degoulet P, Durieux P. Determinants of success for
computerized clinical decision support systems integrated in CPOE systems: a
systematic review. AMIA Annual Symposium Proceedings 2006:594-8.
Balaguer Santamaria JA, Fernandez Ballart JD, Escribano Subias J.
Usefulness of a software package to reduce medication errors in neonatal care.
Anales Españoles de Pediatría 2001;55(6):541-5.
The Audit Commission. A Spoonful of Sugar: Medicines Management in NHS
Hospitals, 2001.
305
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
Abu-Zayed L, Farrar K, Mottram DR. Time spent on drug supply activities in
United Kingdom hospitals. American Journal of Health-System Pharmacy
2000;57(21):2006-07.
Murray MD, Loos B, Tu WZ, Eckert GJ, Zhou XH, Tierney WM. Effects of
computer-based prescribing on pharmacist work patterns. Journal of the
Ross SM, Papshev D, Murphy EL, Sternberg DJ, Taylor J, Barg R. Effects
of electronic prescribing on formulary compliance and generic drug utilization in
the ambulatory care setting: a retrospective analysis of administrative claims
data. Journal of Managed Care Pharmacy 2005;11(5):410-5.
Beer J, Dobish R, Chambers C. Physician order entry: a mixed blessing
to pharmacy? Journal of Oncology Pharmacy Practice 2002;8:119-26.
Wogen SE, Fulop, G., Heller, J. Electronic prescribing: improving the efficiency
of the prescription process and promoting plan adherence. Drug Benefit Trends
2003;15:35-40.
Mekhjian HS, Kumar RR, Kuehn L, Bentley TD, Teater P. Immediate benefits
realized following implementation of physician order entry at an academic
medical center. Journal of the American Medical Informatics Association
2002;9(5):529-39.
Lehman ML, Brill JH, Skarulis PC, Keller D, Lee C. Physician order entry
impact on drug turn-around times. Journal of the American Medical Informatics
Association 2001:359-63.
Bernstein SL, Whitaker D, Winograd J, Brennan JA. An electronic chart prompt
to decrease proprietary antibiotic prescription to self-pay patients. Academic
Emergency Medicine 2005;12(3):225-31.
Siegel C, Alexander MJ, Dlugacz YD, Fischer S. Evaluation of a computerized
drug review system - impact, attitudes, and interactions. Computers and
Biomedical Research 1984;17(5):419-35.
Walton RT, Gierl C, Yudkin P, Mistry H, Vessey MP, Fox J. Evaluation of
computer support for prescribing (CAPSULE) using simulated cases. British
Medical Journal 1997;315(7111):791-95.
Christakis DA, Zimmerman FJ, Wright JA, Garrison MM, Rivara FP, Davis
RL. A randomized controlled trial of point-of-care evidence to improve the
antibiotic prescribing practices for otitis media in children. Pediatrics
2001;107(2):art. no.-e15.
Rivkin S. Opportunities and challenges of electronic physician prescribing
technology. Medical Interface 1997;10(8):77-8, 83.
Chin HL, Wallace P. Embedding guidelines into direct physician order
entry: Simple methods, powerful results. Journal of the American Medical
Informatics Association 1999:221-25.
Overhage JM, Tierney WM, Zhou XH, McDonald CJ. A randomized trial of
''corollary orders'' to prevent errors of omission. Journal of the American
Medical Informatics Association 1997;4(5):364-75.
Smith BJ, McNeely MDD. The influence of an expert system for test ordering
and interpretation on laboratory investigations. Clinical Chemistry
1999;45(8):1168-75.
Kuperman GJ, Teich JM, Tanasijevic MJ, Ma'Luf N, Rittenberg E, Jha A, et
al. Improving response to critical laboratory results with automation: Results of
a randomized controlled trial. Journal of the American Medical Informatics
Association 1999;6(6):512-22.
Han YY, Carcillo JA, Venkataraman ST, Clark RSB, Watson RS, Nguyen
TC, et al. Unexpected increased mortality after implementation of a
commercially sold computerized physician order entry system. Pediatrics
2005;116(6):1506-12.
306
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
Del Beccaro MA, Jeffries HE, Eisenberg MA. Computerized provider order
entry implementation: No association with increased mortality rates in an
intensive care unit. Pediatrics 2006;118(1):290-95.
Keene A, Ashton L, Shure D, Napoleone D, Katyal C, Bellin E. Mortality before
and after initiation of a computerized physician order entry system in a critically
ill pediatric population. Pediatric Critical Care Medicine 2007;8(3):268-71.
Ammenwerth E, Talmon J, Ash JS, Bates DW, Beuscart-Zephir MC, Duhamel
A, et al. Impact of CPOE on mortality rates - Contradictory findings, important
messages. Methods of Information in Medicine 2006;45(6):586-93.
Rosenbloom ST, Harrell FE, Lehmann CU, Schneider JH, Spooner SA,
Johnson KB. Perceived increase in mortality after process and policy changes
implemented with computerized physician order entry. Pediatrics
2006;117(4):1452-55.
Sittig DF, Ash JS, Zhang JJ, Osheroff JA, Shabot MM. Lessons from
"Unexpected increased mortality after implementation of a commercially sold
computerized physician order entry system". Pediatrics 2006;118(2):797-801.
Bradley VM, Steltenkamp CL, Hite KB. Evaluation of reported medication errors
before and after implementation of computerized practitioner order entry.
Journal of Healthcare Information Management 2006;20(4):46-53.
Koppel R, Metlay JP, Cohen A, Abaluck B, Localio AR, Kimmel SE, et al. Role
Jama-Journal of the American Medical Association 2005;293(10):1197-203.
Bates DW. Computerized physician order entry and medication errors:
Finding a balance. Journal of Biomedical Informatics 2005;38(4):259-61.
Morris CJ, Savelyich BSP, Avery AJ, Cantrill JA, Sheikh A. Patient safety
features of clinical computer systems: questionnaire survey of GP views.
Quality & Safety in Health Care 2005;14(3):164-68.
Rotman BL, Sullivan AN, McDonald TW, Brown BW, DeSmedt P, Goodnature
D, et al. A randomized controlled trial of a computer-based physician
workstation in an outpatient setting: Implementation barriers to outcome
evaluation. Journal of the American Medical Informatics Association
1996;3(5):340-48.
Schectman JM, Schorling JB, Nadkarni MM, Voss JD. Determinants of
physician use of an ambulatory prescription expert system. International
Journal of Medical Informatics 2005;74(9):711-17.
Tamblyn R, Huang A, Kawasumi Y, Bartlett G, Grad R, Jacques A, et al. The
development and evaluation of an integrated electronic prescribing and drug
management system for primary care. Journal of the American Medical
Spencer DC, Leininger A, Daniels R, Granko RP, Coeytaux RR. Effect of a
computerized prescriber-order-entry system on reported medication errors.
American Journal of Health-System Pharmacy 2005;62(4):416-19.
Overhage JM, Perkins S, Tierney WM, McDonald CJ. Research paper Controlled trial of direct physician order entry: Effects on physicians time
utilization in ambulatory primary care internal medicine practices. Journal of the
Tierney WM, Miller ME, Overhage JM, McDonald CJ. Physician inpatient order
writing on microcomputer workstations - effects on resource utilization. JamaJournal of the American Medical Association 1993;269(3):379-83.
Bates DW, Boyle DL, Teich JM. Impact of computerized physician order entry
on physician time. Proceedings of Annual Symposium on Computer Application
in Medical Care 1994:996.
Shu K, Boyle D, Spurr C, Horsky J, Heiman H, O'Connor P, et al. Comparison
of time spent writing orders on paper with computerized physician order entry.
307
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
Medinfo 2001: Proceedings of the 10th World Congress on Medical Informatics,
Pts 1 and 2 2001;84:1207-11.
van der Meijden MJ, Tange H, Troost J, Hasman A. Development and
implementation of an EPR: how to encourage the user. International Journal of
Medical Informatics 2001;64(2-3):173-85.
Cordero L, Kuehn L, Kumar RR, Mekhjian HS. Impact of computerized
physician order entry on clinical practice in a newborn intensive care unit.
Journal of Perinatology 2004;24(2):88-93.
Fernando B, Savelyich BSP, Avery AJ, Sheikh A, Bainbridge M, Horsfield P,
et al. Prescribing safety features of general practice computer systems:
evaluation using simulated test cases. British Medical Journal
2004;328(7449):1171-72.
Teich JM, Merchia PR, Schmiz JL, Kuperman GJ, Spurr CD, Bates DW. Effects
of computerized physician order entry on prescribing practices. Archives of
Internal Medicine 2000;160(18):2741-47.
Evans RS, Pestotnik SL, Classen DC, Bass SB, Burke JP. Prevention of
adverse drug events through computerized surveillance. Proceedings of Annual
Symposium on Computer Application in Medical Care 1992:437-41.
Dortch MJ, Mowery NT, Ozdas A, Dossett L, Cao H, Collier B, et al. A
computerized insulin infusion titration protocol improves glucose control with
less hypoglycemia compared to a manual titration protocol in a trauma intensive
care unit. Journal of Parenteral and Enteral Nutrition 2008;32(1):18-27.
Boord JB, Sharifi M, Greevy RA, Griffin MR, Lee VK, Webb TA, et al.
Computer-based insulin infusion protocol improves glycemia control over
manual protocol. Journal of the American Medical Informatics Association
2007;14(3):278-87.
Fortescue EB, Kaushal R, Landrigan CP, McKenna KJ, Clapp MD, Federico F,
et al. Prioritizing strategies for preventing medication errors and adverse drug
events in pediatric inpatients. Pediatrics 2003;111(4):722-29.
Weir CR, Staggers N, Phansalkar S. The state of the evidence for
computerized provider order entry: A systematic review and analysis of the
quality of the literature. International Journal of Medical Informatics
2009;78(6):365-74.
Cresswell K, Bates, D.W., Phansalkar, S. and Sheikh, A. KOpportunities and
challenges in creating an international centralised knowledge base for clinical
decision support systems (CDSS) in prescribing. Quality and Safety in Health
Care Forthcoming.
British TS. British Guideline on the Management of Asthma: Quick Reference
Guide, June 2009.
Aronson JK. A prescription for better prescribing. British Journal of Clinical
Pharmacology 2006;61(5):487-91.
Strom BL, Schinnar R, Aberra F, Bilker W, Hennessy S, Leonard CE, et al.
Unintended Effects of a Computerized Physician Order Entry Nearly Hard-Stop
Alert to Prevent a Drug Interaction A Randomized Controlled Trial. Archives of
Internal Medicine 2010;170(17):1578-83.
Van der Sijs H, Aarts J, Vulto A, Berg M. Overriding of drug safety alerts in
computerized physician order entry. Journal of the American Medical
Procter R, Pappas, Y., Car J., Sheikhh A. and Majeed A. From human factors
to human actors: taking user involvement in eHealth systems seriously.
Forthcoming.
Cafazzo JA, Trbovich PL, Cassano-Piche A, Chagpar A, Rossos PG, Vicente
KJ, et al. Human factors perspectives on a systemic approach to ensuring a
safer medication delivery process. Healthcare Quarterly 2009;12 Spec No
Patient:70-4.
308
114. Vicente KJ, editor. The Human Factor. New York: Routledge, 2004.
115. Saathoff A. Human factors considerations relevant to CPOE implementations.
Journal of Healthcare Information Management 2005;19(3):71-8.
116. Pearson M. Synthesizing qualitative and quantitative health evidence: A guide
to methods. Sociology of Health & Illness 2008;30(2):330-31.
117. Pope C, Mays, Nicholas, Popay, Jennie, editor. Synthesizing Qualitative and
Quantitative Health Evidence: A Guide to Methods. New York: Open University
Press, 2007.
118. Greenhalgh T, Potts HWW, Wong G, Bark P, Swinglehurst D. Tensions and
Paradoxes in Electronic Patient Record Research: A Systematic Literature
Review Using the Meta-narrative Method. Milbank Quarterly 2009;87(4):729-88.
119. Lilford RJ, Foster J, Pringle M. Evaluating eHealth: how to make evaluation
more methodologically robust. PLoS Medicine 2009;6(11):e1000186.
120. Williams R, Edge D. The social shaping of technology. Research Policy
1996;25(6):865-99.
121. Clamp, S, Keen J. Electronic health records: Is the evidence base any use?
Medical Informatics and the Internet in Medicine 2007;32(1):5-10.
122. Durieux P, Trinquart L, Colombet I, Nies J, Walton R, Rajeswaran A, et al.
Computerized advice on drug dosage to improve prescribing practice.
Cochrane Database of Systematic Reviews 2010(10):CD002894.
309
Chapter 11
Case study: ePrescribing for oral anticoagulation
therapy in primary care with warfarin
Summary
•
Anticoagulants are highly efficacious treatments for a range of conditions;
they are, however, also associated with considerable risk of iatrogenic
harm if monitoring of treatment and dosage adjustments are poorly
managed. Monitoring has thus historically taken place in hospital clinics.
•
Widening indications for the use of oral anticoagulation therapy have,
however, in recent years led to overstretch of many hospital-based
anticoagulation services. Coupled with the political imperative to provide
more accessible, patient-centred, models of care, this has catalysed the
recent move of anticoagulation services to primary care.
•
These drivers need to be balanced against concerns about the safety of
prescribing and monitoring of warfarin therapy in primary care.
•
ePrescribing for oral anticoagulation therapy, specifically the use of
computerised decision support systems, has the potential to provide
prescribers with real-time advice on prescribing and monitoring decisions.
•
There is strong and consistent evidence that these theoretical benefits in
relation to computerised clinical decision support for oral anticoagulation
therapy with warfarin can be realised in primary care resulting in greater
therapeutic control than might otherwise be possible.
•
The automation of dosage calculations and determination of time until the
next appointment is acceptable to primary care prescribers.
•
There are strong arguments for the rolling out and integration of this
decision support tool into future upgrades of ePrescribing systems in
English primary care.
•
Understanding the reasons underpinning the success of computerised
decision support for oral anticoagulation therapy—which to a large extent
relate to the fact that this meets a genuine clinical need rather than a
technologically driven “solution”—should provide insights into the contexts
in which it is best to prioritise development of other decision support tools
310
11.1 Introduction
The National Institute for Health and Clinical Excellence (NICE) has estimated
that approximately 1.4% of the population now needs oral anticoagulation
therapy (OAT).1 This high population estimate reflects the fact that indications
for OAT have broadened in recent years, in particular following publication of
the evidence for the effectiveness of warfarin in stroke prevention in patients
with atrial fibrillation.2
3
This substantial increase in the absolute number of
patients treated with OAT led to overburdened anticoagulation services, the
policy response to which has been the development of new models of care,
these centring on an outsourcing of care from specialist to generalist settings.
4 5
When the transition was first being made in the early 1990s, general
practitioners (GPs) were on the whole resistant to the idea of running their
own anticoagulation clinics, this hesitation stemming from: perceived
insufficient time, knowledge and training; lack of appropriate facilities and
equipment; and a lack of financial incentives.6
Regardless of GPs initial apprehension, where OAT was once delivered
exclusively in specialised secondary care clinics, it is now regularly delivered
in primary care in England.2;7 Considering the high numbers of people now
taking OAT, this is in many senses an appropriate
development.8 This
reorganisation of the provision of anticoagulation services, however, remains
controversial, with criticism centring on the quality of OAT when administered
in primary care compared with its delivery in specialist care settings.9
The benefit to risk ratio of anticoagulants depends heavily on keeping
anticoagulation control within narrow limits (i.e. within the therapeutic range,
which is typically measured using the International Normalised Ratio or INR)
in order to minimise the risk of adverse events due either to under- or overtreatment.10 Successful control depends on skilled dosing, effective laboratory
quality control, and regular monitoring. Of particular note in the context of this
chapter is that in its influential report, Improving Medication Safety, the
Department of Health (DH) has highlighted concerns with the prescribing of
anticoagulants in primary care, noting that it is now one of the classes of
medication most commonly associated with fatal medication errors.
11
Also
311
noteworthy are the findings from a systematic review (SR) and meta-analysis
of 45 studies conducted in both primary and secondary care anticoagulation
clinics, which found that improved therapeutic control could reduce the risk of
anticoagulation-related adverse events by almost half.12
In response to a risk assessment of anticoagulation therapy conducted by the
National Patient Safety Agency (NPSA), a report was issued recommending
that all healthcare organisations take a range of actions (Box 11.1) under the
auspices of Patient Safety Alert 18 to make anticoagulant therapy safer.13
Box 11.1: Actions that can make anticoagulant therapy safer
•
Ensure all staff caring for patients on anticoagulant therapy have the necessary
work competences.
Any gaps in competence must be addressed through
training to ensure that all staff undertake their duties safely.
•
Review and, where necessary, update written procedures and clinical protocols
for anticoagulant services to ensure they reflect safe practice and that staff are
trained in these procedures.
•
Audit anticoagulant services using British Society for Haematology and NPSA
safety indicators as part of the annual medicines management audit programme.
The audit results should inform local actions to improve the safe use of
anticoagulants, and should be communicated to clinical governance, and drugs
and therapeutics committees (or equivalent).
Commissioners and external
organisations should use this information as part of the commissioning and
performance management process.
•
Ensure that patients prescribed anticoagulants receive appropriate verbal and
written information at the start of therapy, at hospital discharge, on the first
anticoagulation clinic appointment and when necessary throughout the course of
their treatment.
•
Promote safe practice amongst prescribers and pharmacists to check that
patients’ INR is being monitored regularly and that the INR level is safe before
issuing or dispensing repeat prescriptions for oral anticoagulants.
•
Promote safe practice for prescribers co-prescribing one or more clinically
significant interacting medicines for patients already on oral anticoagulants to
make arrangements for additional INR blood tests and to inform the anticoagulant
service that an interacting medicine has been prescribed.
Ensure that those
312
dispensing clinically significant interacting medicines for these patients check that
these additional safety precautions have been taken.
•
Ensure that dental practitioners manage patients on anticoagulants according to
evidence-based therapeutic guidelines. In most cases, dental treatment should
proceed as normal and oral anticoagulant treatment should not be stopped or the
dosage decreased inappropriately.
•
Amend local policies to standardise the range of anticoagulant products used,
incorporating characteristics identified by patients as promoting safer use.
•
Promote the use of written safe practice procedures for the use of anticoagulants
in care homes. It is safe practice for all dose changes to be confirmed in writing
by the prescriber.
A risk assessment should be undertaken on the use of
Monitored Dosage Systems for anticoagulants for individual patients.
The
general use of Monitored Dosage Systems should be minimised as dosage
changes using these systems are more difficult.
Ensure all staff caring for
patients on anticoagulant therapy have the necessary work competences. Any
gaps in competence must be addressed through training to ensure that all staff
undertake their duties safely.
Source: NPSA Patient Safety Alert 18 (2007)
13
Reprinted with permission from the National Patient
Safety Agency
The NPSA and the British Committee for Standards in Haematology (BCSH)
have identified safety indicators (see Box 11.2) for outpatient OAT.14
Monitoring these indicators should help to identify risks and promote
appropriate action to minimise risk. The safety indicators can also be used to
audit the implementation of the recommendations made in the NPSA’s Patient
Safety Alert 18.13
313
Box 11.2: Patient safety indicators for outpatient oral anticoagulation therapy
•
Proportion of patient time in range (if this is not measurable because of
inadequate decision support software then a secondary measure of percentage
of INRs in range should be used).
•
Percentage of INRs > 5·0.
•
Percentage of INRs > 8·0.
•
Percentage of INRs > 1·0 INR unit below target (e.g. percentage of INRs < 1·5
for patients with target INR of 2·5).
•
Percentage of patients suffering adverse outcomes, categorised by type, e.g.
major bleed.
•
Percentage of patients lost to follow up (and risk assessment of process for
identifying patients lost to follow up).
•
Percentage of patients with unknown diagnosis, target INR or stop date.
•
Percentage of patients with inappropriate target INR for diagnosis, high and low.
•
Percentage of patients without written patient educational information.
•
Percentage of patients without appropriate written clinical information, e.g.
diagnosis, target INR, last dosing record.
14
Adapted from BCSH (2007). Reprinted with permission from the National Patient Safety Agency.
Employing an ePrescribing application (see Chapter 10)–with computerised
decision support system (CDSS) that provides support for determining dosage
and time until next appointment–in primary care has been shown to be
effective in improving therapeutic control.7
support
systems
could
furthermore
15 16 17 18
facilitate
The use of decision
adherence
to
NPSA
recommendations and aid in the monitoring of BCSH safety indicators to
support the provision of anticoagulation services in primary care.
Decision support systems for OAT can utilise input which takes the form of a
target INR, the therapeutic range for a specific condition, history of bleeding
problems, number of prior visits to the clinic, variability of prothrombin over
time, cost of complications and the estimated costs of a visit.19
20
Systems
employ a number of inference mechanisms (see Chapter 10), which draw on
314
pharmacological models of anticoagulation control, to arrive at the
recommended output.
The use of ePrescribing for OAT has the potential to benefit patients,
providers and health services.
For example, patients may benefit from
improved therapeutic control, spending more time within the therapeutic
range, thereby reducing the number of INR tests and visits required to
maintain therapeutic control in comparison to therapy delivered without a
CDSS.
Depending on whether near patient testing (NPT; also sometimes known as
“point-of-care” or POC testing) with coagulometers is employed in primary
care, the need for laboratory monitoring could be almost entirely removed with
the potential benefit of increased patient convenience. This is because use of
NPT also virtually eliminates turnaround time for results compared with
centralised lab-based testing, making the process of anticoagulation
management both faster and more straightforward.
Providing anticoagulation services in primary care should furthermore enable
clinicians to provide more comprehensive and continuous healthcare; and this
more complete clinical knowledge about patients should enable clinicians to
make more informed clinical decisions both in relation to anticoagulation
decisions and care more generally.
As the use of CDSSs allows for the management of patients by non-clinicians,
some workload may be shifted away from clinical staff, thus creating new
roles and responsibilities for other (less expensive) healthcare providers. Use
of these technologies also offer the inherent advantage of opening up the
database of patients to a variety of secondary uses, these including audit,
significant event analysis and research.21
The quality and/or efficiency of healthcare often initially suffers as new
systems are introduced and take time to embed (see Chapter 18); there are
315
also resource and opportunity costs associated with system implementation.
Cognisance of the disruptive effect on clinical workflow associated with
implementation is therefore important. Training of staff is important–as is, for
example, highlighted by The National Centre for Anticoagulation Training,
which cautions that the use of a CDSS should be restricted to those staff that
have completed appropriate training and recommend a training record log is
kept for all staff members working in such clinics–but this is also likely to
disrupt clinical workflows by taking busy professionals out of other core
activities.7
Although there have been no reports of CDSSs providing care inferior to that
delivered without CDSSs, this might occur if the support provided by the
CDSS is predicated on an invalid knowledge base or algorithm. Even though
it is in the manufacturers’ best interests to produce systems that have been
rigorously evaluated, there is as noted in earlier chapters no third party
assurance of validity and safety. As such, patients might actually suffer from
a worsening in therapeutic control, with the associated increased risks of
adverse events.
The provision of less effective OAT might not be detected initially and if
detected, could result in abandonment of the application and subsequent
financial loss for primary care.
Careful monitoring and evaluation of the
impact of the application on the quality and safety of healthcare is thus
imperative in order to reduce risks and attend to problems expeditiously.
11.3 Empirically demonstrated impact
11.3.1 Benefits
The SR by Garg et al. included a number of randomised (and other) trials of
warfarin dosing systems.22 Seven of twelve trials for warfarin dosing improved
practitioner performance as measured in the main by time to achieve a
therapeutic INR, the proportion of time within the therapeutic range, the
proportion of patients with therapeutic INR, the number of days between INR
testing and the number of test measurements. This review, however, included
both initiation and maintenance phases of OAT conducted in inpatient and
316
outpatient settings.
This represents a major challenge to interpreting the
findings from this review as the underlying algorithms differ between the
initiation and maintenance phases.23
An earlier SR and meta-analysis of nine randomised controlled trials (RCTs)
with a total of 1336 patients—the majority of whom were also included in the
review by Garg et al.22—focusing largely on systems for OAT with warfarin,
found that the use of computer programmes for anticoagulation optimisation
increased the proportion of visits where patients were within the therapeutic
range by 29 per cent (pooled odds ratio 1.29; 95%CI 1.12, 1.49).24 There
was, however, significant heterogeneity (P=0.02) between trials making this
summary analysis open to question; to their credit, the authors subsequently
conducted two additional analyses to investigate the source of the
heterogeneity after excluding:
•
The only study for which the medication in question was heparin, the
results remained unchanged from that of the overall analysis
•
One of the smallest studies with the largest effect the pooled odds ratio
decreased slightly to 1.25 (95%CI 1.08, 1.45), but this however
reduced the heterogeneity such that it was no longer significant
(P=0.12); this additional analysis thus uncovered the source of the
heterogeneity and, importantly, demonstrated that even after removing
this trial that CDSS for OAT resulted in significant improvements in INR
control.24
The unit of assessment in this meta-analysis was not the patient, but the
anticoagulation test and the end point of the proportion of tests within the
target range; using this outcome, the sample size was 3416 dosages (carried
out in 1327 patients).24 Further, it should be noted that this SR and metaanalysis also included studies of both initiation and maintenance phases and
both inpatients and outpatients; it therefore poses many of the same
challenges to interpretation as Garg et al’s review.22
317
Another SR on the use of CDSSs for OAT, focusing on its use in primary care
with NPT, was published in 1998.24 Of the seven included studies (with little
overlap of studies included by Garg et al.22 and Chatellier et al.24), only one
was judged to be of high quality. The authors concluded from the one high
quality study that there was evidence that a CDSS could achieve improved
therapeutic control in terms of INR, when compared with human performance
alone.24
The important series of studies by Fitzmaurice and colleagues from the
University of Birmingham, England are worth reflecting on in greater detail.
The first study, published in 1996, assessed 49 patients treated with warfarin
for 12 months using a RCT design.15 There were significant improvements
noted in the proportion of patients with INR control within therapeutic limits,
from 23 per cent to 86 per cent (P<0.001) in the practice where all patients
received dosages with the aid of a CDSS. In the practice where patients were
randomised to either CDSS or hospital dosage, analysis showed a significant
improvement in the CDSS group, which was not apparent in the patients who
received dosage in hospital (P<0.001). Mean recall times were significantly
extended in patients who received dosage instructions with the aid of the
CDSS, from 24 to 36 days (P=0.03). Patient satisfaction with the practice
clinics was also high. This study, however, did not use NPT and specimens
were sent to laboratory with results returned usually on the same day.15
An extension of this study–with the addition of NPT–was published two years
later.16 This was conducted outside trial conditions with data collected over
the course of 12 months from a dedicated nurse-led OAT clinic within primary
care. The cumulative results of this longitudinal study were that the overall
mean percentage of patients within the therapeutic range was 71 per cent and
overall the proportion of INRs within the therapeutic range was 53 per cent.
Importantly, no adverse events were reported and no patients had to be
referred back to secondary care.16
The same group published a second RCT in 2000.17 Of the 248 practices
within Birmingham, England, 12 were randomly selected from a list of 21
318
practices that had expressed interest in the study. The design was somewhat
unusual in that there were two sets of controls: those from the intervention
practices (described as the intra-practice controls) and all patients in the
control practices (described as the inter-practice controls). The intervention
was a nurse-led practice-based anticoagulation clinic using CDSS and NPT,
this being compared to the standard hospital-based assessment. The primary
outcome for this trial was therapeutic control of the INR and this showed
significant improvements for the intervention patients (P<0.01).17
An extension of this study published in 2001 reported the degree of OAT
control for patients from the same practices for an 18-month period after the
completion of the original study.18 the authors found that there were no
significant differences between the practice- and hospital-based populations
in terms of the percentage of time in range (69 per cent practice-based, 64
per cent hospital-based).
The proportion of tests in range was, however,
significantly higher in the practice-based group at 61 per cent vs 57 per cent
for hospital-based (P=0.02). Mean recall time was virtually identical in both
groups at 36 days. There were no significant differences between groups for
the number of clinical outcomes per patient. Overall, the authors found
comparable quality of control when compared to the original study, this
demonstrating that primary care-based OAT supported by CDSS and NPT
was of at least equivalent quality to the hospital-based care.18
In their 1996 study, using average hospital review rates, Fitzmaurice et al.
estimated 148 additional appointments would have been offered to 26
patients in an intervention practice during the 12 months if CDSS had not
been used.15 The authors assessed the total cost savings to be non-existent
in the first year of use (-£476), but estimated a savings of £2,604 for each
subsequent year for the intervention practice with 26 patients. The authors
also concluded that the greater the number of patients seen in a high cost
provider environment, the more economically efficient a CDSS becomes.15
Extending their 1996 study15 outside trial conditions with the addition of NPT,
Fitzmaurice et al. estimated a total cost savings of £539, from £2290 to £1751
319
from the use of their dedicated, nurse-led primary care anticoagulation clinic
using a CDSS and NPT for the 12-month study period.16
In an analysis of patients’ costs in primary versus secondary care, a patient’s
cost per visit was found to be significantly higher when being reviewed in
secondary care: average patient cost per visit in primary care of £6.78 vs
£14.58 in secondary care. These cost differences were to a large extent as a
result of longer journey time and the use of more expensive means of
transport when attending for hospital-based assessments, and shorter clinic
visits in primary care.25
11.3.2 Risks
There was no evidence of an increased risk of adverse events reported in
these studies of primary care delivered CDSS supported OAT. Furthermore,
there were no significant new risks introduced indicating that the underlying
algorithms are on the whole probably well-constructed. The overall costs to
healthcare systems are however likely to be increased with primary care-led
models of care.17 25 26
11.4 Conclusions: Implications for practice, policy and research
The principal outcome measures for any anticoagulation service are the
prevention
of
thrombotic
and
avoidance
of haemorrhagic
events.23
Unfortunately, none of the SRs investigating the role of CDSSs found clear
evidence for improved outcomes for these clinically important endpoints.
There are, more generally, limited data regarding the absolute and relative
risks of OAT as most studies lack sufficient power to detect significant
changes in these relatively uncommon outcomes.17 22 24 This issue is further
complicated by the lack of agreement on definitions and criteria for major and
minor adverse events. It is perhaps for these reasons that evaluations have
tended to focus on time spent on proxy measures such as the proportion of
patients in and time spent within the therapeutic range.
As most of the research pertains to warfarin, further research into the
prescribing
of
other
oral
anticoagulants
such
as
phenindione
and
320
acenocoumarol should also be considered, as the findings might not be
directly transferable to these other oral anticoagulants.26
There are also
outstanding questions about the positioning of newer agents–which do not
require monitoring–such as dabigatran and rivaroxaban, which may in due
course replace use of warfarin.27 28
Finally, both the systematic reviews by Garg et al.22 and Chatellier et al.24 did
not mention whether NPT was part of the intervention in the included primary
studies. This is important as the degree to which an additional technology
such as NPT contributes to the overall effect of CDSSs for anticoagulation is
not yet well understood.
However, the automation of dosage calculation and determination of time until
next appointment has proved to be readily adoptable.
As outlined in the
chapters on CDSSs, ePrescribing human factors and organisational issues in
design, development and deployment (see Chapters 8, 10, 16 and 17) the
success of an application is dependent on a number of variables. It seems
likely that in this case of CDSS-supported OAT this is most likely due to the
relative advantage of using the technology being driven by a recognised need
to improve the quality and safety of OAT and reduce the inconvenience to
patients (rather than primarily a desire to improve technology and application
validity).
That said, there is still the need for further research, specifically to compare in
a head-to-head fashion, the different algorithms that are currently being
employed in CDSS applications for OAT using warfarin, to establish which, if
any, is best at supporting practitioner performance.
A final point worth noting about CDSSs for OAT is that the applications are for
the most part stand-alone systems and this poses potential problems with
regards to integration with existing records (see Chapter 5 for a further
discussion on the matter). Concerns over interoperability are likely to serve
as a deterrent to implementation and use of the application in primary care
and may also potentially compromise safety as a consequence of fragmenting
321
the record of care. This concern should be mitigated by ensuring compatibility
with primary care computing systems or incorporating the functionality within
ePrescribing systems.
Increased provision of OAT opens the potential door to a future move to
patient self-testing and -monitoring based models, however, available data
suggest that whilst effective, these models are not particularly cost-effective.29
In summary, given that the technical infrastructure exists within primary care
in which to readily incorporate this particular ePrescribing application, the fact
that there are also now clear financial incentives to provide OAT services in
primary care and the clearly demonstrated beneficial impact on the quality of
anticoagulation care, it is important that this eHealth application is now used
much more widely in English primary care.30
References
1. National Institute for Health and Clinical Excellence (NICE). Assumptions used in
estimating
a
population
benchmark.
Available
from:
http://www.nice.org.uk/usingguidance/commissioningguides/anticoagulationthera
pyservice/popbench.jsp (last accessed 30/12/10)
2. Ezekowitz MD, Bridgers SL, James KE, Carliner NH, Colling CL, Gornick CC, for
the Veterans Affairs Stroke Prevention in Nonrheumatic Atrial Fibrillation
Investigators. Warfarin in the prevention of stroke associated with nonrheumatic
atrial fibrillation. N Engl J Med. 1992;327:1406–12
3. Aguilar MI, Hart R. Oral anticoagulants for preventing stroke in patients with nonvalvular atrial fibrillation and no previous history of stroke or transient ischemic
attacks. Cochrane Database of Systematic Reviews 2005, 3. Art. No.:
CD001927. DOI: 10.1002/14651858.CD001927.pub2
4. Sudlow CM, Rodgers H, Kenny RA, Thomson RG. Service provision and use of
anticoagulants in atrial fibrillation. BMJ 1995; 311: 558-60.
5. Sweeney KG, Gray DP, Steele R, Evans P. Use of warfarin in non-rheumatic
atrial fibrillation: a commentary from general practice. Br J Gen Pract 1995; 45:
153-58.
6. Taylor F, Ramsay M, Voke J, Cohen H. Anticoagulation in patients with atrial
fibrillation. GPs not prepared for monitoring anticoagulation. BMJ 1993; 307:
1493.
7. National
Centre
for
Anticoagulation
Training.
Available
from:
http://www.anticoagulation.org.uk/courses.php (last accessed 30/12/10)
8. National Institute for Health and Clinical Excellence. Anticoagulative therapy
service.
Commissioning Guides - Supporting Clinical Service Redesign.
Available from: http://www.nice.org.uk/page.aspx?o=370987 (last accessed
30/12/10).
322
9. Wilson SJ, Wells PS, Kovacs MJ, Lewis GM, Martin J, Burton E et al. Comparing
the quality of oral anticoagulant management by anticoagulation clinics and by
family physicians: a randomized controlled trial. CMAJ 2003; 169: 293-98.
10. Oake N, Jennings A, Forster AJ, Fergusson D, Doucette S, van Walraven C.
Anticoagulation intensity and outcomes among patients prescribed oral
anticoagulant therapy: a systematic review and meta-analysis. CMAJ 2008; 179:
235-44.
11. Department of Health. Building a Safer NHS for Patients: Improving Medication
Safety. London: DH, 2004.
12. Oake N, Fergusson DA, Forster AJ, van Walraven C. Frequency of adverse
events in patients with poor anticoagulation: a meta-analysis. CMAJ 2007; 176:
1589-94.
13. National Patient Safety Agency. Patient Safety Alert 18 - Actions that can make
anticoagulant
therapy
safer.
Available
from:
http://www.nrls.npsa.nhs.uk/resources/?entryid45=59814
(last
accessed
30/12/10)
14. Baglin TP, Cousins D, Keeling DM, Perry DJ, Watson HG. Safety indicators for
inpatient and outpatient oral anticoagulant care: recommendations from the
British Committee for Standards in Haematology and National Patient Safety
Agency. Br J Haematol 2007; 136: 26-29.
15. Fitzmaurice DA, Hobbs FD, Murray ET, Bradley CP, Holder R. Evaluation of
computerized decision support for oral anticoagulation management based in
primary care. Br J Gen Pract 1996; 46: 533-35.
16. Fitzmaurice DA, Hobbs FD, Murray ET. Primary care anticoagulant clinic
management using computerized decision support and near patient International
Normalized Ratio (INR) testing: routine data from a practice nurse-led clinic. Fam
Pract 1998; 15: 144-46.
17. Fitzmaurice DA, Hobbs FD, Murray ET, Holder RL, Allan TF, Rose PE. Oral
anticoagulation management in primary care with the use of computerized
decision support and near-patient testing: a randomized, controlled trial. Arch
Intern Med 2000; 160: 2343-48.
18. Fitzmaurice DA, Murray ET, Gee KM, Allan TF. Does the Birmingham model of
oral anticoagulation management in primary care work outside trial conditions? Br
J Gen Pract 2001; 51: 828-29.
19. Ryan PJ, Gilbert M, Rose PE. Computer control of anticoagulant dose for
therapeutic management. BMJ 1989; 299: 1207-09.
20. Poller L, Shiach CR, MacCallum PK, Johansen AM, Munster AM, Magalhaes A et
al. Multicentre randomised study of computerised anticoagulant dosage.
European Concerted Action on Anticoagulation. Lancet 1998; 352: 1505-09.
21. Jones RT, Sullivan M, Barrett D. INRstar: computerised decision support
software for anticoagulation management in primary care. Inform Prim Care
2005; 13: 215-21.
22. Garg AX, Adhikari NK, McDonald H, Rosas-Arellano MP, Devereaux PJ, Beyene
J, et al. Effects of computerized clinical decision support systems on practitioner
performance and patient outcomes: a systematic review. JAMA 2005; 293: 122338.
23. Fitzmaurice DA, Hobbs FD, Delaney BC, Wilson S, McManus R. Review of
computerized decision support systems for oral anticoagulation management. Br
J Haematol 1998; 102: 907-09.
24. Chatellier G, Colombet I, Degoulet P. An overview of the effect of computerassisted management of anticoagulant therapy on the quality of anticoagulation.
Int J Med Inform 1998; 49: 311-20.
25. Parry D, Bryan S, Gee K, Murray E, Fitzmaurice D. Patient costs in
anticoagulation management: a comparison of primary and secondary care. Br J
Gen Pract 2001; 51: 972-76.
323
26. NHS
Economic
Evaluation
Database.
Available
from:
http://www.crd.york.ac.uk/crdweb/ShowRecord.asp?LinkFrom=OAI&ID=2200100
0113 (last accessed 31/12/10)
27. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al.
Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009;
361: 1139-51.
28. EINSTEIN Investigators, Bauersachs R, Berkowitz SD, Brenner B, Buller HR,
Decousus H, et al. Oral rivaroxaban for symptomatic venous thromboembolism.
N Engl J Med 2010; 363: 2499-510.
29. Connock M, Stevens C, Fry-Smith A, Jowett S, Fitzmaurice D, Moore D, et al.
Clinical effectiveness and cost-effectiveness of different models of managing
long-term oral anticoagulation therapy: a systematic review and economic
modelling. Health Technol Assess 2007; 11: 1-66
30. Refreshed
indicators,
December
2010.
Available
from:
http://www.ic.nhs.uk/webfiles/Work%20with%20us/consultations/CQI/IQI%20Wha
ts%20New%20docs/IQI_Update_Dec_2010.pdf (last accessed 31/12/10).
324
Chapter 12
Case study: eHealth-based approaches to reducing
repeat drug exposure in patients with known drug
allergies
Summary
•
There is increasing interest internationally in ways of reducing the high
disease burden resulting from errors in medicine management. Repeat
exposure to drugs to which patients have a known allergy has been a
repeatedly identified error, often with disastrous consequences.
•
Drug
allergies
are
immunologically-mediated
reactions
that
are
characterised by specificity and recurrence on re-exposure. These repeat
reactions should therefore be completely preventable.
•
There is at present insufficient attention being paid to studying and
implementing system-based approaches to reducing the risk of such
accidental re-exposure.
•
A number of eHealth-based interventions may in this context be used to
reduce risk of recurrent exposure, these including:
o ePrescribing
o bar-coding
o radio-frequency identification
o biometric technologies.
•
The most consistent evidence in this respect comes from interventions
which provide prescribers with real-time clinical decision support; whilst
promising, there is however as yet less clear evidence in support of
interventions aiming to enhance patient recognition such as bar-coding,
radio-frequency identification and biometric technologies.
•
ePrescribing support for drug allergies is now routine in UK general
practice. Its use to hospital prescribing now needs to be considered.
•
An important underlying limiting factor, however, is the need to improve
the accuracy of drug allergy coding in electronic records; this is because
ePrescribing systems are fundamentally dependant on the accuracy of
these coded patient data.
325
12.1 Introduction
12.1.1 Key concepts and definitions
An adverse drug reaction (ADR) is an undesired outcome attributable to a
drug. There are two main classes of ADRs recognised: Type A ADRs are
predictable and therefore potentially preventable, while Type B ADRs are
unpredictable and therefore far more difficult to prevent.1
Allergic reactions to drugs are a class of Type A ADRs, which are
immunologically-mediated, this reaction being directed either at the drug in
question or its breakdown product.2 The still widely used Gell and Coombs
classification system can be used to categorise allergic reactions on the basis
of the underlying mechanisms involved.3 4
These immunological mechanisms share the characteristic of specificity,
transferability and, crucially, a high risk of recurrence on re-exposure. Repeat
reactions in individuals with a previous reaction should therefore be
preventable. There are however often a number of important practical
challenges to establishing a diagnosis of drug allergy, these including the
need to obtain a detailed clinical history, the relatively few validated tests
available and the issue of possible cross-sensitivity, all of which often result in
considerable uncertainty in relation to securing a diagnosis of drug allergy.
Despite these difficulties, it is however often possible for clinicians to arrive at
a working diagnosis of drug allergy and it is these individuals–that is, those in
whom a working diagnosis of drug allergy has been made and documented in
the clinical records–that represent the focus of our deliberations in this case
study and who are thereafter described as having a known drug allergy.
12.1.2 Scale of the problem
Obtaining epidemiological data on the frequency of re-exposure to drugs in
those with known drug allergies is difficult for a number of reasons. Firstly,
there is a problem with under-reporting in schemes that collect information on
adverse drug reactions, such as the Yellow Card system run by the UK
Medicines and Healthcare products Regulatory Agency (MHRA),5 the US
326
Food
and
Drug
Administration’s
MedWatch
Program
(http://www.fda.gov/medwatch/index.html), or the Reporting and Learning
System (RLS)–the first national database of patient safety incidents.6
Secondly, retrospective reviews may not provide sufficient information to
clearly establish the background and context of the event raising the
possibility of significant misclassification errors. Thirdly, there is still
uncertainty among many patients and healthcare professionals as to which
reactions are truly allergic and which are, for example, due to intolerances
and other non-allergic mechanisms.
Despite these difficulties, data from studies conducted in US secondary care
suggest that 6-12 per cent of medication errors are due to patients receiving a
medication to which they were already known to be allergic.7
8 9
This
translates into between 90,000 and 180,000 episodes per year in the US
(based on the assumption that medication errors harm about 1.5 million
people per year). Research from US ambulatory care suggests similarly high
figures.10 Crucially, most instances of these repeat exposures have been
judged across studies as being preventable.9 10
More recent epidemiological data comes from an analysis of around 60,000
patient safety incidents reported to the RLS between 2005 and 2006.11 This
found that the administration of medication–despite a known drug allergy–was
the cause of 3.2% of medication-related patient safety incidents in hospitals
and of 2.6% of incidents in family practice settings. Almost one-third of these
resulted in harmful consequences to patients, including death in some cases.
Drugs most frequently involved in these incidents were found to be antibiotics
(in particular, penicillins), opiods and non-steroidal anti-inflammatory drugs
(NSAIDS). Moreover, the Department of Health’s (DH) report of medical
insurance claims found that 11 of 193 (5.7%) Medical Protection Society
claims (for general practice) and 25 of 234 (10.7%) Medical Defence Union
claims (for hospital care) related to allergic reactions to medication.12 Many of
these concerned antibiotics that had been prescribed to patients with known
allergies.11 Other drugs and agents that are commonly responsible for
327
triggering allergic reactions include antihypertensive agents, vaccines,
anaesthetic agents, anti-epileptic and anti-arrhythmia medications.
The conclusions drawn from these studies do however have to be treated with
some caution as data derived from retrospective studies of adverse drug
events may not be directly comparable to studies of global medication errors
or studies of medication administration. Nevertheless, a growing body of
international evidence suggests that accidental re-exposure in patients with
known drug allergies is one of the most common preventable medication
errors, sometimes with disastrous consequences for the patient (see Box
12.1). This appears to be true for both hospital and community care.
Of concern is that despite the frequency of accidental re-exposure and the
likely consequences, systems-based approaches to reducing the risk of
repeat exposure in individuals with known drug allergies are not being given
the attention they warrant, as evidenced, for example, by their failures to be
considered in key relevant international allergy guidelines.13 14
Box 12.1 Specimen cases reported to the Reporting and Learning
System 6
“GP [general practitioner] on call administered amoxicillin to the patient. A
short time later the patient collapsed in cardiac arrest. An ambulance was
called and a CPR [cardiopulmonary resuscitation] protocol followed. Whilst
CPR was being completed, it was ascertained that the patient had been
prescribed an antibiotic ... which a relative stated the patient had told the GP
that they were allergic to. (Reported as resulting in death)”
“A patient was prescribed and given a dose of amoxicillin. The patient
developed itching and felt unwell, and later collapsed requiring antihistamine
and adrenaline injection to treat the reaction. The patient was known to be
penicillin allergic but this had not been transcribed onto the new medicine
chart.”
328
“A patient with a penicillin allergy documented on their medicine chart and
who was wearing a red allergy alert wristband, was prescribed and given a
dose of piperacillin.”
12.2
The
theoretical
and
empirical
evidence
for
eHealth-based
approaches to reducing repeat drug allergy
As discussed in the preceding chapter, eHealth-based solutions are
particularly promising in reducing risk of re-exposure as they can help in
overcoming some of the underlying systemic failings, particularly in relation to
managing, processing, retaining and making accessible large amounts of
disparate data to multiple end-users.
Below we consider theoretical, and where available, empirical evidence of a
number of key eHealth-based approaches currently under investigation with
the aim of reducing the risk of re-exposure in individuals with known drug
allergies. In so doing, where possible, we focus on findings from studies
employing rigorous designs, but as this is very much an area still under
development–particularly in relation to some of the newer technologies–where
relevant experimental studies have yet to be conducted, we also draw on
evidence from formative work and/or research in other disease areas which
may be generalisable to the context of managing patients with known drug
allergies. A detailed summary of the studies on which we have drawn to
inform this case study are included in the Online Repository accompanying
the publication on which this chapter is based.15
12.2.1 Computer systems with warning messages
There is some encouraging evidence for the effectiveness of using computer
systems incorporating hazard messages that alert healthcare professionals to
patients’ allergies (see Figure 12.1). For example, Bates and colleagues
evaluated the effectiveness of ePrescribing alone and in combination with a
329
team intervention in preventing medication errors in a tertiary care hospital
over a six month period.16 In this study, clinicians used a computerised
physician order entry (CPOE) system, which involved electronically selecting
and ordering prescription requests that were chosen from a drop-down menu.
The system also had a component that checked for the most common drug
allergies. It was found that the ePrescribing system reduced known drug
allergy errors by 56% (P=0.009).
In a similar study in an intensive care unit, Evans and colleagues found that a
computerised decision support system (CDSS) programme, which was linked
to patient records and issued recommendations as well as warnings,
significantly reduced (from 146 to 35, P<0.01) the number of instances in
which drugs were prescribed to which the patient had a recorded allergy.17
Studies of ePrescribing systems in general practice have pointed in a similar
direction.10
Figure 12.1: Example of a drug allergy alert
Despite these encouraging results there are several issues with computerised
prescribing support tools that often prevent them for realising their potential.
330
Firstly, they require relevant and accurate drug allergy information to be
entered into the record in order to be effective. Secondly, the production of
hazard alerts of known drug allergies depends on the level of system
sophistication.16 Thirdly, studies have shown that there is often a lack of
training amongst healthcare professionals as to how best to use these
systems, sometimes resulting in a false sense of security if data on allergies
are, for example, entered into an inappropriate section of the records or as a
free text entry rather than as a code that will be recognised by the system.18
Fourthly, they often lack specificity, which can result in spurious hazard
messages being generated. This in turn may lead to frustration amongst users
with the consequent danger that hazard messages are accidentally overridden.19 Conversely, an over-reliance on warnings may contribute to errors
with clinicians “blindly” trusting system alerts.18 Fifthly, there are still
sometimes problems with the underlying design of the prescribing software,
as demonstrated in a study by Fernando and colleagues in which they tested
the safety features of four widely used family practice computer systems in
the UK and found that only three out of four systems produced alerts when
penicillin was prescribed to a dummy patient with a penicillin allergy
previously recorded in the system.20 Thus efforts still need to concentrate on
improving the safety features of such prescribing support systems in order to
further enhance their effectiveness. And finally, as discussed in earlier
chapters, it is still unknown whether this evidence of effectiveness from
“home-grown” systems studied in “bench-mark” institutions is generalisable to
routine care settings, which will in most cases be using commercial systems;
it is also unknown whether these positive impacts on errors in prescribing will
translate into actual improved outcomes for patients i.e. a reduced incidence
of reactions.
Kuperman and colleagues have recommended several approaches to ensure
more effective checking in relation to drug allergies.21 Included in these is the
suggestion that computer systems should store allergy information centrally in
order to ensure it is up-to-date and readily available in different settings (e.g.
family practice and hospital). The authors also suggest assigning levels of
importance to allergy warnings so that the likelihood of truly important alerts
331
being over-ridden will be minimised. Creating a common language for groups
of allergies and strategies to prompt healthcare professionals to enter all
allergic reactions into the system are additional helpful suggestions.
Investigations into the effectiveness and potential improvement of existing
computer systems incorporating hazard messages alerting healthcare
professionals to patients’ allergies are ongoing.
12.2.2 Bar-codes in wristbands
Another stream of efforts has focused on investigating the use of bar-coded
wristbands in hospital settings. While colour-coded wristbands can help to
identify a particular risk in a patient (such as a drug allergy), bar-coded
wristbands can provide more comprehensive patient-specific information on
the class of drug to which the patient is allergic. Here, the system includes a
bar-code reader that is connected to a host computer that retrieves the
desired patient information when the wristband is scanned. The scanned
information can then be checked against the medication packaging, which
may also have a bar-code. Although there is a broader literature surrounding
bar-codes on medication packaging, we focus here on bar-codes in
wristbands as this is most relevant for preventing repeat exposure in patients
with drug allergies.
A number of different wristband bar-coding systems have now been
developed and are at various stages of implementation in hospitals across the
world. For example, Franklin and colleagues have recently investigated the
effectiveness of what is described as a “closed-loop electronic prescribing and
administration system” in a surgical ward of a UK teaching hospital.22 This
involves giving patients bar-coded wristbands, which are connected to a
reader on a drug trolley. After the wristband is scanned, the trolley releases
medication through a drawer; details of items dispensed in this way are stored
on computer. This has been shown to significantly reduce prescribing and
medication administration errors; it also resulted in increased identity checking
of the patient by hospital staff.
332
The majority of studies investigating the effectiveness of bar-coded
wristbands come from the area of transfusion medicine. For example, an
investigation by the Department of Haematology at the Oxford Radcliffe
Hospital in the UK found that the introduction of patients wearing a bar-coded
identification wristband that could be scanned into a handheld computer for
compatibility resulted in a significant improvement in patient identification.23
When the same model was tested in two other hospitals, one in Italy and the
other in the US,24 similar results were obtained, thereby demonstrating the
potential generalisability of these findings.
Whilst promising, there is however as yet a paucity of high quality evidence
from randomised controlled trials (RCTs) for the effectiveness of bar-coded
wristbands. Systems have also so far not been tested specifically with regard
to reducing known drug allergies. In addition, some have questioned the
practicality of such devices. The National Patient Safety Agency (NPSA), for
example, refers to healthcare staff’s lack of compliance with using bar-coded
wristbands (see Box 12.1) and a lack of standardisation across institutions.25
There is furthermore a danger of patients inadvertently being given a
wristband belonging to another patient and the associated risk that may ensue
from this misidentification.26 Other issues include the problem of access as
wristbands need to be scanned with a handheld device from relatively close
proximity. This might under some circumstance prove problematic, for
example if the position of the patient does not permit scanning (e.g. they are
sleeping or unconscious) or if the handheld scanner is impractical to use (e.g.
in the operating room).27
Bar-code technology may also be used for Medic-Alert identification jewellery.
These are simple bracelets or necklaces with the medical condition of the
patient (e.g. allergy information) is recorded along with an identification
number and a phone number for someone who can be contacted in an
emergency. Although Medic-Alert bracelets are commonly recommended to
allergic patients and highly valued by allergists (especially in patients with
anaphylaxis), we are not aware of any studies evaluating their effectiveness.
A likely benefit of these bracelets is that they can be carried around by
333
patients and therefore used outside hospital settings. Nevertheless, it has to
be kept in mind that in addition to problems mentioned above, there may be
issues with compliance of wearing Medic-Alert bracelets.28
12.2.3. Radio-frequency identification
A more recent development still is radio-frequency identification (RFID).
These are chipped tags that are connected to a transceiver via radio-waves,
which allow the storage of information (such as drug allergies) remotely.
Radio-frequency identification chips allow more information to be stored than
bar-codes, are more user-friendly (as they can be scanned through fabric)
and are commonly regarded as more fail-safe (as they can work
independently of any user input). They are also relatively cheap at a cost of
between 5-10 pence each. Pilot studies in hospitals support the feasibility of
such systems, but as with bar-codes, more rigorous evidence of effectiveness
is still needed. Most activity in the developing and implementing RFID chips
can currently be found in the US. For example, the Jacobi Medical Center in
New York has already implemented a system using RFID tags in wristbands
to improve patient identification and reduce medication errors in two acute
care departments.29 Patients are issued with a so-called “Smart Band” on
admission, which can be scanned by a handheld reader that brings up all
necessary patient information on a screen at the bedside (see Figure 12.2).
This information can then be updated and modified as necessary. The Jacobi
Medical Center has reported a reduction of medication errors as well as staff
time associated with data entry as a result of introducing the system.
334
Figure 12.2: Handheld scanner attached to a screen and RFID tagged
wristband used at the Jacobi Medical Center in New York
(Adopted from http://www.ti.com/rfid/docs/news/eNews/enewsvol40.htm)
Again, there are several pilot studies using RFID chips to try to reduce errors
in blood transfusion. The San Raffaele Hospital in Milan (Italy) is currently
piloting a system, where RFID tags are incorporated into wristbands given to
donors.30 These are matched with a tag on the blood, which is then scanned
at the time of transfusion (see Figure 12.3).
Similar systems are also
currently being piloted in hospitals in the German Saarbrucken Clinic
Winterberg, in Washington and Boston in the US, in Taiwan and in England.31
32 33 34
Figure 12.3: Blood handling process using RFID technology at the San
Raffaele Hospital in Italy
335
(Adapted
from
Cisco
systems
at
http://www.cisco.com/web/IT/local_offices/case_history/rfid_in_blood_transfus
ions_final.pdf)
In the UK, the Birmingham Heartlands Hospital has been implementing RFID
wristbands in surgical wards after successful piloting. Here the whole
operating team has wireless personal digital assistants that can be used to
scan a patient’s wristband giving instant access to operating schedules and
patient records.35
Taking the use of RFID chips one step further, the Hackensack University
Medical Center in the US is currently piloting a patient identification system
using implantable RFID chips for patients with chronic conditions, which
provides
detailed
emergencies.
36
and
instantly
accessible
patient
information
in
A similar pilot has been started in the Alzheimer's Community
Care headquarters in the US.37 The advantages of these implants include the
comfort of wearing them and their potential application in the community. Here
the family physician will merely have to scan the patient, gaining instant and
up-to-date access to medical records. These implants do however raise
important ethical concerns and their effectiveness remains to be evaluated in
high quality clinical trials.
12.2.4. Biometric technologies
Another possible way to facilitate patient identification (and therefore potential
drug allergies) is the use of biometric technologies such as fingerprint, face or
iris scanning. These systems are not as invasive as implanted RFID chips and
have the distinct advantage of being easily and instantly accessible. Obviating
the need for patients to carry or wear an external object there are also no
issues with compliance. Such systems have already been piloted around the
world, again mainly in the US. The majority of applications have been
implemented in outpatient clinics helping to identify patients coming in to
collect medication. For example, the Netherlands are using fingerprint
scanners to identify heroin addicts who come to methadone distribution
336
services.38 Also, an increasing number of healthcare providers in the US are
using fingerprint scanning to prevent recipient and provider fraud.39 Moreover,
South Africa has launched a project allowing for patient identification using
fingerprints, and the “Methadose” system in Australia is scanning the irises of
patients who collect their methadone at St Vincent's Hospital in Sydney.40
In England, the Patient Access to Electronic Records System (PAERS) gives
patients electronic access to their medical records in general practices via
fingerprint scanning. The system is now implemented in some first-wave
practices across England.41
However, again robust scientific evidence of the effectiveness of such
systems is lacking. These remain as yet untested with regard to reducing
repeat exposure in patients with known drug allergies–an area in which their
application might have significant potential.
12.2.5 Patient-managed electronic health records
The approaches outlined above all involve some type of IT application. They
also rely, to a lesser or greater extent, on patient involvement and it is
important in this respect that parallel attempts are made to improve providerpatient communication and more actively involve patients (and carers) in their
care with a view to reducing such reactions.42 Patients, after all, have the
most to lose through accidental re-exposure and the most to be gained
through getting involved with care delivery in general. Patient-generated and
managed electronic health records will thus be particularly important in this
respect, with the onus on patients taking responsibility for ensuring that their
records are accurate and up-to-date and that they are not given treatments
that place them at unnecessary risk.43 Examples of the successful
implementation of personal health records include HealthConnectOnline in the
US, where patients can access and update their medical records and manage
treatments online (http://www.healthconnectonline.com/). A similar service
has been introduced in some parts of Europe (including Germany,
Switzerland,
Austria
and
Bulgaria)
under
the
name
LifeSensor
337
(https://www.lifesensor.com/us/us/), which is a web-based personal health
record accessible for both health professionals and patients. Patients in
England
can
use
HealthSpace
(https://www.healthspace.nhs.uk/visitor/default.aspx; see Chapter 3), a web-based
health organiser.
Although some concerns have been voiced regarding the accuracy of
personal health records and the limited usage to-date (in England),44 their
potential to help improve delivery of care is increasingly being recognised,
particularly if their introduction can be aligned better with the needs of patients
and care pathways adopted by professionals. In the context of patients with
known drug allergies, they can provide a valuable additional safety net for use
in combination with other established and emerging technologies discussed
above.
Which of the approaches or combinations of approaches discussed will
ultimately prove most effective in reducing instances of repeat exposure in
patients with known drug allergies remains to be determined. Given the
(relatively) strong empirical evidence of their impact on relevant surrogate
markers, it is important that ePrescribing systems are considered for
implementation, particularly in hospital settings to help reduce accidental
repeat exposure. Similarly, it is anticipated that patient-managed personal
electronic records will in due course occupy a more central position in the
delivery of care.45 Amongst the other emerging technologies discussed, RFID
chips and biometrics hold potential for the integration of allergy information
across care settings. But the lack of formal evaluations of the effectiveness of
such devices with regard to patients with known drug allergies, their high
capital costs and the considerable training needs of healthcare staff and
patients may slow developments in this area. Also, ethical issues, including
the accessibility of confidential information that patients carry with them at all
times may prove problematic.
338
12.3 Implications for practice, policy and research
There is an increasing body of epidemiological work indicating that adverse
drug reactions due to known drug allergies are relatively common and
potentially preventable.
Alongside the need for increased research into
investigating the underlying mechanisms and processes involved in allergic
reactions to drugs, and the accompanying need to develop improved
screening and diagnostic tests,4 there is, we believe, the pressing need to
investigate and, where found to be effective, implement systems-based
approaches to reducing these events into routine care. We have discussed
existing and emerging information technology-based interventions, which are
proven and have the potential to prove successful in minimising the risk of reexposure.
ePrescribing systems with carefully-refined alerts are likely to continue to play
an important role in general practice and the wider use of this technology
should also be considered in hospital practice, albeit within an evaluative
context.
Newer technological developments in this area as bar-codes in
wristbands and Medic-Alert bracelets, RFID chips, biometrics and patientmanaged electronic records all have potential, but the evidence-base is as yet
too immature to contemplate their widespread deployment. There is thus an
urgent need to evaluate the effectiveness of these newer technologies in the
context of managing patients with a history of drug allergy.46
Whilst these developments are important, they do not detract from the onus of
responsibility on the prescribing and administering clinician, and the patient
who agrees to take the drug.
Acknowledgements: This chapter is based on our publication: Cresswell KM, Sheikh A.
Information technology-based approaches to reducing repeat drug exposure in patients with
known drug allergies. J Allergy Clin Immunol 2008; 121: 1112-17. Permission to reproduce
this material has been applied for.
339
References
1. Adkinson NF, Jr., Friedmann PS, Pongracic JA. Drug allergy. In: Holgate ST,
Church MK, Lichtenstein LM, editors. Allergy. 3rd ed. London: Mosby Elsevier;
2006: 157-166.
2. Vervloet D, Durham S. ABC of allergies: Adverse reactions to drugs. BMJ 1998;
316: 1511-14.
3. Coombs RRA, Gell PGH. Classification of allergic reactions responsible for drug
hypersensitivity reactions. In: Coombs RRA, Gell PGH, editors. Clinical Aspects
of Immunology. Philadelphia, PA: Davis; 1968: 575-96.
4. Adkinson NF, Jr., Essayan D, Gruchalla R, Haggerty H, Kawabata T, Sandler JD
et al. Task force report: future research needs for the prevention and
management of immune-mediated drug hypersensitivity reactions. J Allergy Clin
Immunol 2002; 109: S461-78.
5. Monitoring the safety and quality of medicines: the Yellow Card Scheme. London:
Medicines and Healthcare Product Regulatory Agency. Available from:
http://www.mhra.gov.uk/index.htm (last accessed 30/12/10)
6. The
National
Reporting
and
Learning
Service.
Available
from
http://www.nrls.npsa.nhs.uk/report-a-patient-safety-incident/
(last
accessed
30/12/10).
7. Lesar TS, Briceland LL, Delcoure K, Parmalee JC, Masta-Gornic V, Pohl H.
Medication prescribing errors in a teaching hospital. JAMA 1990; 263: 2329-34.
8. Lesar TS, Briceland L, Stein DS. Factors related to errors in medication
prescribing. JAMA 1997; 277: 312-17.
9. Leape LL, Bates DW, Cullen DJ, Cooper J, Demonaco HJ, Gallivan T, et al.
Systems analysis of adverse drug events. ADE Prevention Study Group. JAMA
1995; 274: 35-43.
10. Gandhi TK, Weingart SN, Borus J, Seger AC, Peterson J, Burdick E. Adverse
drug events in ambulatory care. N Engl J Med 2003; 348: 1556-64.
11. National Patient Safety Agency. Safety in Doses: Medication Safety Incidents in
the NHS. London: National Patient Safety Agency, 2007.
12. Chief Pharmaceutical Officer. Building a Safer NHS for Patients - Improving
Medication Safety. London: Department of Health, 2004.
13. Bernstein I, Gruchalla RS, Lee R, Nicklas R, Dykewicz M. Executive summary of
disease management of drug hypersensitivity: a practice parameter. Joint Task
Force on Practice Parameters, the American Academy of Allergy, Asthma and
Immunology, the American Academy of Allergy, Asthma and Immunology, and
the Joint Council of Allergy, Asthma and Immunology. Ann Allergy Asthma
Immunol 1999; 83: 665-700.
14. Gomes RE, Pichler WJ, Demoly P, Aberer W, Frew AJ, de Weck A. The Drug
Ambassador Project - the diversity of diagnostic procedures for drug allergy
around Europe. Allergy Clin Immunol Int - J World Allergy Org 2005; 17: 9-18.
15. Cresswell KM, Sheikh A. Information technology-based approaches to reducing
repeat drug exposure in patients with known drug allergies. J Allergy Clin
Immunol 2008; 121: 1112-17.
16. Bates DW, Leape LL, Cullen DJ, Laird N, Petersen LA, Teich JM et al. Effect of
computerized physician order entry and a team intervention on prevention of
serious medication errors. JAMA 1998; 280: 1311-16.
17. Evans RS, Pestotnik SL, Classen DC, Clemmer TP, Weaver LK, Orme JF et al. A
computer-assisted management program for antibiotics and other anti-infective
agents. N Engl J Med 1998; 338: 232-38.
340
18. Morris CJ, Savelyich BSP, Avery AJ, Cantrill JA, Sheikh A. Patient safety
features of clinical computer systems: questionnaire survey of GP views. Qual
Saf Health Care 2005; 14: 164-68.
19. Hsieh TC, Kuperman GJ, Jaggi T, Hojnowski-Diaz P, Fiskio J, Williams DH et al.
Characteristics and consequences of drug allergy alert overrides in a
computerized physician order entry system. J Am Med Inform Assoc 2004; 11:
482-91.
20. Fernando B, Savelyich BSP, Avery AJ, Sheikh A, Bainbridge M, Horsfield P, et al.
Prescribing safety features of general practice computer systems: evaluation
using simulated test cases. BMJ 2004; 328: 1171-72.
21. Kuperman GJ, Gandhi TK, Bates DW. Effective drug-allergy checking:
methodological and operational issues. J Biomed Inform 2003; 36: 70-79.
22. Franklin BD, O'Grady K, Donyai P, Jacklin A, Barber N. The impact of a closedloop electronic prescribing and administration system on prescribing errors,
administration errors and staff time: a before-and-after study. Qual Saf Health
Care 2007; 16: 279-84.
23. Turner CL, Casbard AC, Murphy MF. Barcode technology: its role in increasing
the safety of blood transfusion. Transfusion 2003; 43: 1200-09.
24. Marconi M, Langeberg AF, Sirchia G, Sandler SG. Improving transfusion safety
by electronic identification of patients, blood samples, and blood units.
Immunohematol 2000; 16: 82-85.
25. National Patient Safety Agency. Standardising Wristbands Improves Patient
Safety. 2007. London: NPSA, 2007.
26. McDonald CJ. computerization can create safety hazards: a bar-coding near
miss. Ann Intern Med 2006; 144: 510-16.
27. Patterson ES, Cook RI, Render ML. Improving patient safety by identifying side
effects from introducing bar coding in medication administration. J Am Med
Inform Assoc 2002; 9: 540-53.
28. Siminerio L, Clougherty M, Gilliland A, Kelly K. Evaluating children with diabetes
and their parent's preparedness and involvement in diabetes care for the school.
Diabetes 2000; 49: A173.
29. PDC.
RFID
System:
Jacobi
Medical
Center.
Available
from
http://www.pdcorp.com/en-us/healthcare/case-study-jacobi-medical-center.html
(last accessed 31/12/10)
30. Dalton J, Ippolito C, Poncet I, Rossini S; 2005. Using RFID Technologies to
Reduce Blood Transfusion Errors. White Paper by Intel Corporation, Autentica,
Cisco
Systems,
and
San
Raffaele
Hospital.
Available
from
http://www.cisco.com/web/IT/local_offices/case_history/rfid_in_blood_transfusion
s_final.pdf (last accessed 28/12/10)
31. Wessel R. German Clinic Uses RFID to Track Blood. Available from
http://www.rfidjournal.com/article/articleview/2169/1/1/. (last accessed 30/12/10)
32. Dzik S. Radio frequency identification for prevention of bedside errors.
Transfusion 2007; 47: 125S-29S.
33. Precision Dynamics Corporation. RFID Technology Integration - 100% Accuracy
in Patient Identification; 2005. Available from http://www.pdcorp.com/enus/healthcare/case-study-chang-gung-hospital.html (last accessed 30/12/10)
34. Goddard K, Shabestari O, Kay JD, Roudsari A. Formatively evaluating the
importance of different aspects of an electronic blood transfusion system from the
end users' point of view: a questionnaire study. Stud Health Technol Inform
2010;160: 1211-15.
35. Bacheldor B. Tags Track Surgical Patients at Birmingham Heartlands Hospital;
2007. Available from http://www.rfidjournal.com/article/articleview/3222/2/1/ (last
accessed 30/12/10).
341
36. Medical News Today. Horizon Blue Cross Blue Shield Of New Jersey To Sponsor
Two-Year Pilot With Hackensack University Medical Center; 2006. Available from
http://www.medicalnewstoday.com/articles/47480.php (last accessed 30/12/10)
37. More RFID. VeriChip Corporation and Alzheimer's Community Care Officially
Launch VeriMed Patient Identification Project on August 28; 2007. Available from
http://www.morerfid.com/details.php?subdetail=Report&action=details&report_id=
3442 (last accessed 30/12/10).
38. Schmitz
PE.
Biometrics
in
Europe.
Available
from:
http://www.libertysecurity.org/IMG/pdf/Trend_Report_2006.pdf
(last accessed
31/12/10)
39. Kohl G. As records go digital, biometrics integrates into healthcare sector.
Available from: http://www.securityinfowatch.com/online/Biometrics/As-RecordsGo-Digital--Biometrics-Integrates-into-Healthcare-Sector/8706SIW417
(last
accessed 30/12/10).
40. Adwoa Serwaa-Bonsu,1 Abraham J. Herbst,2* Georges Reniers,3,4 Wilfred
Ijaa,5 Benjamin Clark,3 Chodziwadziwa Kabudula,3 and Osman Sankoh6. First
experiences in the implementation of biometric technology to link data from
Health and Demographic Surveillance Systems with health facility data. Glob
Health Action 2010; 3: 10.3402/gha.v3i0.2120
41. Head B. The eyes have it - Identification using biometrics is finding a range of
applications.
Available
from:
http://cio.co.nz/cio.nsf/depth/52FF4B475BD4DAFACC2573540082A2C7
(last
accessed 30/12/10)
42. E Health Insider. Patient information - a powerful prescription; 2006. Available
from:
http://www.e-health-insider.com/Features/item.cfm?docId=86 (last
accessed
30/12/10)
43. Cohen MR. Medication Errors: Causes, Prevention and Risk Management.
Sudbury, MA: Jones and Bartlett Publishers, 2000.
44. Greenhalgh T, Hinder S, Stramer K, Bratan T, Russell J. Adoption, non-adoption,
HealthSpace. BMJ 2010; 341: c5814.
45. Pagliari C, Detmer D, Singleton P. Potential of electronic personal health records.
BMJ 2007; 335: 330-33.
46. Catwell L, Sheikh A. Evaluating eHealth interventions: the need for continuous
systemic evaluation. PLoS Med 2009; 6: e1000126.
342
Chapter 13
Telehealthcare
Summary
•
Current systems of healthcare are inadequate to deal with rapidly
increasing populations, many of whom have one or more long-term
conditions.
•
Telehealthcare can be defined as “personalised healthcare delivered
over a distance.” It involves transfer of data from the patient to the
professional using information and communication technologies and
personalised feedback from the professional to the patient.
•
Whilst potentially attractive, these new models of care can however
prove very costly and may also cause a considerable change to
consultation dynamics and workflow processes; these should therefore
be introduced only after a careful assessment of effectiveness, costeffectiveness and safety considerations.
•
Most patients have in general a positive view of telehealthcare, so long
as it is offered in addition to, rather than as a substitution for, the faceto-face consultation. Qualitative studies have highlighted that the needs
of the carer should also be taken in account when designing
telehealthcare systems.
•
Overall, for people with long-term conditions such as asthma, COPD
and diabetes, there is now encouraging evidence that telehealthcare
can reduce hospital admissions in those with more severe disease
without increasing mortality; the evidence in relation to improving
quality of life in those with milder disease is less encouraging.
•
There is a need for more cost-effectiveness studies which take into
account the full range of perspectives i.e. that of the patient, healthcare
services and society.
•
Potential pitfalls of telehealthcare include user interface issues such as
older people not understanding how to operate systems, technical
breakdown, and safety concerns such as data loss and breach of
confidentiality.
343
Summary contdinued…
•
The use of telehealthcare often results in a reshaping of the doctorpatient relationship and attention should be paid to the opportunities for
humanising this interaction. Workflows also need to be considered to
minimise the risk of unintended disruptions to normal working routines.
13.1 Introduction
Telehealthcare has been used by healthcare professionals since the invention
of the telephone over 100 years ago, in Canada. Whenever practitioner and
patient are not physically present in the same place and some form of
telecommunications technology is used to enable the consultation, an action
of “telehealthcare” may be considered to have taken place. What is new is the
plurality of devices that now facilitate telehealthcare consultations.
Also new, is the increasing interest on the part of policymakers who now tend
to see telehealthcare implementations as central to their drive to modernise
health systems.1 This is because ageing populations and the related
inexorable increase in the numbers of people living for extended periods of
time with long-term conditions is placing a major financial and staffing burden
on health systems. Older people, have need of more medical and social
support to ensure that they are functioning to their maximum potential. This
change in demography has increased pressure on health and social care
services to manage more patients with fewer resources.
Health systems internationally are looking to solutions such as telehealthcare
in order to save time and money and improve efficiency. Telehealthcare
providers often promise that their products will do exactly this:2 however, it is
necessary to examine the evidence of the effectiveness of telehealthcare, its
risks and benefits and what is known regarding its implementation, before
drawing such conclusions. Telehealthcare is anticipated to have other benefits
as well, such as increasing patient satisfaction and even improving health
outcomes.
344
This chapter aims to provide an overview of telehealthcare, and the studies
undertaken so far which demonstrate the contexts and settings in which it is
likely to represent a useful development. There then follows a discussion of
policy, practice and research considerations in relation to telehealthcare. The
important question of how best to implement telehealthcare is discussed in
more detail in Chapter 17. The case study on telehealthcare in the context of
asthma in the following chapter (Chapter 14), provides an opportunity for a
much more detailed discussion of the opportunities and risks associated with
use of telehealthcare in the context of long-term conditions, which is a key
area of policy focus.
13.2.1 Definition
There are a multiplicity of definitions of telehealthcare and related terms. Here
we discuss what we mean by the term “telehealthcare” and what is meant by
related terms, telehealth, telemedicine, telecare etc.
Term
Origin
Definition
Telecare
Barlow et al ‘The use of communications technology to
20073
provide health and social care directly to the
user (patient). This excludes the exchange of
information
solely
between
professionals,
generally for diagnosis or referral.’
Telehealth
Glueckauf et ‘The
al 20024
use
of
telecommunications
and
information technologies to provide access to
health information and services across a
geographical
limited
to)
distance,
including
consultation,
(but
not
assessment,
intervention, and health maintenance.’
Telemedicine
Commission
(1)
of
European
‘The provision of health care services, through
the the use of information and communications
technology (ICT), in situations where the health
Communities professional and the patient (or two health
345
COM
professionals) are not in the same location. It
(2008)6895
involves secure transmission of medical data
and information through text sound, images or
other
forms
needed
for
the
prevention,
diagnosis, treatment and follow-up of patients.’
Telemedicine
Sood et al ‘Telemedicine being a subset of telehealth,
(2)
20076
uses communications networks for delivery of
healthcare services and medical education
from one geographical location to another,
primarily to address challenges like uneven
distribution and shortage of infrastructural and
human resources.’
Home
Paré et al7 ‘An automated process for the transmission of
telemonitoring 2007
data on a patient’s health status from home to
the respective health care setting.’
Table 13.1 Some examples of definitions of telehealthcare and related
terms
We prefer to use the term “telehealthcare”, rather than “telehealth” or
“telecare”, “telemedicine” or “home telemonitoring” as this incorporates both
the health or medical element of the interaction and the caring element
whether nursing care or social care. The term “telehealthcare” allows the
interaction to contain many elements without role-based delineations. One
might expect “telemedicine” to be delivered by a doctor, telecare by a carer, a
nurse or a social worker and telenursing by a nurse. Telehealthcare is a much
more flexible and encompassing term and so is preferred.
In this sense, “telehealthcare” is a pragmatic, rather than an academic term. It
synthesises across the health and social care boundary and implies the use of
distance communication technologies for supporting the care of people with
long-term conditions in an interactive manner.
346
Telehealthcare does, however, bracket some things outwith its definition.
Telehealthcare is not the use of distance technologies for inter-professional
communication,
(sometimes
described
as
business-to-business
or
B2B,communication–in contrast to business-to-consumer or B2C) or passive
information provision (as in traditional online health education or marketing
tools).
Used in this sense, “telehealthcare” thus refers to “the provision of
personalised healthcare over a distance”.8 It has the three following essential
components:9
1. Information obtained from the individual, whether voice, video,
electrocardiography, oxygen saturation or other similar data, that
provides information on their condition.
2. Electronic transfer of such information–over a distance–to a healthcare
professional.
3. Personalised feedback tailored to the individual from a healthcare
professional who exercises their clinical skills and judgement.
“Telehealthcare” is the provision of “healthcare”, “over a distance.” “Over a
distance” implies the delivery of healthcare by means of the use of some tool
of distance communication, which operates without the simultaneous physical,
geographical presence of the two or more participants in the healthcare
interaction.
The postal service, although a means of communication at a distance does
not use information technology (IT), but relies on another human acting as a
relay messenger and therefore it is conventional to exclude the postal service
from discussions on telehealthcare.
The telephone, sometimes referred to in the literature as the “Plain Old
Telephone System (POTS)”, to emphasise that telehealthcare is a novel
employment of what is now an established technology; provides an often
reliable means of providing telehealthcare. In addition, the Internet; email;
mobile networks; radio transmission and other wired and wireless networks;
347
short message service (SMS), i.e. text messaging by phone and instant
messaging over the Internet are all included in telehealthcare interventions.
Video-conferencing allows the addition of the visual element to the healthcare
interaction.
The most common model of telehealthcare is shown in the diagram below:
data flows firstly from the patient to the healthcare professional at a distance
by means of IT facilitating communication across a distance; the patient then
receives tailored advice.
Data
Patient:
At home,
Work,
Care
home,
In a car,
•
•
•
•
•
•
•
Telephone
Mobile phone
Email
Internet
Video-conferencing
Text
Electronic
monitoring devices
such as blood
glucose monitor,
peak expiratory flow
meter which relay
data via a modem
• Daily questionnaires
Healthcare
professional
in distant
office or
hospital
Tailored
advice
Figure 13.1 Diagrammatic representation of telehealthcare with flow of
information from the patient to professional across a distance with
personalised feedback
Telehealthcare can be delivered by both synchronous and asynchronous (i.e.
store-and-forward) technologies (see Figure 13.1). For example, telephone
and
video-conferencing
enable
consultations
in
real
time,
whereas
348
asynchronous communication includes, for example, storing two weeks of
spirometry results in a batch and then forwarding these on, via a modem, to a
healthcare provider who responds by email or telephone.
13.2.3 Scope for use
Telehealthcare has been piloted in many different situations, however, it
would seem that its greatest potential is in the management of chronic
conditions away from the settings of traditional healthcare such as general
practice or the hospital or some other clinic. It thus enables patients to be
seen and receive clinical input without incurring the time and cost of travel.
Even more important, the inconvenience of travel is avoided–this is especially
pertinent when considering future population demographics with an increased
proportion of frail individuals suffering from more that one long-term condition.
Such individuals may have several appointments a week with different
hospital departments for different healthcare needs. They may have mobility
and memory issues, which render them vulnerable when travelling and
increase the time it takes them to complete their normal activities of daily
living even before one takes into account the time required for healthcare,
physiotherapy, podiatry etc. Telehealthcare to assist such people to receive
their healthcare in their home or a familiar local clinic setting is potentially very
welcome from their perspective.
Similarly, people living in remote and rural areas distant from the nearest
setting of healthcare provision can use telehealthcare to enable consultations
from their own home.10 And where there is limited specialist coverage for rare
diseases, for example in paediatrics, expertise can be more easily accessed
by patients.
In terms of the definition above we are concentrating on telehealthcare as
represented in the systematic review (SR) literature, which focuses on
telehealthcare in relation to long-term conditions. Although access to
emergency services such as
NHS Direct and 999 calls fall within our
definition of telehealthcare, studies on these issues have not as yet been
summarised in the SR literature, hence they will not be considered further in
349
this chapter. Before moving on, it is important to note that this represents an
important gap in the evidence base for health systems. We do however
include the use of the telephone for delivery of telehealthcare in non-triage
situations.
Our focus on long-term conditions is also justified in terms of the international
policy context. For example, the 2008 Darzi report1 emphasised the need to
concentrate resources on planning for an increasingly elderly population with
increased frailty and multiple long-term conditions. Darzi was very clear that
there may well be untapped potential in telehealthcare with regard to the more
efficient and cost-effective management of limited resources to supply health
and social care for this population. With the change to the coalition
government, policies,11 such as decentralisation, increased emphasis on selfcare and driving-up efficiency are to an extent predicated on telehealthcare
delivering on some of its potential benefits. There may however be a danger
of adopting technology for its own sake, or because it is well-marketed.
13.3.1 Benefits
In theory, telehealthcare can have both preventive and supportive roles. An
individual with a long-term condition may be well enough to manage their
activities of daily living with minimal support, potentially delivered via
telehealthcare. In the event of an acute deterioration in their condition,
telehealthcare enables swift intervention to prevent a prolonged or deeper
deterioration and dependence. This rapid intervention is enabled by
telehealthcare’s potential to deliver an early warning to the healthcare
professionals to enable them to take the necessary action.
Other potential uses of telehealthcare fall into the category of “upstream”
interventions or, in medical terminology, “primary prevention” including for
example improved monitoring of glucose and of blood pressure to prevent
heart attacks and strokes. Secondary prevention is improved by ensuring
medical management is taken as prescribed.
350
The potential supportive role of telehealthcare is demonstrated by the help
people can get in interpreting their day-to-day symptoms (e.g. pain), readings
(such as blood pressure) and test results (e.g. blood glucose), all of which
may be improved and the patient’s anxiety alleviated when such problems are
discussed with healthcare professionals. In the absence of regular clear
guidance from health professionals regular blood glucose monitoring in Type
II diabetes leads to feelings of guilt and self-blame especially in women.
Patients attribute the monitoring readings to whether they have been “bad” or
“good” with their diet and ignore the contribution of factors such as exercise
and medication. Without supportive, ideally tailored, education “monitoring
fatigue” sets in and patients abandon monitoring. Patients question the point
of monitoring if they are not seeing positive changes as a result of it.12 In
addition, in the event of a deterioration in symptoms or readings, the need for
an increase in the intensity or frequency of clinical contact may be enabled
without the upheaval and stresses generated by a hospital admission. The
risks of cross-infection or loss of function which may also be associated with a
bed-bound hospital admission are also avoided. This type of management at
home may be more in keeping with the patient and their family’s own wishes
for privacy and dignity at a time of vulnerability.
It is also hoped that telehealthcare will enable care to be tailored to the
individual. There is a policy push towards more personalised care and the use
of telehealthcare may enable patients to have their individual needs assessed
and then met in a more personalised manner.13
13.3.2 Risks
In theory, there is a risk with telehealthcare that the level of care will be underestimated or over-estimated. Over-estimating the level of care that is required
is more likely during a preventive phase of management. There are risks that
patients find the care obtrusive, imposing on their independence with
reminders regarding what to eat, what to drink, what exercises to do and
when to take medication. This may undermine autonomy to vulnerable groups
of people.14 Smart homes which detect people’s movements from room to
room might be said to be looking out for elderly patients in case of falling,
351
however, some people feel as though their privacy is compromised and they
are “under surveillance”.15
16
Or, alternatively, there is a risk of medicalising
the spectrum of normality and reducing an individual to a “patient” in a “sick
role” when they were coping quite well alone and independently or with
traditional support.
Under-estimating the level of care is also possible, for example, when a
patient with an acute exacerbation of illness is managed at home where
traditionally they would be managed in hospital. If the illness is more unstable
and less predictable than at first thought and the patient lives a long way from
more comprehensive medical care, then there are potential risks to patient
safety. However, competent patients should be encouraged to weigh up these
benefits and risks and decide for themselves whether they would rather be at
home or in hospital according to their individual priorities.
Staff may have to spend additional time learning to use the telehealthcare
technology. Staff users may then find that the time required to operate the
new system on a day-to-day basis may be greater than the systems and
processes that were replaced. Also during replacement staff will have to
duplicate their workload as the two systems run in parallel. There is a risk that
staff may become demoralised and overburdened when trying to implement
an innovation that is not “fit-for-purpose”. Similarly, if the telehealthcare
operates as planned and allows increased access to staff by patients staff
may become overwhelmed with the increased demands on their time as
without further resources it may be difficult for them to provide an enhanced
service.
There is also the risk that the systems and processes of care will need to be
redesigned to incorporate the technology which can lead to realignment of
staff roles and political problems.
There are technical risks: of breakdown, of unreliable telephone/network
connection and of poor interoperability.13
352
Vast amounts of data will be collected to enable telehealthcare. On the one
hand, if granted access, researchers may be able to interpret these data and
increase understanding of diseases to the benefit of society. On the other
hand, other groups may request access to these data for example in the
United States (US) there is a request to enable targeted marketing by
commercial companies for monetary gain, without the explicit consent of the
patients. Specifically, insurance and healthcare marketing may be anticipated
to be interested in such databases.17
Healthcare is a safety critical infrastructure industry, meaning that its smooth
running is imperative to the integrity of the nation. If vast numbers of
vulnerable people are living at home dependent on telehealthcare and there is
a sudden problem with the telephone or mobile network which links them,
then services will no longer be at a scale that can cope in looking after such
people. Similarly if the Internet server has a problem, there is a power failure
or a cyber-attack many people may be left without services they rely on. This
possibility demands careful contingency planning, including planning for more
than one problem occurring at a single point in time.17
We identified 41 relevant SRs3
7 18-56
investigating use of a variety of
telehealthcare interventions in different contexts and settings. These ranged
from connections between hospital specialist services and patients’ homes to
Internet discussion groups which remotely connected several peer patients
and a professional facilitator who may be based in the community or in a
specialist service, and included primary studies employing different designs. A
fuller description of these SRs and our assessment of their quality is available
in Appendices 4 and 5. Below, we provide a synopsis of the main findings to
emerge from this body of work.
13.4.1 Benefits
Satisfaction
There is evidence that patients are largely satisfied with telehealthcare. For
example, patients reported high levels of satisfaction and empowerment in
353
two overlapping systematic reviews of telehealthcare by the same author.7 57
Satisfaction tended to be reported on a Likert (or similar) scale.36
37 46
However, it may be that qualitative evidence gives greater insights regarding
how to achieve and maintain satisfaction. For example, participants in a focus
group discussed telehealthcare and agreed that it remained acceptable as
long as it was not compulsory. They wanted to still be able to have the option
of face-to-face care.9 A way forward may be to have the first visit/assessment
face-to-face with the option of telehealthcare follow-up.
Effectiveness
The effectiveness of telehealthcare is often assumed by industry and
policymakers. However, when scrutinised, research on effectiveness depends
very much on the context of the introduction of a specific telehealthcare
system. It is thus crucial that the aim of introducing the telehealthcare is clear
from the start.58 For example, is the aim to widen access, improve clinical
endpoints, aid the early detection of disease exacerbations and reduce the
risk of hospital admissions, reduce mortality, or is it to reduce the degree of
dependency in old age? These aims are not necessarily mutually exclusive,
but agreement on why telehealthcare has been introduced, followed by
focused implementation, and then evaluation to see if the specific need has
been addressed should help to determine if these aims have been realised.
Improving access
Access to specialist care in rural locations was enabled by telehealthcare
which improved the quality of care, especially for rare diseases.38 Teleradiology improved access to radiological services, saving transfer time, cost
and inconvenience.38 Likewise access to renal dialysis services was improved
with telehealthcare, especially as patients require several treatments a week.
There was also improved access to cardiology, colposcopy, neuropsychology,
minor injuries, dermatology and nutrition services.38 Another review found that
using the web as the medium of delivery improved access to the benefits of
tailored therapies for mental health problems regardless of geographical and
temporal barriers.43
354
Similarly, using the telephone improved accessibility to primary care asthma
reviews in one study, with the authors concluding that it was possible to
review almost double the number of patients via telephone in comparison to
face-to-face assessments.59 This telephone-based approach, with limited
start-up costs, was also found to be cost-effective.60
Further supporting evidence comes from a review, which reported that 20% of
included studies had found an improvement in access.31 And so overall it
appears that
telehealthcare can, in carefully selected contexts, improve
access to care.
There is some suggestive evidence that access can also be enhanced for
carers. For example, peer support and relief of social isolation have been
reported, both for people with mental health difficulties and for carers.
Improving clinical endpoints
The evidence is less clear cut regarding the improvement of clinical
endpoints. Firstly, attention should be drawn to the fact that many of the
clinical endpoints reported in reviews are surrogate markers rather than “hard”
clinical outcomes such as mortality. For example, telehealthcare may be
shown to increase adherence to therapy and although it is assumed that this
is beneficial, studies rarely mentioned adverse events and side-effects.
Similarly, although slight improvements in blood pressure (BP),49 glycated
haemoglobin (HbA1c)50 and lipid profiles have been demonstrated, studies
were neither powerful, nor long enough to demonstrate a translation of these
results into fewer heart attacks, strokes or other relevant cardiovascular
morbidity.
Despite this difficulty some improvements were demonstrated, but not
consistently. One difficulty is in deciding whether the comparisons were valid.
Many studies used a before-and-after design in which it is possible that only
regression to the mean is being detected or the natural fluctuations of the
illness. Where comparisons were made across telehealthcare and control
groups, the groups were often not significantly different at the end of the
355
study, which may be taken to mean that patients were not significantly worse
off for receiving telehealthcare, however, not significantly better either.
In terms of considering whether a comparison is valid it is worth pondering on
what is the best control. For example, for depression, one study compared
usual care with cognitive behavioural therapy via telephone. However, it was
not stipulated what the control consisted of. Did they receive any cognitive
behavioural therapy, or perhaps a lesser form of counselling? How often?
From whom? Did they receive medication? Were they given a leaflet? These
variables all need to be considered in deciding whether telehealthcare has
helped or hindered a patient’s recovery or deterioration.
A wide range of disease specific clinical endpoints were studied, however,
they each need to be considered in context given the caveats above. Even
then, the literature is somewhat inconsistent. It may be anticipated that
increased frequency and intensity of therapy delivered by whatever means
whether via telehealthcare or face-to-face would correlate with improvement.
However, one then needs to factor in side effects, quality of life, dependence
and loss of self-efficacy to complete the complex picture. In this review,
overall answers remained elusive.
Reducing mortality
There was very little data on mortality. A meta-analysis of telehealthcare
interventions in coronary heart disease found a non-significant lower all cause
mortality compared with controls (RR 0.70, 95%CI 0.45, 1.1).49 A review of
studies in patients with chronic obstructive pulmonary disease (COPD) found
increased mortality in the telehealthcare arm; however, we believe this is due
to a data entry error and this is discussed in our forthcoming Cochrane
review.
Reducing risk of hospital admission
Our systematic review (see Chapter 14) identified a number of telehealthcare
trials in asthma.45 It showed that hospital admissions can be reduced in
356
carefully selected patients who have severe asthma and/or have recently
been admitted to hospital and are consequently at high risk of re-admission.45
A reduction in the number of patients who were hospitalised, the number of
hospitalisations and the length of stay associated with these admissions was
also found with telehealthcare in diabetes.50 Similarly, monitoring of chronic
heart failure and pacemaker patients reduced hospital readmissions and
emergency department visits at a similar cost to conventional interventions.38
Therefore in selected high risk subgroups with appropriate conditions it is
possible to reduce hospital admissions.
Cost-effectiveness
Studies which considered cost savings generally had methodological flaws in
that they looked only at list prices and did not calculate opportunity costs. In
general, cost savings were analysed from the perspective of the healthcare
body, with the greatest saving coming from the prevention of hospital
admission.55 There were very few cost-effectiveness studies reported in terms
of Quality Adjusted Life Years(QALYs).56 Interventions which may have
proved cost-effective over the long term were generally not followed up for
long enough to derive any reliable estimates.45
Overall, in the absence of a convincing impact on mortality or morbidity, at this
moment in time, it appears that telehealthcare may only prove cost-effective in
carefully considered clinical situations in which there are substantial
opportunities for reducing risk of premature death or hospitalisation.
The
relative costs of the telehealthcare intervention adopted is also relevant.
13.4.2 Risks
Confidentiality
There has been much publicity recently concerning the loss of confidential
data from medical and other record systems.61 62 Such events become more
likely if people’s sensitive medical data are available on networks and mobile
storage devices which are not protected with encryption facilities. In addition,
with the increased use of mobile phones, people often phone the doctor from
357
their place of work, or when in a public place, this can lead to problems with
private conversations being overheard. There is some evidence to show that
in situations in which doctors call patients unexpectedly, then there may be
situations in which their privacy may at times be compromised through being
overheard and that patients may be inclined to answer sensitive questions in
“yes/no” terms.63
With telephone technology older patients with problems with vision and
hearing may need a younger person to help to make a telephone call, this
leads to problems of third party confidentiality, which are often ignored by
professionals.63 The use of online technology has also been associated with
significant breaches of confidentiality, for example, through the Kaiser
Permanente Internet Patient Portal, when details of e-mail conversations
between patients and their doctors were accidentally sent to multiple other
patients.64 There needs to be reporting of such events so that organisational
processes can respond to minimise repeat events and a “safety culture” can
be encouraged in healthcare organisations.
Human contact
Contact with machines which look after people in place of human and family
carers can result in loneliness, especially acute in those who have lost a
spouse where there may be a risk of over-reliance on devices. Rather than
empowering patients to become more active in looking after themselves,
telehealthcare may result in an increase in the patient’s passivity.65 In
addition, older people with long-term conditions, reduced dexterity, poor
eyesight and deteriorating cognition sometimes have difficulty operating
modern technology interfaces without help from a younger person.65
13.5.1 Policy
It used to be that physical co-location was a prerequisite for the definition of a
healthcare consultation according to various insurance companies within the
US and the European Union (EU). This has changed and now the European
Commission provides some guidance regarding the quality of the remote
358
healthcare encounter and the safety nets required.5
66
Within the UK, the
Medical Defence Union however reminds doctors that they are contractually
obliged to enable a physical examination where necessary.67 Until GP
contracts catch up with the implications of telehealthcare, it seems unlikely
that professional bodies such as the British Medical Association (BMA) will
support the introduction of telehealthcare. Doctors and nurses need to ensure
that they are professionally insured for any increased risk in caring for patients
via telehealthcare.
The General Medical Council (GMC) in the UK describes how best to carry
out and document a remote prescription. Although this guidance was written
with regard to telephone consultations, it can also be applied to email, Internet
and video-consultations.68
Early signs are that the UK’s coalition government is enthusiastic regarding
telehealthcare and the White paper ‘Equity and excellence: Liberating the
NHS’11 and the consultation document ‘An Information Revolution’69 discuss
the potential of telehealthcare in very optimistic terms. Its themes include the
following:
•
Consortia of General Practitioners will be able to commission new
ways of delivering care, including telehealthcare, in order to improve
services and manage the increasing need for healthcare more
efficiently.
•
Patients will have the opportunity to choose the providers of their
health
and
geographical
social care. This
boundaries
so
will include flexibility regarding
that,
especially
in
the
case
of
telehealthcare, there will be fewer restrictions.
•
Health and social care delivery will overlap to some degree depending
on local needs, including a single telephone number for every kind of
urgent health and social care.
•
Patients will have access to their own health records and be
participants in the forming of the health record. They may also choose
to share their health records with third parties such as charities,
359
advocacy organisations and information management service providers
and interpreters.
•
Determination of central standards by the NHS Commissioning Board
regarding record keeping, data sharing capabilities and efficiency of
data transfer. This should enable locally commissioned systems to be
interoperable.
•
An emphasis on new types of information such as quality standards for
patients to help them make choices, for clinicians as feedback on their
performance and for the public to hold the government to account
regarding improved service delivery.
•
The use of personal health budgets and an emphasis on responsibility
for increased self-care for long-term conditions.
These over-arching themes will have an important impact on how quickly
telehealthcare provision is scaled up. There is a strong policy drive but many
aspects remain to be thought through. For example, if GP consortia working in
partnership with local authorities are able to commission telehealthcare
services according to their patient population, this will result in many different
providers and technology packages. The government anticipates that
competition between providers will drive innovation and lower prices for
commissioners. However, the need for interoperability is crucial in order to
enable many of the presumed benefits of telehealthcare (see Chapter 5) and
this can only be delivered with careful technical specification during
procurement procedures.
The implications of our research on policy appear to be that using the
telephone to deliver telehealthcare is the most mature technology. It is
certainly the best understood in terms of the way in which it works and its
potential shortcomings and challenges. Scaling up of different telehealthcare
projects should be done with the uncertainties and expectancies of new
technology, firmly in mind. The use of the telephones should be encouraged
and the use of more complex telehealthcare equipment delayed until further
large scale trial results are available.
360
13.5.2 Practice
Introducing telehealthcare into practice involves a number of important
considerations. Firstly, this review has found that the most robust evidence
exists
for
telephone
consultations
and
follow-up.
This
medium
of
telehealthcare has been around the longest and is the most well studied. It is
cheap to provide and access for both patients and clinicians. It does not
necessitate expensive new technology, but only uses reliable and widely
available technology. For example, when telephones are used to provide
asthma review appointments a greater proportion of the eligible patient
population are reviewed and the intervention is cheaper than a face-to-face
review and overall more cost-effective.70 Telephone follow-up after hospital
discharge, initiated by a hospital team, has the potential to replace follow-up
face-to-face visits and help address post-discharge problems.46 However,
when telephone consultations were used as an alternative to acute general
practice face-to-face consultation the evidence is that telephone consultations
were shorter, presented fewer problems and included less data gathering,
counselling and rapport-building than traditional face-to-face consultations.71
This comparison also found that doctors conducting telephone consultations
were less likely to have gathered sufficient information to exclude important
serious illnesses.71 A further study found that rather than replacing face-toface consultation the encounter was postponed as patients managed by
telephone were 50% more likely to re-consult in the following two weeks and
so an inefficiency is introduced into the workflow, although the precise
reasons for this repeat consultation remain to be clarified.72 Such evidence
casts doubt on whether to use telephone consulting for triage of acute
emergencies.
Patients have had little say in the widespread adoption of telephone
consulting. There are differing contextual reasons for its use, including access
targets, which, when combined with high demand for appointments and
inflexible working practices, resulted in an increased use of telephone
appointments in urban practices. Whereas, in rural practices, the telephone is
used to overcome distances and maintain continuity of care.73
361
Alternatives to telephone appointments include workflows designed around a
patient with a telehealthcare device which accesses directly to secondary
care. These have important implications for practice. Consider, for example, a
programme in which patients monitor their own weight, blood pressure and
record their own electrocardiogram to transmit to a hospital-based specialist
nurse. The nurse, in consultation with a cardiologist, provides tailored advice
to the patient over the telephone on managing their heart failure.74
•
There is a change in the role and responsibility of the nurse; they are
involved increasingly in decision-making and so there is a need for
training to reflect this.
•
The GP is somewhat marginalised as patients with increasingly
complex management regimens, default to a trusted specialist
relationship.
•
There is a need to maintain accountability, protocols and guidelines
proliferate for case managers, describing how much they can
undertake themselves and when to involve a senior medical
practitioner.
•
There is an issue regarding multi-morbidity, when patients have more
than one long term condition. Patients may find that they receive
conflicting and complex advice from multiple specialists, however, their
GPs are no longer well placed to guide them holistically when
continuity of care is ruptured and they also provide predominantly
telehealthcare care.
13.5.3 Research
As can be seen from the summary of SRs on telehealthcare, there is much
about this technology which is promising. However, many aspects of the
technology remain poorly researched. This is accompanied by a pressure for
implementation for the demographic reasons given at the beginning of this
chapter. There is also pressure from commercial companies to implement
their telehealthcare solutions while regulation remains minimal in order to
maximise their opportunity to profit financially.
362
A recent independent industry report, ‘Healthcare without walls’75 proclaims
that ‘telehealth can enable an incredible 40% reduction or more in the level of
unplanned hospital admissions’. There is no citation for this figure and it does
not mention whether this is an absolute or relative reduction. The brief
summaries of evidence cited in the appendices are the only quantitative data
included in this report. These data are presented out of context and without
discussion of the limitations and accuracy of the figures. No randomised
control trial evidence is cited. The report recommends a rapid broad strategic
“push” for telehealthcare by addressing a business case to primary care
providers before they become consortia of general practitioners. It is proposed
that telehealthcare be adopted at scale by the NHS in order to share best
practice and ensure efficient procurement and economies of scale. We feel
that the note this paper strikes is very optimistic and somewhat premature.
We do not know enough about telehealthcare models of care yet to anticipate
that adoption at scale would not be an expensive catastrophe. The report from
the Whole System Demonstrator Project76 is eagerly awaited this year (2011)
to help to fill this knowledge gap. However, it is clear that telehealthcare
should be studied in projects of greater size than the large number of
published small studies e.g. 500+ participants before such widespread
adoption goes ahead.
In future, research needs to concentrate on large comparative studies
between telehealthcare options and usual care controls. The studies need to
fully describe and report both controls and interventions. The controls need to
be real-world, usual-care controls in order to allow meaningful comparisons –
to see if the introduction of telehealthcare would result in a significant
improvement to care. Understandably, the validity of many studies is
compromised by enhancements in the control arm; this does however make
interpreting this evidence difficult. Such studies should be randomised
controlled trials with follow-up of at least 6-12 months, preferably longer.
It is also important that studies include meaningful clinical endpoints rather
than surrogate and process measures, wherever possible. For example, the
coronary heart disease trials need longer follow-up and comparison of the
363
incidence rates of cardiac and cardiovascular morbidity and readmission
rates. Such clinical endpoints are necessary to allow clinicians to decide if
telehealthcare does indeed improve outcomes for patients or whether there is
a risk that the reduction in the intensity of care of telehealthcare causes more
problems than it solves.
One recent large study which fulfilled these criteria of large study population
and meaningful clinical endpoints was a study of 1653 patients with heart
failure.77 This followed a pilot study where a highly skilled and motivated
specialist nurse case manager had help to reduce readmission rate by 44%.78
The authors were very concerned regarding the generalisability of this pilot.
Their misgivings were confirmed in this large multicentre heart failure trial
when they found no reduction in the risk of readmission or death from any
cause with telehealthcare as compared with usual care. This study77 had
overcome the methodological challenges listed above and provides a high
quality contradiction to the systematic review on heart failure by Inglis et al.,
which pooled a number of lower quality studies and found a positive effect of
telehealthcare.36
There is furthermore a need to find effective ways of safely and reliably
integrating the vast amounts of data generated by telehealthcare initiatives
into electronic patient records, rather than as stand-alone datasets.
In addition, there needs to be improved study of the cost-effectiveness of
telehealthcare. Accurate costs from different perspectives, including that of
the healthcare provider, the patient and society should be combined with cost
per QALY analysis to inform the incremental adoption of telehealthcare
techniques into existing and adapted workflows. When resources are tight it is
important that the costs of every new technology are substantially justified in
terms of proven clinical benefit.
13.6 Conclusions
Telehealthcare is widely anticipated as the future of healthcare internationally
in order to relieve demographic pressure on healthcare services. There are
364
many potential benefits from “personalised healthcare, over a distance”,
including tailoring advice, and reducing time and cost of travel. Access can be
enabled for those in remote and rural areas and for frail and vulnerable
populations, in particular, those suffering from long-term conditions. Support
in managing long-term conditions should prevent deterioration and reduce the
risk of hospital admission, which may in turn result in cost savings.
However, will this lead to dependency on cold technology in place of warm
human interaction? Or to underestimating the seriousness of an acute illness?
Will systems be easy for staff and patients to operate and if so will access to
healthcare be so easy as to overwhelm the ability of staff to respond? What of
the data generated by the system and its related security and privacy issues?
These are some of the more pressing questions raised.
The SRs we identified have demonstrated that telehealthcare can have a
measurable impact in certain specific contexts including long term conditions,
patient assessment and social outcomes. Cost-effectiveness has not yet been
demonstrated and further research in this area is urgently required.
There is a very focused policy push for telehealthcare including modern digital
technologies and the Internet, but the best evidence is for the telephone. The
widespread introduction of telehealthcare into practice will engender major
changes
to
the
roles,
responsibilities and
workflows
of
healthcare
organisations and some of these are discussed further in Chapter 17.
Finally, more, longer term, larger research studies are required in order to
address these uncertainties so that the potential benefits will be realised if
telehealthcare is implemented. Without such studies, implementation of
telehealthcare risks being expensive, inefficient, and even dangerous to
vulnerable patients.
Acknowledgments: This chapter is partially based on: McLean S, Protti D, Sheikh A.
Telehealthcare for long-term conditions.
BMJ (in press). Permission to reproduce this
material has been granted.
365
References
1. Darzi A. High Quality Care for All: NHS Next Stage Review Final Report.
Department of Health 2008
2. Cisco. United Healthgroup, Cisco Launch Connected Care, First National
Telehealth Network. http://newsroom.cisco.com/dlls/2009/prod_071509b.html
2009
3. Barlow J, Singh D, Bayer S, Curry R. A systematic review of the benefits of
home telecare for frail elderly people and those with long-term conditions. J
Telemed Telecare 2007;13(4):172-9
4. Glueckauf R, Fritz S, Ecklund-Johnson E, Liss H, Dages P, Carney P.
Videocounseling for rural teens with seizure disorders: Phase I findings.
Rehabilitation Psychology 2002;47(49-72)
5. European-Commision. Communication from the commission to the European
Parliament, the Council, the European Economic and Social Committee and
the Committee of the regions on telemedicine for the benefit of patients ,
healthcare systems and society. 2008:1-13
6. Sood S, Mbarika V, Jugoo S, Dookhy R, Doarn CR, Prakash N, et al. What is
telemedicine? A collection of 104 peer-reviewed perspectives and theoretical
underpinnings. Telemed J E Health 2007;13(5):573-90
7. Pare G, Jaana M, Sicotte C. Systematic Review of Home Telemonitoring for
Chronic diseases: the evidence base. Journal of the American Medical
Informatics Association 2007;14:269-277
8. McLean S, Sheikh A. Does telehealthcare offer a patient-centred way forward
for the community based management of long-term respiratory disease?
Primary Care Respiratory Journal 2009;18(3):125-126
9. Miller E. Solving the disjuncture between research and practice, Telehealth
trends in the 21st Century. Health Policy 2007;82:133-41
10. Rygh EM, Hjortdahl P. Continuous and integrated health care services in rural
areas. A literature study. Rural Remote Health 2007;7(3):766
11. DoH. Equity and excellence: Liberating the NS: Department of Health, 2010.
12. Peel E, Douglas M, Lawton J. Self monitoring of blood glucose in type 2
diabetes: longitudinal qualitative study of patients' perspectives. BMJ
2007;335:493-8
13. Goodwin N. The State of Telehealth and Telecare in the UK: Prospects for
Integrated Care. Journal of Integrated Care 2010;18(6):3-10
14. Hensel BK, Demiris G, Courtney KL. Defining obtrusiveness in home
telehealth technologies: a conceptual framework. J Am Med Inform Assoc
2006;13(4):428-31
15. Dinesen B, Nohr C, Andersen SK, Sejersen H, Toft E. Under surveillance, yet
looked after: telehomecare as viewed by patients and their spouse/partners.
Eur J Cardiovasc Nurs 2008;7(3):239-46
16. Hughes JC, Louw SJ. Electronic tagging of people with dementia who
wander. BMJ 2002;325:847-8
17. Johnson CW. Politics and Patient Safety Don't Mix; On the Interaction
between Patient Safety , Politics and Software Engineering.
18. Akesson KM, Saveman B-I, Nilsson G. Health care consumers' experiences
of information communication technology--a summary of literature.
International Journal of Medical Informatics 2007;76(9):633-45
19. Bailey Julia V, Murray E, Rait G, Mercer Catherine H, Morris Richard W,
Peacock R, et al. Interactive computer-based interventions for sexual health
366
promotion. Cochrane Database of Systematic Reviews. Chichester, UK: John
Wiley & Sons, Ltd, 2010.
20. Balas EA, Jaffrey F, Kuperman GJ, Boren SA, Brown GD, Pinciroli F, et al.
Electronic communication with patients. Evaluation of distance medicine
technology. JAMA 1997;278(2):152-9
21. Bartoli L, Zanaboni P, Masella C, Ursini N, x00F. Systematic review of
telemedicine services for patients affected by chronic obstructive pulmonary
disease (COPD). Telemedicine Journal & E-Health 2009;15(9):877-83
22. Bessell TL, McDonald S, Silagy CA, Anderson JN, Hiller JE, Sansom LN. Do
Internet interventions for consumers cause more harm than good? A
systematic review. Health Expectations 2002;5(1):28-37
23. Biem HJ, Turnell RW, D'Arcy C. Computer telephony: automated calls for
medical care. Clinical & Investigative Medicine - Medecine Clinique et
Experimentale 2003;26(5):259-68
24. Boren SA, Puchbauer AM, Williams F. Computerized prompting and feedback
of diabetes care: a review of the literature. Journal of Diabetes Science &
Technology 2009;3(4):944-50
25. Bowles KH, Baugh AC. Applying research evidence to optimize
telehomecare. Journal of Cardiovascular Nursing 2007;22(1):5-15
26. Brebner JA, Brebner EM, Ruddick-Bracken H. Accident and emergency
teleconsultation for primary care--a systematic review of technical feasibility,
clinical effectiveness, cost effectiveness and level of local management.
Journal of Telemedicine & Telecare 2006;12 Suppl 1:5-8
27. Bunn F, Byrne G, Kendall S. Telephone consultation and triage: effects on
health care use and patient satisfaction. Cochrane Database of Systematic
Reviews 2004(4):CD004180
28. Currell R, Urquhart C, Wainwright P, Lewis R. Telemedicine versus face-toface patient care:effects on professional practice and health care outcomes. .
Cochrane Database of Systematic Reviews 2000(2)
29. Dellifraine J, Dansky K. Home-based telehealth: a review and meta-analysis.
Journal of Telemedicine and Telecare 2008;14:62-66
30. Demiris G, Hensel BK. Technologies for an aging society: a systematic review
of "smart home" applications. Yearbook of Medical Informatics 2008:33-40
31. Durrani H, Khoja S. A systematic review of the use of telehealth in Asian
countries (Structured abstract). Journal of Telemedicine and Telecare,
2009:175-181.
32. Duvvuri V, Jianhhong W. Information and communication technology
developments in asthma management: a systematic review. Indian Journal of
Medical Science 2007;61(4):221-241
33. Eysenbach G, Powell J, Englesakis M, Rizo C, Stern A. Health related virtual
communities and electronic support groups: systematic review of the effects
of online peer to peer interactions. BMJ 2004;328(7449):1166
34. Hersh WR, Helfand M, Wallace JA, Kraemer DF, Patterson P, Shapiro S, et
al. Clinical outcomes resulting from telemedicine interventions: a systematic
review. BMC Medical Informatics & Decision Making 2001;1(5)
35. Hersh WR, Wallace JA, Paterson PK, Shapiro SE, Kraemer DF, Eilers GM.
Telemedicine for the medicare population: Paediatric, obstetric and clinician
indirect home interventions. Evid Rep Technol Assess (Summ) 2001;24
Suppl:1-32
36. Inglis Sally C, Clark Robyn A, McAlister Finlay A, Ball J, Lewinter C,
Cullington D, et al. Structured telephone support or telemonitoring
programmes for patients with chronic heart failure. Cochrane Database of
Systematic Reviews. Chichester, UK: John Wiley & Sons, Ltd, 2010.
37. Jackson CL, Bolen S, Brancati FL, Batts-Turner ML, Gary TL. A systematic
review of interactive computer-assisted technology in diabetes care.
367
Interactive information technology in diabetes care. Journal of General
Internal Medicine 2006;21(2):105-10
38. Jennett P, Hall L, Hailey D, Ohinmaa A, Anderson C, Thomas R, et al. The
socio-economic impact of telehealth: a systematic review. Journal of
Telemedicine and Telecare 2003;9:311-320
39. Jones JF, Brennan PF. Telehealth interventions to improve clinical nursing of
elders. Annual Review of Nursing Research 2002;20:293-322
40. Kairy D, Lehoux P, Vincent C, Visintin M. A systematic review of clinical
outcomes, clinical process, healthcare utilization and costs associated with
telerehabilitation (Structured abstract). Disability and Rehabilitation,
2009:427-447.
41. Leach L, Christensen H. A systematic review of telephone-based
interventions for mental disorders. Journal of Telemedicine and Telecare
2006;12:122-129
42. Lehmann S, Domdey A, Bramesfeld A. [Telephone case management: is it
beneficial for the care of depression patients in Germany? A systematic
literature survey]. Gesundheitswesen;72(5):e33-7
43. Lustria MLA, Cortese J, Noar SM, Glueckauf RL. Computer-tailored health
interventions delivered over the Web: review and analysis of key components.
Patient Education & Counseling 2009;74(2):156-73
44. Martin S, Kelly G, Kernohan W, McCreight B, Nugent C. Smart home
technologies for health and social care support. cochrane Database of
Systematic
Reviews
2008(4):Art.
No.
CD006412.
DOI:
10.1002/14651858.CD006412.pub2
45. McLean S, Nurmatov U, Chandler D, Liu J, Pagliari C, Sheikh A.
Telehealthcare for Asthma. cochrane Database of Systematic Reviews
2009(2):Art. No.: CD007717.DOI:10.1002/14651858.CD007717
46. Mistiaen P, Poot E. Telephone follow-up, initiated by a hospital-based health
professional, for postdischarge problems in patients discharged from hospital
to home. Cochrane Database Syst Rev 2006(4):CD004510
47. Montori VM, Helgemoe PK, Guyatt GH, Dean DS, Leung TW, Smith SA, et al.
Telecare for patients with type 1 diabetes and inadequate glycemic control: a
randomized controlled trial and meta-analysis (Provisional abstract). Diabetes
Care, 2004:1088-1094.
48. Murray E, Burns J, See TS, Lai R, Nazareth I. Interactive Health
Communication Applications for people with chronic disease. Cochrane
Database of Systematic Reviews 2005(4):CD004274
49. Neubeck L, Redfern J, Fernandez R, Briffa T, Bauman A, Freedman S.
Telehealth interventions for the secondary prevention of coronary heart
disease: a systematic review. European Journal of Cardiovascular Prevention
and Rehabilitation 2009;16(3):281-289
50. Polisena J, Tran K, Cimon K, Hutton B, McGill S, Palmer K. Home telehealth
for diabetes management: a systematic review and meta analysis. Diabetes
Obes Metab 2009;11(10):913-30
51. Polisena J, Tran K, Cimon K, Hutton B, McGill S, Palmer K, et al. Home
telemonitoring for congestive heart failure a systematic review and metaanalysis. Journal of Telemedicine and Telecare 2010;16:68-76
52. Revere D, Dunbar PJ. Review of computer-generated outpatient health
behavior interventions: clinical encounters "in absentia". Journal of the
American Medical Informatics Association 2001;8(1):62-79
53. Roine R, Ohinmaa A, Hailey D. Assessing telemedicine: a systematic review
of the literature. CMAJ Canadian Medical Association Journal
2001;165(6):765-71
54. Samoocha D, Bruinvels DJ, Elbers NA, Anema JR, van der Beek AJ.
Effectiveness of web-based interventions on patient empowerment: a
368
systematic review and meta-analysis. Journal of Medical Internet
Research;12(2):e23
55. Stinson J, Wilson R, Gill N, Yamada J, Holt J. A Systematic Review of
Internet-based Self Management Intervetions for Youth with Health
Conditions. Journal of Pediatric Psychology 2008;34(5):495-510
56. Whitten P, Mair F, Haycox A, May C, Williams T, Hellmich S. Systematic
review of cost effectiveness studies of telemedicine interventions. BMJ
2002;324:1434-7
57. Pare G, Moqadem K, Pineau G, St-Hilaire C. Clinical effects of home
telemonitoring in the context of diabetes, asthma, heart failure and
hypertension: a systematic review. Journal of Medical Internet
Research;12(2):e21
58. Catwell L, Sheikh A. Evaluating eHealth Interventions: The Need for
Continuous Systemic Evaluation. PLoS Medicine 2009;6(8):e1000126
59. Pinnock H, Bawden R, Proctor S, Wolfe S, Scullion J, Price D, et al.
Accessibility, acceptability, and effectiveness in primary care of routine
telephone review of asthma: pragmatic, randomised controlled trial. BMJ
2003;326(7387):477-9
60. Pinnock H, McKenzie L, Price D, Sheikh A. Cost-effectiveness of telephone or
surgery asthma reviews: economic analysis of a randomised controlled trial.
Br J Gen Pract 2005;55(511):119-24
61. Mail. Data blunder. Mail Online 2009:http://www.dailymail.co.uk/news/article495188/Now-MORE-discs-containing-personal-data-missing-bunglingMinistry-Mayhem.html
62. BBC. UK families put on fraud alert.http://news.bbc.co.uk/1/hi/7103566.stm
63. McKinstry B, Watson P, Pinnock H, Heaney D, Sheikh A. Confidentiality and
the telephone in family practice: a qualitative study of the views of patients,
clinicians and administrative staff. Family Practice 2009;26:344-350
64. Collmann J, Cooper T. Breaching the Security of the Kaiser Permanente
Internet Patient Portal: The Organizational Foundations of Information
Security. Journal of the American Medical Informatics Association
2007;14(2):239-243
65. Mort M, Finch T, May C. Making and Unmaking Telepatients: Identity and
Governance in New Health Technologies. Science, Technology and Human
Values 2008;34(9):8-33
66. EU. Opinion of the European Economic and Social Committee on the
"Communication from the Commission to the European Parliament, the
Council, the European Economic and Social Committee and the Committee of
the Regions on telemedicine for the benefit of patients, healthcare systems
and society". Official Journal of the European Union 2009;COM (2008) 689
final:C 317/84
67. Cuzner E. Medico-legal dilemmas posed by advances in computer
technology. Good Practice 2010;1(1):4-6
68. GMC.
Guidance
on
remote
prescribing.
In:
http://www.gmcuk.org/guidance/ethical_guidance/prescriptions_faqs.asp,
editor.
Good
Medical Practice: General Medical Council, 2006:Paras 39-43.
69. DoH. Liberating the NHS: An Information Revolution: Department of Health,
2010.
70. Pinnock H, Adlem L, Gaskin S, Harris J, Snellgrove C, Sheikh A.
Accessibility, clinical effectiveness, and practice costs of providing a
telephone option for routine asthma reviews: phase IV controlled
implementation study. Br J Gen Pract 2007;57(542):714-22
71. McKinstry B, Hammersley V, Burton C, Pinnock H, Elton R, Dowell J, et al.
The quality, safety and content of telephone and face-to-face consultations: a
comparative study. Qual Saf Health Care 2010;19:298-303
369
72. McKinstry B, Walker J, Campbell C, Heaney D, Wyke S. Telephone
consultations to manage requests for same-day appointments: a randomised
controlled trial in two practices. Br J Gen Pract 2002;477:306-10
73. McKinstry B, Watson P, Pinnock H, Heaney D, Sheikh A. Telephone
consulting in primary care: a triangulated qualitative study of patients and
providers. British Journal of General Practice 2009;June:e209-e218
74. Nicolini D. Stretching out and expanding work practices in time and space:
the case of telemedicine. Human Relations 2007;60:689
75. Cruickshank J, Beer G, Winpenny E, Manning J. Healthcare without walls, A
framework for delivering telehealth at scale: 2020health.org, 2010.
76. DoH. Whole system demonstrator project., 2008.
77. Chaudhry SI, Mattera JA, Curtis J, Spertus Ja, Herrin J, Lin Z, et al.
Telemonitoring in patients with heart failure. N Eng J Med 2010;363(24):23019
78. Krumholz H, Amatruda J, Smilth G, etal. Randomized trial of an education
and support intervenion to prevent readmission of patients with heart failure. J
Am Coll Cardiol 2002;39:83-9
370
Chapter 14
Case study: Telehealthcare for asthma
Summary
•
Asthma is now one of the commonest long-term conditions in the UK and
other economically-developed countries.
•
A range of telehealthcare interventions have been developed and trialled,
but the effectiveness and cost-effectiveness of these remain unclear.
•
We conducted a Cochrane systematic review, which identified 21 eligible
randomised controlled trials.
•
These trials included study of the telephone, video-conferencing, the
Internet and other networked communications; and text messaging.
•
Meta-analysis showed that these interventions did not result in clinically
important improvements in asthma quality of life (minimum clinically
important difference=0.5): mean difference in Juniper’s Asthma Quality of
Life Questionnaire (AQLQ) 0.08 (95%CI 0.01, 0.16).
•
Telehealthcare resulted in a non-significant increase in the odds of
emergency department visits over a 12-month period: OR 1.16 (95%CI
0.52, 2.58).
•
Meta-analysis however showed a significant reduction in hospitalisations
over a 12-month period: OR 0.21 (95%CI 0.07, 0.61), the effect being
most marked in people with more severe asthma managed predominantly
in secondary care settings.
•
There were only limited data on cost-effectiveness of telehealthcare
interventions.
•
Overall, it seems that telehealthcare interventions are unlikely to result in
clinically relevant improvements in health outcomes in those with relatively
mild asthma, but they may have a role in those with more severe disease
who are at high risk of hospital admission.
•
Further trials evaluating the effectiveness and cost-effectiveness of a
range of telehealthcare interventions are needed.
371
14.1 Introduction
There is no gold standard objective definition of asthma; its diagnosis is
clinical, based on the presence of characteristic symptoms (wheeze,
breathlessness, chest tightness and nocturnal or exercise induced cough) and
of variable airflow obstruction (BTS/SIGN 2008)1. The features of asthma are
so heterogeneous that, in both children and adults, it seems that what is
currently termed ’asthma’ is unlikely in the future to be regarded as a single
disease entity2. Much research is still needed to answer the following three
fundamental questions:
1. What is asthma?
2. Who gets asthma and why?
3. Which factors predict exacerbations and treatment response?3
The Global Initiative for Asthma (GINA),4 run in collaboration with the World
Health Organization and the U.S. National Heart Lung and Blood Institute
(NHLBI) and National Institute for Health (NIH), estimates that 300 million
people have asthma.4 Asthma is thus now a very common long-term condition
and there has been an increase in prevalence in recent decades.5-7 The
highest prevalence rates, as high as 30%, are amongst certain age groups in
economically developed English-speaking countries.8-10 However, there has
also been an increase in asthma prevalence in many economically developing
countries11
12
ISAAC 1998;13 This increase affects both children and adults.
Worldwide asthma presents substantial challenges. The high disease burden
demands improvements in the development of and access to treatments.10 14
15
Patterns of help-seeking behaviour are also relevant, as delayed reporting
is associated with greater morbidity and the need for costly emergency care.
There is also a significant indirect cost burden associated with asthma
through school and work absences.
We summarise in this chapter the main findings from our recently completed
systematic review; the methods are detailed in our Cochrane review.
372
Telehealthcare interventions may help to address some of the above
challenges by enabling remote delivery of patient-centred care, facilitating
timely access to health advice and medications, prompting self-monitoring
and medication compliance, and educating patients on trigger avoidance.16-19
Telehealthcare is a complex intervention and, as such, it is quite difficult to
specify exactly why it works or does not work, i.e. what is/are the ’active
ingredient(s)’ within the intervention.20 Some potential mechanisms through
which the use of telehealthcare may enhance the quality of care and achieve
cost savings include (adapted from Finkelstein 200021):
•
Providing patient education and counselling for primary prevention and
early detection of disease;
•
Replacing face-to-face nursing/doctor visits;
•
Improving adherence to medications and other treatment regimens;
•
Monitoring patients’ health parameters remotely;
•
Enabling early detection of incipient disease exacerbation and timely
intervention for early symptom management;
•
Reducing unscheduled/unnecessary visits to the physician and
emergency room;
•
Preventing repeat hospitalisations.
There are now in many parts of the world an increasing array of electronic
tools for remotely helping people with asthma and often many are now
beginning to be implemented in the absence of an explicit evidence base.19 22
A recent Cochrane Review of generic tele-consultations compared with faceto-face consultations found little evidence of clinical benefit.23 There were also
a lack of analysable data for assessing the cost-effectiveness of these
interventions. The authors concluded that further research is required.23
Another
systematic
review24
of
studies
of
patient
satisfaction
with
telehealthcare raised a number of important questions, these included:
•
What types of consultation are suitable for remote consulting?
373
•
What are the effects of this mode of healthcare delivery on the
clinician-patient relationship?
•
How do communication issues affect the delivery of healthcare via
telehealthcare?
•
What are the possible limitations of telehealthcare in clinical practice?
Answers to such questions are urgently needed or we risk blindly
implementing a non-proven way of working which may have a negative effect
on
patients
and
professionals.
One
commonly
used
argument
for
telehealthcare is that long-term running costs will be lower than in
conventional care because disease will be detected and treated early,
preventing ensuing morbidity and hospitalisations and allowing patients to be
cared for in their own home. However, the initial start-up costs of
telehealthcare may be substantial.25 The cost-effectiveness of telehealthcare
interventions therefore also needs to be established. In asthma, patients often
have a high level of responsibility for their own health. In some people it can
also be a life-limiting and challenging disease to manage. Telehealthcare
interventions may make this easier for patients by providing timely
professionally guided feedback on their condition. Such interventions may
help patients to identify and address triggers and to optimise their medication
regimens to address the fluctuations in their illness - and at low cost. This is
the ideal situation, however such results cannot be presumed and a robust
critique of the evidence base is overdue.
14.3 Empirically demonstrated main findings
14.3.1 Description of studies
Our searches found 525 titles and abstracts and following review we
considered 76 to be relevant. After detailed examination of the 76 full texts, 21
trials satisfied our inclusion criteria. In addition, we found 14 ongoing trials that
have reported only as abstracts and a further 21 trials that have yet to report
in any format. Two studies had to be translated from Japanese and one from
Italian. It was only possible to obtain partial translations of the Japanese
374
reports and so the information is taken from the studies’ figures which were
published largely in English.
Two studies used pharmacists as the main deliverer of the telehealthcare
intervention26
27
and the rest used a combination of doctors (both general
practitioners and specialists) and nurses, including specialist nurses. The
most common model for intervention was to have an initial face-to-face
introductory session and then follow-up using telephone, telephone and web,
web/other networked system or text message. This approach featured in the
following studies: 27 28 29 30 31 32 33. Pinnock34 and Donald29 published follow-up
papers dealing with the costs and cost-effectiveness of the interventions (35
and
36
). Willems37
evaluation,
38
38
and
33
refer to only one trial of which
publishes cost-effectiveness data and
Kokubu39 was expanded on in
40
33
37
is a process
is the main report.
with the addition of more data and a section
on costs; however, this was hard to interpret given the incomplete
translations.
In terms of the major telecommunication devices used in the studies, overall
nine studies used the telephone (27 41 28 29 42 43 34 44 and
video26
46
. In
26
45
). Two studies used
video-conferencing was used to deliver education on inhaler
technique and in
46
participants submitted repeated videos for checks of their
inhaler technique via modem. One study used text messaging32. Two studies
used the Internet47 48. Other networked systems were used by six trials;49 30 31
39 40 33
. Van der Meer50 used text or internet.
There were a number of reasons for excluding studies. Most often studies
were excluded because they did not fulfil our definition of telehealthcare, i.e.
there was not a two-way exchange of information between patient and
healthcare professional. If the intervention involved only education without
feedback or if feedback was only mechanical in nature, e.g. from a peak flow
meter and not involving a professional, then the study was excluded. In
addition, we excluded studies if they were found not to employ a randomised
controlled design or if they were not studying an asthma population.
375
14.3.2 Risk of bias in included studies
Allocation
Fifteen trials used appropriate randomisation27
26 46 28 47 49 42 31 43 40 32 34 44 50
33
. For the remaining studies the methods of allocation were unclear
41 29 30 39
45 48
.The most common method of randomisation was to use a random
number table, often computer-generated; however, other acceptable methods
were also used, including tossing a coin. In most studies which included
details of concealment, sealed envelopes were used. However, some studies
appeared to use a centralised randomisation hub, but this was often unclear.
Blinding
Four studies
27 28 29 43
made some attempt to blind researchers as to the
group allocation of their participants. The remainder did not and this may have
introduced bias. Guendelman30 used self-reporting of outcomes to the nurse
co-ordinator so that the same person was both involved with delivering the
intervention and assessing outcomes, thus substantially increasing the risk of
bias. In Pinnock34 where blinding was not feasible due to the pragmatic nature
of the trial, an independent researcher validated a 20% sample of the results.
Incomplete outcome data
Several studies had high drop-out rates
26 29 42 31 43
. In Pinnock
44
the uptake
rate and patient population dynamics were being studied as a primary
outcome measure because it was a pragmatic phase IV implementation trial
(as per Medical Research Council 2008).20
Selective reporting
Research protocols were not sought for any of the studies. There was
nonetheless some evidence of selective reporting of results. In most studies,
all outcomes specified in the methods section were reported in the results,
however there were some exceptions, for example:
31
the data on satisfaction
were not reported and 47 did not report quality of life data. These last points do
not seem to relate to selective reporting, but rather other problems.
Other potential sources of bias
376
Variable efforts to recruit from ethnically diverse and marginalised populations
may have impacted on the external validity of the findings. There was variable
consideration of smoking. Patients with paper diaries filled in more than one
day’s entry at a single time point thereby opening the data to recall bias.
Some studies recruited from academic centres rather than from primary care
which may limit the generalisability of findings.
14.3.3 Effectiveness of interventions
Primary outcomes – clinical: Asthma quality of life
The impact of telehealthcare interventions on disease-specific quality of life
was assessed in 14 trials46
28 49 29 42 31 43 40 34 44 48 50 45 33
. The effect of
intervention is shown in the forest plot Figure 14.1.
Figure 14.1 Asthma quality of life questionnaires AQLQ Juniper mean
scores
Five of these studies
28 42 34 44 45
used Juniper’s validated Mini-AQLQ.51 This
instrument contains 15 items which are scored from 7 (no impairment) to 1
(maximum impairment), so high scores indicate better quality of life. Three
studies
48 50 33
used Juniper’s validated full 32-item Adult-AQLQ in which
similarly high scores indicate better quality of life. Two studies used Juniper’s
validated 23-item Paediatric Asthma Quality of Life Questionnaire (PAQLQ)
46
33
, high scores again indicate better quality of life. Jan31 used the Paediatric
Asthma Caregiver’s Quality of Life Questionnaire (PACQLQ), which is filled in
by the patients’ parents as did
49
and
43
. There are 13 items in this instrument
and, in keeping with the other instruments, higher scores represent better
377
quality of life and less impairment by asthma52
51
. As Juniper’s quality of life
instruments are similarly structured with each question answered on a Likert
scale with a minimum value of one and a maximum value of seven, we
considered it appropriate to perform meta-analysis of the data derived from
these instruments. We performed a meta-analysis of
46 28 49 42 53 44 50 45 33
.
This meta-analysis of nine studies, yielding a total of 1566 intervention and
1553 control patients, revealed a mean difference of 0.08 point improvement
on this scale (95%CI 0.01,
0.16) in those randomised to intervention
compared with controls. This is lower than the minimal clinically important
difference of 0.5 points on the Juniper scale. It was not possible to include
other studies in the meta-analysis because:
•
48
had used the AQLQ, but could not be included because the data
were not normally distributed and so median scores were supplied by
the author on request. AQLQ score was 6.42 (IQR 3.62 to 7.00) in the
Internet group, 6.31 (IQR 3.98 to 7.00) in the GP group and 6.17 (IQR
1.41 to 7.00) in the specialist group.
•
35
used the Modified Marks Asthma Quality of Life Questionnaire, a
validated scale in which a higher score indicated a less detrimental
impact on quality of life. A clinically important difference was seen in
the intervention group from recruitment to 12 months. There was no
clinically important difference seen in the control group in this time.
•
40
did not use a validated instrument for the measurement of quality of
life. However, they showed a greater improvement in the intervention
group than in the control group (P=0.04). These results, however, need
to be interpreted with caution because of the unvalidated nature of the
scale and, furthermore, the inability to determine what constituted a
minimal clinically important difference.
•
31
used the Juniper’s PACQLQ, however insufficient summary statistics
were reported for meta-analysis. Only the caregivers of asthmatic
children randomised to the intervention group showed significant
improvement after the study compared to before the study.
•
43
used the Juniper’s validated PACQLQ. Small increases in median
score after the study were not statistically significant for either the
378
control (P=0.11) or the intervention group (P=0.6). These data were not
normally distributed and so the mean was not used for comparison.
Overall, across all these quality of life studies, these results suggested that
telehealthcare did not result in clinically important improvements in quality of
life. The mean difference (fixed effect) was 0.08 (95%CI 0.001, 0.16).
Emergency department visits
Ten studies reported data on emergency department visits
46 28 29 30 43 40 34 48
45 33
. The effect of telehealthcare interventions on emergency department
visits over 12 months is shown in Figure 14.2. This meta-analysis included
five trials
46 29 40 48 33
representing 619 patients in total. It revealed a non-
significant increase in the odds of emergency department attendance: OR
1.16 (95%CI 0.52, 2.58).
Figure 14.2 One or more emergency dept visit in 12 months
In
40
the authors reported a greater reduction of night and daytime emergency
room visits per patient in the control group than in the intervention group. It
appears from their data that all patients in both arms were admitted to the
emergency department at least once at some point during the study; it was
therefore not possible to include these data in the meta-analysis. The
absolute numbers of visits (presumably more often than one per patient) are
not given in the study’s English tables. Of the other studies not included in the
meta-analysis:
379
•
43
reported emergency department visits over a six-month interval and
so could not be included in the meta-analysis, but again numbers were
very small with only one or two patients attending from each arm over
the study period.
•
28
•
There were no emergency department consultations in 34.
•
45
reported only within group analyses.
did not distinguish between emergency department care and full
hospitalisation, therefore it was not possible to include these results in
the meta-analysis.
•
49
reported survival analysis by means of a Kaplan-Meier curve of time
to first prednisolone course, emergency visit or hospitalisation or to
whichever came first. There were a total of 31 events in these
categories, but the detailed breakdown of data was not reported and so
the data could not be included in meta-analysis. However, the time to
the first emergency department visit was considered comparable
across the two arms of the study (P=0.13).
•
The remaining trials27
26 41 47 35 42 39 31 32 44 50
did not include data on
emergency department visits.
Hospitalisations
Six studies 46 35 30 40 32 48 presented data on hospitalisations. For two studies30
32
these hospitalisations occurred over a three-month period. For the
remaining four studies the hospitalisations were recorded as occurring over a
12-month period46
35 40 48
. Meta-analysis of the two studies30
32
that reported
data from hospitalisations over three months of study duration is shown in the
forest plot in Figure 14.3. This includes data from 138 patients. Overall, there
was no significant difference in the odds of hospitalisation in the intervention
groups when compared to the control group (OR 0.47, 95%CI 0.010, 36.46).
The confidence intervals were very wide.
380
Figure 14.3 Hospitalised once or more in 3 months of study
Meta-analysis of the four studies
46 35 40 48
which reported on hospitalisation
within 12 months of randomisation is shown in Figure 14.4. This includes data
from 499 patients. This gave an OR of 0.21 (95%CI 0.07, 0.61) indicating that
telehealthcare reduces the risk of hospitalisation.
46 35
and
48
, cross the line of
no difference. These studies are large and of reasonable quality. In contrast,
Kokubu40, which contributes a weight of 55.6%, shows a clearly beneficial
effect. This trial focused on patients with severe asthma. Patients who had
visited the night emergency department room three times or more within a
year in spite of corticosteroid therapy were selected and so these patients
were not representative of a typical asthma population. In the Kokubu39,40
study, 2/32 intervention patients were hospitalised and 11/34 of the control
patients
were
hospitalised.
Overall,
it
therefore
seems
as
though
telehealthcare interventions may prevent hospitalisations in selected high-risk
populations studied over long timescales.
Figure 14.4 Hospitalised once or more over 12 months of study
There were no hospitalisations during the period of follow-up in
34
Numbers
emergency
of
hospitalisations
department visits in
45
. But
41
were
indistinguishable
from
or in
43
.
reported hospitalisations over the three-month
period of its trial, and it was unclear as to whether only data for the control
381
group were presented and so the study was not included in the meta-analysis.
One study49 reported survival analysis (Kaplan-Meier curves) of time to first
hospitalisation for both groups with P=0.13 for intention-to-treat analysis.
From the curve, approximately 5% of the intervention group were hospitalised
once or more and 15% of the control group, however as these figures are
rough visual estimates they were not used in the meta-analysis. Another42 did
not clearly report on the number of hospitalisations, referring instead to mean
length of inpatient stay. Also 47 31 and 44 did not report on hospitalisations.
Overall, hospitalisation was an infrequent outcome. In the forest plot for
hospitalisations in studies over a three-month period it can be seen that
telehealthcare is not associated with a reduction in hospital admissions.
However, there is a reduction over 12 months. This may point to a role for
telehealthcare to reduce hospitalisation in high-risk individuals.
Other primary outcomes
Symptoms
The following 17 studies reported on symptoms as an outcome measure:
41 28 49 42 30 31 43 39 32 34 44 48 50 45 33
27 46
In Barbanel27 symptom scores improved in
the intervention group when compared to the control group over the three
months of the study; the difference between groups at three months adjusted
for baseline scores was 7.0 (95%CI 4.4, 9.5; P<0.001), on a scale of 10 to 40.
This difference remained significant when adjustments were made for missing
data. Another study46 reported the number of symptom-free days recorded by
each group, but the difference between groups was not significant (P=0.13,
our calculation).In Clark28 study, there was a small drop in the average
number of nights with night-time symptoms per month experienced by the
women following the intervention (from 5.1 to 3.7, i.e. -1.4 nights). However,
across groups there was no difference in average number of nights with nighttime symptoms per month: control group 3.8 nights, intervention group 3.7
nights. Wheezing-related sleep disturbances were studied in
49
not use a validated questionnaire. Study
41
, but they did
calculated the within-group
percentage of symptom-free days. This was found to improve significantly
with P<0.001 in both groups. The authors speculate that this may have been
382
due to frequent monitoring and telephone contacts in both arms, which could
not be further improved by adding the nitric oxide telemonitoring. There was
no difference across groups, the baseline-adjusted difference in mean
percentage of symptom-free days was 0.3%, (SD -10% to 11%, P=0.95).
30
reported asthma control problems between groups as not significantly
different at 12 weeks (P=0.07).
Study43 reported that there was no significant difference between groups in
their primary outcome of wheezing in the last three months and study42
reported the mean of individual changes in the Asthma Control Questionnaire
(ACQ), a validated questionnaire, over 12 months. The clinic group changed
by -0.11 (-0.32 to 0.11) and the telephone group by -0.18 (0.38 to 0.02), this
representing a non-significant improvement in asthma control (P=0.35 when
adjusted for baseline differences) (a negative change in ACQ is an
improvement).
Between group differences were significant in Jan 31 for the Paediatric Asthma
Control Test scores change from baseline at 12 weeks: the intervention group
had a significant decrease of night-time (P=0.028) and daytime (P=0.009)
symptoms compared with the children in the control group. There were no
between group comparisons in this study.
Kokubu40 method for analysing the symptom score remains obscured by the
lack of a full translation of the Japanese paper.
Ostojic32 used a bespoke symptom score which produced significant results
across scores for the study group and control group demonstrating that the
control group had more symptoms during the study. Scores for cough were
1.42 (SD=0.28) for the study group and 1.85 (SD=0.43), P=0.028, and scores
for sleep quality were: study group 0.85 (SD=0.32) and control group 1.22
(SD=0.23), P=0.021. However, it is not reported whether this symptom
scoring system had been validated.
383
Pinnock34 measured symptom scores using the ShortQ asthma morbidity
score three months after randomisation and found these to be similar in the
two groups. Face-to-face consultations had a mean score of 1.96 (SD=1.96)
and telephone consultations had a mean score of 2.10 (SD=2.16), difference
0.41 (95%CI -0.41, 0.68) (P=0.62), i.e. a non-significant difference.
In 2007, study44 used the validated Asthma Control Questionnaire and found
a non-significant mean difference of 0.12 (95%CI -0.06, 0.31) (P=0.19) for the
telephone option versus face-to-face only.
Rasmussen48 reported an improvement in symptoms in the Internet versus
specialist groups: OR 2.64 (95%CI 1.43, 4.88, P=0.002); and also in the
Internet versus GP groups: OR 3.26 (95%CI 1.71, 6.19, P<0.001).Here the
Internet group showed the greatest likelihood of improvement over six
months. Another study50 used the validated ACQ to compare groups. The
Internet group showed greater improvement of asthma control than did the
usual care group: change from baseline -0.54 versus -0.06; adjusted
difference -0.47 (CI -0.64 to -0.3) which was just slightly smaller than a
clinically relevant difference of -0.5 (where negative changes represent
improvements).
Vollmer45 assessed several markers of symptom status including the Asthma
Therapy Assessment Questionnaire (ATAQ), asthma impact score, selfreported health status, self-assessed severity of asthma in the past four
weeks and nocturnal waking in the past four weeks. No significant difference
between the telephone and usual care groups were found for any of these
outcomes (ATAQ P=0.56, asthma impact score P=0.46, self-reported health
status P=0.08, self-assessed severity of asthma in the past four weeks P=
0.22 and nocturnal waking in past four weeks P=0.84).
Willems33 recorded the changes in asthma symptoms of coughing, production
of sputum and shortness of breath/wheezing, however no statistically
significant differences in improvement in any of the symptoms were observed.
The above results suggest that symptom scores may be improved with
384
telehealthcare. However, in many cases there was no difference between
groups. Future research should use validated scoring systems which can then
be pooled in meta-analysis to give a more accurate picture of the extent to
which telehealthcare interventions improve symptom scores.
Improved access
Improved access was clearly demonstrated in Pinnock34. Of patients
randomised to receive the telephone review, 74% of patients were reviewed,
whereas only 48% of patients in the surgery-only group were reviewed. There
was therefore a significant improvement of 26% (95%CI 14, 37; P<0.001).
Asthma-related morbidity at three months (in terms of acute exacerbations of
asthma) P=0.68; emergency broncho-dilation P=0.97; and steroid courses for
asthma P=0.64) were not significantly different across groups, and neither
was quality of life (Juniper scores P=0.69). And so in this study access was
improved with no loss in quality and at no additional cost.
Access in
44
was also improved: the proportion reviewed was 66.4% of
patients in the telephone option group compared with 53.8% in the face-toface only group. Chan46 treated the control group with an ambulatory asthma
clinical pathway with six visits scheduled over the period of 12 months after
the initial visit. The intervention group had three in-person visits at 0, 26 and
52 weeks and the rest as virtual visits using video-conferencing technology. In
Gruffydd-Jones’ study, 35% more asthma patients received their annual
review in the telephone group than in the clinic group and the cost of the
telephone group was less.
In Vollmer45 there was evidence of a negative reaction to automated
computerised calling and the intervention was more successfully delivered in
the live caller arm (P<0.001).
Secondary outcomes
Study withdrawal
Barbanel 27 lost one patient when (s)he moved away.
385
Bynum26: 49 students were randomised. Three did not attend any visit, eight
had never used a metered dose inhaler (MDI) before and so did not meet the
inclusion criteria (the study does not explain why they were not screened out
before randomisation) and two students did not attend the follow-up visit.
Chan46: 127 children were assessed for eligibility: one did not meet the
inclusion criteria, five refused to participate and one was not able to
participate. The study report does not give further details of why not. One
hundred and twenty children were randomised, 11 were lost to follow-up,
seven discontinued because they moved, and so 55 were included in the
analysis for the control group and 47 remained for analysis in the intervention
group.
Chatkin41: 293 patients were screened for participation, four were excluded for
not fulfilling the inclusion criteria, eight for not responding to the telephone
calls and 10 for not returning their inhaler disk to the study office. In the final
analysis control group n = 131 and intervention group n = 140.
Clark28: 808 women provided baseline data and mailed a consent form back
following the mailing of 2336 invitation letters. Four hundred and twenty-four
were randomised to the intervention group and 384 to the control group.
There was then attrition of 133 participants from the intervention group and 87
participants from the control group. The reasons for this were not reported.
Cruz-Correia47: 21 patients were included from 28 patients assessed in the
outpatient clinic. Patients were only included if they used the Internet. Sixteen
patients provided their opinions on the monitoring instruments. Four patients
did not use the monitoring tools because of technical problems with the
instruments.
de Jongste49: 151 children were randomised and four children were excluded,
two due to non-compliance, one because he or she had been inappropriately
included and did not have allergy and the final child had moved abroad and
was unable to transfer data. All others completed the study.
386
Donald29: 660 patients were assessed for eligibility, 385 could not be
contacted, 154 did not want to take part in the study 31 were excluded and 19
did not attend the first meeting. Following randomisation there were 36 in the
intervention group with 31 left at the end of the study for analysis and 35
controls, of which 29 were left in the final analysis. The study did not discuss
why participants withdrew following randomisation.
Gruffydd-Jones42: 97 patients were randomised to the control group, only 82
attended their first visit and 62 completed the study, of which 28 completed all
visits and 34 completed two visits only. Ninety-seven patients were similarly
randomised to the telephone group, 90 attended their baseline visit, three
patients left the area, two patients died and 84 completed all three visits. An
explanation for the final patient was missing from the patient flow diagram.
Guendelman30: Reasons for withdrawing from the study included moving out
of the area (n = 3) or life crisis (n = 4). Five families who dropped out were
unavailable for contact by the study authors to find out their reasons.
Jan31: 184 families were eligible with access to the Internet. Five families
declined to participate as they were “too busy”, “not interested” or found it “too
complex to perform the diary card”. Fifteen participants were excluded
because of either a request from the parents or a lack of data due to Internet
failure.
Khan43: 310 children were enrolled in this study, 266 completed the follow-up
questionnaires. The reasons for non-response were not explored although in
two cases a search of the telephone directory enabled the discovery of a new
address.
Kokubu39: Two patients withdrew from the intervention group and one from
the control group, however the reasons for this are not described.
387
Kokubu40: Five patients withdrew from the intervention group and four from
the control group but the reasons for this are not described in the report
according to the translation we have available.
Ostojic32: Patients unfamiliar with text SMS or without consistent access to a
cell telephone were excluded. After enrolment no patient withdrew from the
study.
Pinnock44: This was a phase IV implementation study and so there is no
selection of participants. It instead takes place within the real-world situation
of general practice where patients move, die or their asthma becomes more
active or becomes inactive or is re-diagnosed as COPD or patients refuse to
take part.
Rasmussen48: 300 subjects were enrolled who fulfilled the criteria for asthma,
253 subjects completed the trial. Reasons for the withdrawals were not given.
Van der Meer50: recruitment was undertaken from 37 general practices and
continued until there were 200 patients entered in the study. Results were
available for 183 patients. Twenty patients did not complete the three-month
questionnaire, eight patients were lost to follow-up and nine patients withdrew
consent.
Vollmer45: It is difficult to assess withdrawal in this study as it was devised to
compare usual care with automated telephone outreach and the outcome
measures were taken from a representative sample. However, participation
statistics were low for the intervention arm, only 38% participated in the first
round of calls, 32% in the second round and 18% in the third round. Overall
47% of the intervention group had one call and 12.1% completed all three
calls. Compared with those who did not participate in any call, participants
were more likely to be female, take more inhaled corticosteroids and report
worse asthma-specific quality of life at baseline. Interestingly 59.9% of the
live-caller arm completed at least one call and 27.6% all three calls,
suggesting that patients preferred a live call to the automated calls.
388
Willems33: 274 patients were assessed as potentially eligible from patient
records. Eighteen did not have a house phone connection and so were
excluded as not meeting inclusion criteria. One hundred and forty-seven
refused to participate because of the following reasons: no time (n=63), not
interested (n=28), no symptoms (n=18), too confronting (n=13) and other
reasons (n=25). Therefore, 109 were enrolled, seven were lost to follow-up,
six refused participation continuation and one person emigrated.
Time off school or work
Clark28: reported the yearly average number of days missed work per month
as 3.0 (SD=7.1) for the control group and 2.3 (SD=6.2) for the intervention
group (P=0.14).
In the Donald29 trial, 24 intervention group participants worked or studied and
lost 10 days over 12 months, and 25 of the control group participants lost a
total of 11 days; the difference between the groups was not significant
(P=0.58).
In Guendelman,30 the odds of missing school in the past six weeks in the
Health Buddy group were 0.74 (95%CI 0.37, 1.5). In summary, it appears that
these telehealthcare interventions did not reduce time off work or school.
PEF monitoring and diary monitoring
46 31 33
and32 report PEF flow monitoring. The initial improvements in inhaler
technique seen in both arms of
46
can be attributed to the monitoring of
participants by healthcare professionals. As monitoring by the health
professionals dropped in the second 26 weeks of the study to a telephone call
once a week in the control group, the participants’ inhaler technique scores
dropped off. Mean peak flow (+/-SD) as a percentage of personal best was
reported as 91.6% (+/- 27.2) in the intervention group and 100% (+/-17.6) in
the control group at the end of the study.
389
Monitoring, or more accurately “prompting”, was also important in 41 a study in
Brazil. A twice-weekly telephone call made by a trained student nurse to the
patients in the experimental arm resulted in an inhaler adherence rate of 74.3
% whereas the rate in the control arm was only 51.9% (number needed to
treat to benefit 4.5, i.e. telephone calls to 4.5 patients were required to prevent
one non-adherence or missed inhaler dose). The content of the telephone call
was individualised to each patient. In keeping with this advantage of
monitoring, Kokubu39 also found that those patients who checked their peak
flow most regularly tended to have the best peak flows. Patients who did not
check or transmit peak flow data regularly remained poorly controlled.
In Jan31 at week 12, children in both groups had a significantly increased PEF.
Between-group differences are reported as non-significant.
Willems33 reported rank correlations of lung function values (PEF and FEV1)
with morning symptoms and evening symptoms as moderate to high.
Absolute values were not reported in the study.
Peak flow in L/min was measured by Ostojic32 as mean PEF by time of day
(morning, afternoon, evening) none of which were significant. However, PEF
variability (%) in the text message intervention group (16.12 +/- 6.93) and the
control group (27.24 +/- 10.01) showed a significant difference (P=0.049).
Therefore, overall it can be seen that telehealthcare improved PEF in some
studies, but that this was not a consistent finding. However, monitoring in itself
appeared to generate an advantage.
Spirometry FEV1/FVC
46 49 32 48
and
50
reported data on FVC and FEV1 as follows. There were no
differences between groups in FVC, forced expiratory volume in 1 second or
forced expiratory flow in mid-expiratory phase at the end of the Chan46 study.
FEV1 was similar at baseline in both groups in de Jongste49 and there was no
significant difference in change from baseline over the course of the study
between the groups: both groups improved.
390
Neither FVC nor FEV1 as a percentage of predicted was significantly different
across the groups in Ostojic32.
Rasmussen48 reported odds ratios for improved FEV1 > 300 ml over baseline
of 3.26 (1.50 to 7.11) for Internet versus specialist, with a significance of
P=0.002. The odds ratio of Internet versus GP group was also significant at
P<0.001 (OR 4.86, 95%CI 1.97, 11.94). These odds ratios were calculated
after six months of intervention.
Van der Meer50 reported that mean pre-bronchodilator FEV1 changed by 0.24
L for the Internet and -0.01 L for the usual care group thus indicating an
improvement in the telehealthcare arm’s FEV1.
Patient satisfaction
In the Cruz-Correia47 study, questionnaires were used to assess patient
satisfaction and it was found that overall patients preferred the web-based
system for monitoring their asthma to the paper based system.
Gruffydd-Jones42 triaged patients in the intervention group by telephone call
then allocated follow-up appointments accordingly. Of patients in this group
88% expressed a strong preference for care in this form compared to the
previous standard form.
Pinnock36 is an additional report of patients who had completed the RCT
comparing telephone and face-to-face consultations for routine asthma
reviews; they were sent a questionnaire to better understand their preference
for future reviews. Thirty-three percent preferred telephone consultations for
future reviews, 17% preferred surgery and 50% expressed no preference.
Thematic analysis of the free-text responses identified five themes including
convenience of telephone consultations, specific problems with telephone
consultations (e.g. confidentiality when phoning from a public place), the
human dimension of face-to-face encounters, that the mode of consultation
391
might change according to the clinical situation and wider implications such as
using email for attachment of PEF information. In summary, patient
satisfaction appeared to be high, but these newer approaches did not appear
to suit all patients.
In Bynum26 and Pinnock34 there was no significant difference in the
satisfaction scores of each of the arms (Bynum26 P=0.132; and Pinnock34
P=0.51).
Kokubu40 asked the following question:
What did you think about the frequency of telephone consulting?
1. Adequate 87%
2. Too much 7%
3. Too few 3%
4. No response 3%
Insufficient detail was reported by Chan46 to interpret the satisfaction scores
published.
Willems37 found no clinically important differences between a satisfaction
questionnaire administered at four months and again at 12 months. Both arms
of the trial used a PEF monitor and both arms answered the satisfaction
questionnaire. Only 4% of patients experienced “a lot” of symptoms over the
previous months. Forty percent of patients said that they would like to use the
monitor again in the future; 36% said maybe. Ninety-four percent of patients
were either much or very much appreciative that lung function deterioration is
immediately noticed by the nurse. Eighty-four percent said that they felt “not at
all less safe contacting the nurse instead of the doctor”. Eighty-seven percent
found that they had a qualitative improvement by playing an active role in their
asthma management. Sixty-five percent felt safer while using the monitor.
Eighty-four percent said the monitor did not interfere with their daily activities.
With regard to questions on the application of the monitor both children and
adults were highly satisfied with using the monitor, 87% finding it “not difficult
at all” to perform the spirometry test.
392
Costs from the healthcare perspective
This refers to the costs of providing the intervention compared with
differences in outcomes between the intervention and control groups (all
currency conversions were undertaken in February 2010 to US dollars).
In Pinnock36 the cost to the practices of face-to-face asthma reviews was
$17.31 each per consultation achieved as there was a higher default rate than
for telephone reviews. Telephone reviews reached more patients at $11.20
per consultation achieved. Access was therefore improved at lower cost per
consultation, however overall costs were similar because more patients were
interviewed in the telephone group.
In Pinnock44 the cost per patient reviewed was significantly lower in the
telephone option group; $15.63 versus $19.85. This generated a cost saving
per patient reviewed of between $4.02 and $4.41 per patient reviewed (i.e.
per six-monthly review achieved). Data on costs were published in the
Donald35 paper. The overall cost of delivering the intervention from the
healthcare perspective was $1693.91, i.e. the additional six phone calls each
over the study time period for all intervention patients. These calls resulted in
the intervention group having six readmissions overall, as opposed to the
control group’s 20 readmissions to hospital. The total cost of the hospital
readmissions in the control group was $35,878.52. Therefore there was a
significant cost saving solely on the basis of reduced hospital admissions. The
intervention group also showed a clinically significant increase in quality of life
scores over the 12-month follow-up period in comparison to no change in the
control group’s scores. Formal cost-effectiveness analysis looking at Quality
Adjusted Life Years (QALYs) or Disability Adjusted Life Years (DALYs) was
not undertaken. In the Kokubu40 study, it was estimated that a saving of
$7074 per year per patient would be achieved if they were to use the
telehealthcare system rather than conventional treatment, these savings
largely being achieved by a reduction in hospitalisation costs.
393
In Willems38 costs from the societal perspective were calculated by costing
from the healthcare perspective and adding the cost of over-the-counter
medication, family informal care and productivity losses due to time off work.
Cost-effectiveness data were calculated from the costs and the quality of life
data from EQ-5D utility (adults and children) and the SF-6D utility (adults
only). Cost-effectiveness depends on what society is prepared to pay per
QALY gained. Among adults the healthcare costs were a mean of $695.54
higher in the intervention group. After adjustment for baseline differences by
multiple regression an average 0.03 QALY were gained from the intervention.
This equates to $42,520.39 (31,035 Euros) per QALY gained from the societal
perspective. Among children the healthcare costs were $829.56 higher in the
intervention group. On average 0.01 QALY were gained from the intervention,
equating to $80,437.16 (59,071 Euros) per QALY from the societal
perspective. Overall it therefore appears that the studies which analysed costs
found that where hospitalisation was prevented, costs were favourable to
continuing the intervention. However, this did not hold true for all studies.
14.4 Empirically demonstrated risks
In Gruffydd-Jones42 study, two patients in the telephone triage arm died.
Following contact with the author it was confirmed that these were not
asthma-related deaths. Overall, studies did not report adverse events, with
the exception of Rasmussen48 which found that an increase in corticosteroid
dose was likely to result in oral candidiasis or dysphonia. In this study at
follow-up an increase in inhaled corticosteroids was found in all groups, but
significantly more patients in both the Internet and specialist groups were
using inhaled corticosteroids than in the general practice group.
This review included 21 trials measuring the effects of telehealthcare
interventions. Overall, this review indicates that telehealthcare-based
interventions do not have an appreciable impact on disease-specific quality of
life or risk of emergency department attendance for asthma; they may,
however, result in a reduction in the risk of hospitalisation for asthma,
particularly in high-risk populations.
394
This latter finding is consistent with the high hopes that policymakers hold for
telehealthcare in terms of its ability to prevent asthma patients requiring
hospital admission. However, this meta-analysis is highly reliant on the results
from the Kokubu40 and Donald studies. The Kokubu study40 selected a group
of patients who had very poorly controlled asthma, requiring oral steroids at
least three times in the previous 12 months or having had a previous hospital
admission, and so they were much more likely to be admitted to hospital at
baseline than patients in other studies which were based in primary care. In
addition, our knowledge of this study is incomplete due to difficulties in
translating this article. When the data were analysed without the Kokubu
study40 the rate of hospitalisations over 12 months became non-significant
(P=0.1). Similarly, the Donald35 study recruited from a group of adults who
had been admitted to hospital with their asthma at some point previously.
These two studies together suggest that telehealthcare might be most useful
for high-morbidity asthma groups selected from secondary care over longer
intervals (i.e. greater than 12 months). However, overall hospitalisations
represent infrequent events and so generalisation from this limited number of
patient outcomes should only be undertaken cautiously. Symptom scores data
were inconsistent and often reported as within-group changes rather than
across-group changes. In some instances telehealthcare related to improved
symptoms, but mostly the reporting of data was not robust enough to draw
any clear conclusions.
There were few adverse events. There is always a risk in reducing the level of
care from face-to-face to telehealthcare that if patients are inexpertly triaged
they will receive insufficient support for their needs and their safety will be
compromised. We, however, found no evidence of this having happened. This
suggests that authors were aware of this risk and managed it appropriately.
This review did identify a tendency for patients to abandon the technology and
cease self-monitoring when they felt well; for example in Chan46 children’s
adherence to submission of inhaler technique videos decreased over time.
This observed pattern calls into question not only what works for asthma in
telehealthcare and in what contexts, but for whom too. It seems that many of
395
the primary care population with asthma do not see themselves as having a
chronic illness needing constant medication but as having an occasional
acute inconvenience–this is a separate issue that is best explored using
qualitative designs.
Nine of the 21 studies included in this review studied the telephone as an
intervention. However, Tulu54 found that only 6.6% of the telemedicine
projects listed in the Telemedicine Information Exchange survey used the
“Plain Old Telephone System” (sic). It would seem that more modern
technologies such as video-conferencing, Skype® and web 2.0 technologies
are attracting interest but have not yet made it into the asthma literature.
There were several trials which had a very low participation rate for collection
of their follow-up data, e.g. approximately 38% in Vollmer45 and 27% in
Delaronde55.
Studies varied in their recruitment either from hospital outpatients or
emergency departments in which case patients with more severe forms of
asthma were recruited than in primary care where the patients with milder
asthma were recruited. The type of patient in the study had implications for
the findings as we saw with the Kokubu40 and Donald35 studies.
There is a plethora of ongoing research and research that has as yet only
been published in abstract form. This emerging literature includes a number of
studies looking at networked telehealthcare tools. It is worth noting that in the
situations in which such solutions might prove most useful, e.g. remote and
rural settings, there may be anticipated difficulties with broadband linkage
required for networks. Similarly, frail, elderly people who could benefit from
telehealthcare to help to maintain their independence may lack the cognitive
skills to be able to adapt to network interfaces and so have difficulties using
the technology at all.
The research in this review has mostly taken place in a developed world
setting with the equivalent of primary and secondary care, sometimes
transferring the responsibility for care from secondary to primary, or setting up
396
a type of preventative secondary care. As such it is likely to be broadly
transferable to other developed world settings. It is quite possible that some
types of telehealthcare may work well in the developing world as well. This is
because the developing world often has good mobile telephone network
coverage. However, the devices for interfacing with patients require reliable
electric power and skilled labour which might be more difficult to secure in
such countries.
In general, the biases seen included a lack of proper randomisation, problems
with allocation concealment and a lack of overt statements of specific
methodological processes, such as blinding. Therefore many of the risk of
bias tables were often populated with the judgement ’unclear’ due to
insufficient information.
In summary, this review has found 21 randomised controlled trials researching
the effects of telehealthcare intervention for asthma. In addition, we found 21
ongoing studies and unpublished trials. Despite some lack of consistency in
the way telehealthcare has been researched thus far (see below), this
represents a substantial body of reasonably high quality research in different
settings internationally and assessing different technologies. Some positive
conclusions can be drawn. Telehealthcare improves access and is no worse
than normal care in carefully selected and triaged primary and secondary care
patients. It does not, however, appear to have the desired impact on quality of
life and these interventions have little or no significant impact on emergency
department visits. However, given careful candidate selection conditions,
telehealthcare may reduce hospital admission rates and associated costs.
There is also some evidence for improved symptoms in telehealthcare trial
arms where symptoms are dealt with quickly and exacerbations are prevented
in a way not open to the control arm.
One of the problems with telehealthcare research that we have seen in this
review is that the control arm is not always usual care, but often receives an
enhanced input from the clinicians–for example, a greater number of visits or
face-to face contacts and so this results in greater adherence to medications
397
and surveillance of disease fluctuations in both arms of the trial and improved
outcomes for both arms. Telehealthcare fits into the Medical Research
Council’s model of a “complex intervention”20 and as such it seems that each
telehealthcare intervention is very diverse. This can make it very hard to
pinpoint the “active ingredients” of a telehealthcare intervention. Confusion
also comes from the different modes of delivery within the bracket of
telehealthcare. For example, both Internet- and telephone-based “helplines”
improve access to clinicians for patients. In this example it is the function of
improving access which leads to reduced hospitalisations–and so far the
evidence shows that this is not to the detriment or advantage of quality of life.
The form of the intervention, whether telephone or Internet, seems to have
less of an impact than the function of the access which results in timely advice
to prevent exacerbation in a way that outpatient clinics cannot deliver in
comparison. It is important then in future studies that interventions are
described as fully as possible, that relevant process or intermediary measures
are studied, and consideration is also given to embed qualitative work within
these complex intervention trials to help assess how these interventions exert
their effects.
Future telehealthcare interventions studied are likely to be even more
complex and include the features of web 2.0. Such research is important to
find how best to harness these innovative technologies without inadvertently
causing harm. It is, however, important that these interventions are patientcentred, and that they are developed through close consultation both with
patients, but also with healthcare professionals to maximise the chances of
success.56 57
Acknowledgements: This chapter is based on our review which is published in The
Cochrane Library: McLean S, Chandler D, Nurmatov U, Liu J, Pagliari C, Car J, et al.
Telehealthcare for asthma. Cochrane Database Syst Rev. 2010 Oct 6;(10):CD007717.
Margaret Fletcher translated the Ricci paper from Italian, for which we are grateful. We are
grateful to our co-authors Joe Liu and David Chandler for their help with this Cochrane
review. Finally, our thanks to colleagues in the Cochrane Airways Group for their help and
support in undertaking this review. Permission to reproduce this material has been applied
for.
398
References
1. BTS/SIGN. British Guideline on Asthma. British Thoracic Society, Scottish
Intercollegiate Guidelines Network, National Guideline, 2008.
2. Editorial. A plea to abandon asthma as a disease concept. Lancet
2006;368(9537):705
3. Editorial.
Asthma:
still
more
questions
than
answers.
Lancet
2008;372(9643):1009
4. Masoli M, Fabian D, Holt S, Beasley R. Global Burden of Asthma: The Global
Initiative for Asthma, 2003.
5. Anderson H, Gupta R, Strachan D, Limb E. 50 years of asthma: UK trends from
1955 to 2004. Thorax 2007;62(1):85-90
6. Asher M, Montefore S, Bjorksten B, Lai C, Strachan D, Weilland S, et al.
Worldwide time trends in the prevalence of symptoms of asthma allergic
rhinoconjunctivitis and eczema in childhood: ISAAC Phases One and Three
repeat muli-country cross-sectional surveys. Lancet 2006;368:733-743
7. Pearce N, Sunyer J, Cheng S, Chinn S, Bjorksten B. Comparison of asthma
prevalence in the ISAAC and the ECRHS. European Respiratory Journal
2000;16:420-6
8. Anandan C, Nurmatov U, Sheikh A. Is the prevalence of asthma declining,
systematic review of epidemiological studies. Allergy 2010;65(2):152-67
9. Punekar Y, Sheikh A. Establishing the incidence and prevalence of cliniciandiagnosed allergic conditions in children and adolescent using routinely collected
data from general practices. Clinical Experimental Allergy 2009;39(8):1209-16
10. Simpson C, Sheikh A. Trends in the epidemiology of asthma in England: a
national study of 333,294 patients. J R Soc Med 2010;103(3):98-106
11. Stewart A, Mitchell E, Pearce N, Strachan D, Weiland S, ISAAC. The relationship
of per capita gross national product to the prevalence of symptoms of asthma
and other atopic diseases in children (ISAAC). International Journal of
Epidemiology 2001;30:173-9
12. Weiland S, Bjorksten B, Brunekreef B, Cookson W, von Mutius E, Strachan D.
Phase II of the International Study of Asthma and Allergies in Childhood (ISAAC
II): rationale and methods. Eur Respir J 2004;24(3):406-12
13. Marks G. Geographical heterogeneity of asthma International Journal of
Epidemiology 2001;30:179-80
14. Anandan C, Gupta R, Simpson C, C F, Sheikh A. Epidemiology and disease
burden from allergic disease in Scotland: analyses of national databases. Journal
of the Royal society of Medicine 2009;102(10):431-42
15. Gupta R, Sheikh A, Strachan D, Anderson H. Burden of allergic disease in the
UK: secondary analyses of national databases. Clinical Experimental Allergy
2003;34(4):520-526
16. Car J, Sheikh A. Telephone consultations. BMJ 2003;326:966-9
17. Car J, Sheikh A. Email consultations in healthcare: 1- scope and effectiveness.
BMJ 2004;329:435-438
18. Car J, Sheikh A. Email consultation in healthcare: 2- acceptability and safe
application. BMJ 2004;329:439-42
19. McLean S, Sheikh A. Does telehealthcare offer a patient-centred way forward for
the community based management of long-term respiratory disease? Primary
Care Respiratory Journal 2009;18(3):125-126
20. Craig P, Dieppe P, MacIntyre S, Michie S, Nazareth I, Petticrew M. Developing
and evaluating complex interventions: new guidance.: Medical Research Council,
2008.
21. Finkelstein J. INHALED STEROID ADHERENCE IN MODERATE & SEVERE
ASTHMA. CRISP (Computer Retrieval of Information on Scientific Projects)
2000;31/07/2005 End date
399
22. McKinstry B, Pinnock H, Sheikh A. Telemedicine for management of patients with
COPD? Lancet 2009;374(9691):672-3
23. Currell R, Urquhart C, Wainwright P, Lewis R. Telemedicine versus face to face
patient care:effects on professional practice and health care outcomes. Cochrane
database of systematic reviews 2008;2000(2)
24. Mair F, Whitten P. Systematic review of studies of patient satisfaction with
telemedicine. BMJ 2000;320:1517-1520
25. Whitten P, Mair F, Haycox A, May C, Willliams T, Hellmich S. Systematic review
of cost effectiveness studies of telemedicine interventions. BMJ 2002;324:1434-7
26. Bynum A, Hopkins D, Thomas A, Copeland N, Irwin C. The Effect of
Telepharmacy Counselling on Metered-Dose Inhaler Technique Among
Adolescents With Asthma in Rural Arkansas. Telemedicine Journal and E-Health
2001;7(3):207-217
27. Barbanel D, Eldridge S, Griffiths C. Can a self-management programme
delivered by a community pharmacist improve asthma control? A randomised
trial. Thorax 2003;58(10):851-854
28. Clark NM, Gong ZM, Si JW, Lin X, Bria WF, Johnson TR. A randomized trial of a
self-regulation intervention for women with asthma. Chest 2007;132(1):88-97
29. Donald KJ, McBurney H, Teichtahl H, Irving L. A pilot study of telephone based
asthma management. Australian Family Physician 2008;37(3):170-3
30. Guendelman S, Meade K, Benson M, Chen YQ, Samuels S. Improving asthma
outcomes and self-management behaviors of inner-city children: a randomized
trial of the Health Buddy interactive device and an asthma diary. Archives of
Pediatrics & Adolescent Medicine. 2002;156(2):114-20
31. Jan RL, Wang JY, Huang MC, Tseng SM, Su HJ, Liu LF. An internet-based
interactive telemonitoring system for improving childhood asthma outcomes in
Taiwan. Telemedicine Journal and e-Health 2007;13(3):257-268
32. Ostojic V, Cvoriscec B, Ostojic SB, Reznikoff D, Stipic-Markovic A, Tudjman Z.
Improving asthma control through telemedicine: A study of short-message
service. Telemedicine Journal & E-Health 2005;11(1):28-35
33. Willems D, m J, J H, van Duurling R, Wouters E, J S. The effectiveness of nurseled telemonitoring of asthma results of a randomized controlled trial. Journal of
Evaluation in Clinical Practice 2008;14:600-608
34. Pinnock H, Bawden R, Proctor S, Wolfe S, Scullion J, Price D, et al. Accessibility,
acceptability, and effectiveness in primary care of routine telephone review of
asthma: pragmatic, randomised controlled trial. BMJ 2003;326(7387):477-9
35. Donald K, McBurney H, Teichtahl H, Irving L, Browning C, Rubinfeld A, et al.
Telephone based asthma management; Financial and individual benefits.
Australian Family Physician 2008;37(4):212-275
36. Pinnock H, McKenzie L, Price D, Sheikh A. Cost-effectiveness of telephone or
surgery asthma reviews: economic analysis of a randomised controlled trial. Br J
Gen Pract 2005;55(511):119-24
37. Willems DC, Joore MA, Hendriks JJ, van Duurling RA, Wouters EF, Severens JL.
Process evaluation of a nurse-led telemonitoring programme for patients with
asthma. Journal of Telemedicine & Telecare 2007;13(6):310-7
38. Willems DCM, Joore MA, Hendriks JJE, Wouters EFM, Severens JL. Costeffectiveness of a nurse-led telemonitoring intervention based on peak expiratory
flow measurements in asthmatics: Results of a randomised controlled trial. Cost
Effectiveness and Resource Allocation 2007;5
39. Kokubu F, Suzuki H, Sano Y, Kihara N, Adachi M. Tele-medicine system for highrisk asthmatic patients. Arerugi - Japanese Journal of Allergology
1999;48(7):700-12
40. Kokubu F, Nakajima S, Ito K, Makino S, Kitamura S, Fukuchi Y, et al.
Hospitalization reduction by an asthma tele-medicine system. Arerugi - Japanese
Journal of Allergology 2000;49(1):19
400
41. Chatkin JM, Blanco DC, Scaglia N, Wagner MB, Fritscher CC. Impact of a lowcost and simple intervention in enhancing treatment adherence in a Brazilian
asthma sample. Journal of Asthma 2006;43(4):263-266
42. Gruffydd-Jones K, Hollinghurst S, Ward S, Taylor G. Targeted routine asthma
care in general practice using telephone triage. British Journal of General
Practice 2005;55(521):918-923
43. Khan MSR, O'Meara M, Stevermuer TL, Henry RL. Randomized controlled trial of
asthma education after discharge from an emergency department. Journal of
Paediatrics & Child Health 2004;40(12):674-677
44. Pinnock H, Adlem L, Gaskin S, Harris J, Sheikh A. Accessibility clinical
effectiveness and practice costs of providing a telephone option for routine
asthma reviews: phase IV controlled implementation study. British Journal of
General Practice 2007;57(714-722)
45. Vollmer WM, Kirshner M, Peters D, Drane A, Stibolt T, Hickey T, et al. Use and
impact of an automated telephone outreach system for asthma in a managed
care setting. American Journal of Managed Care 2006;12(12):725-733
46. Chan DS, Callahan CW, Hatch-Pigott VB, Lawless A, Proffitt HL, Manning NE, et
al. Internet-based home monitoring and education of children with asthma is
comparable to ideal office-based care: results of a 1-year asthma in-home
monitoring trial. Pediatrics 2007;119(3):569-78
47. Cruz-Correia R, Fonseca J, Lima L, Araujo L, Delgado L, Castel-Branco MG, et
al. A comparison of web-based and paper-based self management tools for
asthma: Patients' opinions and quality of data in a randomised crossover study.
Journal on Information Technology in Healthcare 2007;5(6):357-371
48. Rasmussen LM, Phanareth K, Nolte H, Backer V. Internet-based monitoring of
asthma: A long-term, randomized clinical study of 300 asthmatic subjects.
Journal of Allergy & Clinical Immunology 2005;115(6):1137-1142
49. de Jongste J, Carraro S, Hop W, Group TCS, Baraldi E. Daily telemonitoring of
exhaled nitric oxide and symptoms in the treatment of childhood asthma.
American Journal of Respiratory and Critical Care Medicine 2009;179:93-97
50. van der Meer V, Bakker M, van den Hour W, Rabe K, Sterk P, Kievit J, et al.
Internet-based self management plus education compared with usual care in
asthma. Annals of Internal Medicine 2009;151:110-120
51. Juniper E, Guyatt GH, Cox FM, Ferrie PJ, King DR. Development and validation
of the Mini Asthma Quality of Life Questionnaire. European Respiratory Journal
1999;14(32-8)
52. Juniper EF, Guyatt GH, Feeny DH, Ferrie PJ, Griffith LE, Townsend M.
Measuring quality of life in the parents of children with asthma. Qualty of Life
Research 1996;5:27-34
53. Pinnock H, Sheikh A, Bawden R, Proctor S, Wolfe S, Scullion J, et al. Cost
effectiveness of telephone vs face to face consultations for annual asthma
review: randomised controlled trial in UK primary care [Abstract]. European
Respiratory Journal 2003;22(Suppl 45):Abstract No: [2250]
54. Tulu B, Chatterjee S, Maheshewari M. Telemedicine taxonomy: a classification
tool. Telemed Journal E Health 2007;13(3):349-58
55. Delaronde S, Peruccio D, Bauer B. Improving asthma treatment in a managed
care population. The American Journal of Managed Care 2005;11(6):361-368
56. Catwell L, Sheikh A. Evaluating eHealth interventions: the need for continuous
systemic evaluation. . Plos med 2009;6(8):e1000126
57. Catwell L, Sheikh A. Information technology system users must be allowed to
Prim Care 2009;17(1):1-4
401
Chapter 15
Case study: Internet-based interventions for smoking
cessation
Summary
•
Smoking is the leading cause of preventable morbidity and mortality.
•
With the increasing penetration of the Internet, online resources represent
a potentially important means of promoting smoking cessation, particularly
amongst younger people.
•
A variety of educational and behavioural support tools have been
developed to try and promote smoking cessation, but the effectiveness of
these approaches remains unclear.
•
Our Cochrane review identified 20 eligible trials, which studied Internetbased interventions of a range of intensities and complexities, comparing
these with either no intervention or a different Internet-based resource.
•
This review suggested that more engaging, tailored, intensive approaches
may have modest short-term effects (at three months), particularly if used
alongside other pharmacological measures; the impact on longer-term
abstinence however remains unclear.
•
Future studies should focus on establishing the longer-term (i.e. over at
least six and preferably 12 months) effectiveness and cost-effectiveness of
Internet-based smoking cessation programmes.
•
It is important that future trials also seek to describe the likely mechanisms
through which these interventions may (or may not) be exerting their
effects–they should therefore also report data on patient satisfaction,
changes in knowledge, motivation, dependency, quit attempts and safety
considerations.
402
15.1 Introduction
Tobacco smoking is a major preventable cause of death in both developed
and developing countries. Every day over 13,000 people die from tobaccorelated diseases.1 If current trends continue, by 2025 tobacco will contribute
to the death of 10 million people worldwide each year, with seven million of
these deaths occurring in developing countries.2 People who smoke are more
prone to developing various types of cancer, such as those of the oral cavity,
larynx, bladder and particularly lung cancer. Tobacco smokers are also at
substantially increased risk of developing heart disease, stroke, emphysema
and other fatal diseases.1 Smoking also imposes a huge economic burden on
society–currently up to 15% of the total healthcare costs in developed
countries.3 Additionally, passive smoking is associated with serious
morbidity.4 To reduce the growing global burden of tobacco-related mortality
and morbidity, and the impact of tobacco use on economic indicators, tobacco
control has become a world-wide public health imperative.1
Prevention and cessation are the two principal strategies in the battle against
tobacco smoking. Nicotine is highly addictive.5 There is evidence that, for
example, although 70% of US smokers say they want to quit, only five per
cent are able to sustain cessation for one year.6 The balance between the
individual's motivation to stop smoking and his or her dependence on
cigarettes influences smoking cessation success. Dependence in smokers
and their motivation to stop smoking can be assessed by simple questions.7
Maximising the delivery of smoking cessation interventions can achieve more
in terms of years of life saved and economic benefits than most medical
interventions for smoking-related illnesses.8
There is good evidence for the effectiveness of brief, therapist-delivered
interventions, such as advice from a clinician.9 There appears to be additional
benefit from more intensive behavioural interventions, such as group
therapy,10 individual counselling11 and telephone counselling.12 However,
these more intensive therapies are usually dependent on a trained
professional delivering or facilitating the interventions. This is both expensive
and time-consuming for the health providers, and often inconvenient to the
403
patient, because of lengthy waiting times and the need to take time off work.
Another major limitation of these more intensive types of interventions is that
they reach only a small proportion of those who smoke.
It is estimated that in 2009, there were 1.73 billion Internet users worldwide,13
and the number of Internet users is likely to increase rapidly over a relatively
short time-frame.14 The Internet has the potential to deliver behaviour change
interventions.15-18 Internet-based material is an attractive intervention tool,
because of relatively low costs per user, resulting in high cost-effectiveness.17
The Internet can be accessed in people's homes, in public libraries and
through other public Internet access points, such as Internet cafes and
information kiosks. The Internet is available 24 hours a day and 365 days a
year, even in areas where there are not the resources for a smoking cessation
clinic (such as some rural or deprived areas and low-income countries).
Online treatment programmes are convenient from the users' perspective,
because content can be accessed at any time, and they also offer a greater
level of anonymity than in-person or phone-based counselling. They have the
potential to reach audiences who might not otherwise seek support, because
of limited health care provision or possible stigmatisation. Existing smoking
cessation services such as advice from health professionals and nicotine
replacement therapy (NRT) are under-utilised by young people.19 Internet use
by young people has grown exponentially and has a powerful role in
influencing youth culture. It may therefore reach a target population of young
people who smoke more effectively than the more traditional providers are
able to do.
The Internet is a promising vehicle for delivering smoking cessation treatment
either as a stand-alone programme or as an adjunct to pharmacotherapy.17 18
According to the Pew Internet & American Life Project,20 seven per cent of
adult US Internet users, (approximately eight million people), reported having
404
searched online for information on 'how to quit smoking'. In the US, 18% of
those with less than high school education searched the web for information
on how to quit smoking, which represents a higher proportion than those with
more education.20
Materials tailored for individual smokers are more effective than untailored
ones, although the absolute size of the effect is still small.21 Internet
programmes can be highly tailored to mimic the individualisation of one-to-one
counselling. A web-based programme that collects relevant information from
users and tailors the intervention to their specific needs had significant
advantages over a web-based, non-tailored cessation programme.22 A Dutch
study explored existing self-help materials which are currently available in the
Netherlands, and found them to be ineffective for smoking cessation.
However, this study suggested that computer-tailored interventions could
potentially be successfully designed, and may be a promising means of
communicating information on smoking and cessation.23 Another study on the
efficacy of web-based tailored behavioural smoking cessation materials for
nicotine patch users showed that participants in the tailored programme
reported significantly higher 12-week continuous abstinence rates (22.8%)
than those in a non-tailored programme (18.1%).22
Using the Internet for smoking cessation programmes may also have
limitations. There are a large number of smoking cessation web sites, but they
do not all provide a direct intervention. Some studies of popular smoking
cessation web sites and their quality24
25
suggest that smokers seeking
tobacco dependence treatment online may have difficulty discriminating
between the numerous sites available.25 In addition, web sites that provide
direct treatment often do not fully implement treatment guidelines and do not
take full advantage of the interactive and tailoring capabilities of the Internet.24
Furthermore, a study on rates and determinants of repeat participation in a
web-based health behaviour change programme suggested that such
programmes may reach those who need them the least. For example, older
405
individuals who had never smoked were more likely to participate repeatedly
than those who currently smoke.26 The Internet is also less likely to be used
by people
on
lower income,
who
are
more
likely to
smoke.27,28
There have been two recent reviews of this area.29 30 Myung et al. evaluated
both web-based and computer-based cessation programmes.29 The review
included 22 trials in a single random effects meta-analysis that showed a
significant effect on cessation; similar effects were also seen in a range of
subgroups including one limited to nine trials that used a web-based
intervention. They concluded there was evidence to support the use of web
and computer-based cessation programmes for adults who smoked, but not
adolescent smokers. Shahab et al focused on interactive online interventions,
and also sought to identify treatment effect moderators and mediators.30 Their
review included eleven randomised trials. The authors concluded that web
based tailored and interactive interventions increased abstinence compared to
booklet or email control, based on three trials. Both reviews identified large
amounts of heterogeneity in both study designs and effect sizes.
Our review focused on determining the effectiveness of Internet-based
interventions for smoking cessation. Details of the methods employed can be
found in our Cochrane review.31
15.3 Main findings
15.3.1 Description of studies
Twenty studies met the inclusion criteria. Trialists were mostly based in the
US and recruited participants based there. One trial was conducted in
Switzerland, one in Norway, one in the Netherlands, one in England and one
in the Republic of Ireland. The studies by Munoz and colleagues recruited
from multiple countries.32
406
Recruitment was mainly web-based; participants found the sites through
search engines and browsing. Several trials used press releases, billboards,
television advertisements and flyers in addition to the web-based recruitment.
Therefore, participants included in these trials were smokers motivated to quit
smoking, who chose the Internet as a tool for smoking cessation
support. Clark et al33 recruited people undergoing chest computerised
tomography (CT) as a screening assessment for lung cancer at their first
follow-up visits. Strecher et al.34 recruited from members of two health
management organisations (HMOs) participating in the National Cancer
Institute's Cancer Research Network, Swan et al.35 recruited participants from
a large healthcare organisation. Sixteen studies recruited a full adult age
range. An et al.36 recruited young adult college students and three studies
recruited adolescents.37-39 Sample sizes ranged from less than 15037
38
to
nearly 12,000.40 In some studies, participants were offered financial
compensation for completing assessment surveys and for biochemical
analysis.32 36 37 38 41 47
A range of types of Internet interventions was tested in the included studies,
from a very low intensity intervention, providing a list of web sites for smoking
cessation,33 to highly intensive interventions consisting of Internet, email and
mobile phone delivered components.42,43 Tailored Internet interventions
differed by the amount of tailoring, from a bulletin board facility,44 a multimedia
component (McKay 2008),45 tailored and personalised access
22,46
) to very
high-depth tailored stories and highly personalised message sources.34
Sixteen studies assessed smoking status at least six months after the start of
the intervention.15,32-39,
41-43,45-48
All of these except two34 41 also assessed
smoking at one or more intermediate follow-up points. Four studies followed
participants for less than six months.17,22,40,44 Studies reported a range of
definitions of abstinence at the time of follow-up. Where studies reported
abstinence rates for more than one definition we displayed the effect using the
most conservative outcome (with the exception of An et al36). For most
studies, seven-day smoking abstinence was the main outcome measure.
15,17,32-34,40,41,44,45,47,48
A few studies reported 30-day self-reported smoking
407
abstinence.35,38,39 One study assessed six-month prolonged abstinence from
smoking;36
this study also reported seven-day and 30-day prevalence
abstinence. We used 30-day rates as our primary outcome because the
programme did not involve setting a quit date, and the prolonged abstinence
was based on self-report of time since last cigarette rather than repeated
assessments of abstinence. Three of the four short-term studies assessed
self-reported point prevalence abstinence at three-month follow-up only17,40,44
whilst Strecher et al.22 assessed 28-day continuous abstinence rates at sixweek follow up, and 10-week continuous abstinence rates at 12-week follow
up. In one study, seven-day smoking abstinence was a secondary outcome,
while time spent on the website, utilisation of pages, cessation aids used in
the past and during the study period were the main outcome measures.44
Due to the limited face-to-face contact and due to data collection through
Internet or telephone interview, biochemical validation to confirm self-reported
smoking abstinence was conducted in only five trials.15,33,36,38,39 All these
measured carbon monoxide (CO) in expired air.
Utilisation of the Internet site or programme use was a secondary outcome
measure in 10 studies.15,17,33-35,38,43,45,46,48
Satisfaction with treatment
condition was assessed in five trials.22,37,41,44,48 Use of NRT or other
pharmacotherapies was a secondary outcome measure in five trials.
34,35,39,43,45
Two of the studies in adolescents assessed reductions in the
number of cigarettes or in smoking frequency as secondary outcomes.37,39
Munoz et al48 reported the impact of method of obtaining follow-up data on
quit rates; comparing phone with online follow-up procedure.
We first grouped studies in subgroups according to whether: (1) they
compared an Internet intervention with no intervention or a non-Internet
intervention; or (2) compared a more complex or tailored Internet intervention
with a less complex one. At the suggestion of peer reviewers we added a third
comparison that included studies in which the intervention offered access to
an interactive website and the control could be either a static website, or a
control without Internet access. In five trials, all participants were using, or
408
were offered, pharmacotherapy15,22,34,35,43 and the Internet component was
thus being evaluated as an adjunct to pharmacotherapy. These were grouped
in comparisons based on the nature of the Internet component and the
control.
15.3.2 Internet intervention compared to no Internet intervention or no
intervention at all
Ten trials compared an Internet intervention to a non-Internet based smoking
cessation intervention or to a no intervention control.15,17,33,35-39,42,43 Six trials
recruited adults, one targeted young adult university students
36
and three
were conducted among adolescents.37-39
Of the six trials that targeted adults, two studies
42,43
evaluated 'Happy
Endings', a 1-year programme delivered via the Internet and cell phone,
consisting of more than 400 contacts by email, web pages, interactive voice
response, and short message service (SMS) technology. Brendryen
et al
42
recruited people attempting to quit without NRT, whilst in their other trial
43
they offered a free supply of NRT to all participants. Follow-up was after 12
months in both studies. Japuntich et al15 evaluated a web-based system
incorporating information, support and problem-solving assistance; this was
tested as an adjunct to bupropion and brief face-to-face counselling, with
follow up after six months. Swartz et al.17 recruited participants via worksites
to test a video-based Internet site that presented strategies for smoking
cessation and motivational materials tailored to the user’s race/ethnicity, sex
and age. This study only reported short follow up (i.e. at 90 days); after this
time the control group had access to the programme. Clark et al.33 tested a
very low intensity intervention for smokers having CT lung screening; a
handout with a list of 10 Internet sites related to stopping smoking with a brief
description of each site compared to printed self-help materials. Follow-up
was at 12 months. Swan et al.35 was a three arm trial comparing an
established proactive telephone counselling intervention, an interactive
website based on the same programme, and a combination of phone and
Internet components, all providing behavioural support in conjunction with
409
varenicline
use.
An et al36 recruited college students who reported smoking in the past 30
days; the sample smoked an average of four cigarettes per day. Intervention
group participants received $10 a week to visit an online college magazine
that provided personalized smoking cessation messages and peer email
support. The control group received only a confirmation email containing links
to online health and academic resources. Both groups were informed about a
campus-wide ‘Quit&Win’ contest sponsored by University Health Service.
The three small studies in adolescents recruited populations of relatively light
smokers. Mermelstein et al38 evaluated the effectiveness of enhancing the
American Lung Association’s Not on Tobacco programme (NOT) with a webbased adjunct (NOT Plus), which included access to a specially designed
web-site for teenagers, along with proactive phone calls from the group
facilitator to the participant. Twenty-nine high schools were randomly
assigned to either the NOT programme alone or to the NOT Plus one. Selfreported smoking behaviour was assessed at the end of the 10-week
programme and three months later. Patten et al39 compared a home-based,
Internet-delivered treatment for adolescent smoking cessation with a clinicbased brief office intervention (BOI) consisting of four individual counselling
sessions. Adolescents assigned to the Internet condition had access to the
web-site for 24 weeks and abstinence was assessed at the end of this
period. Woodruff et al.37 evaluated an Internet-based, virtual reality world
combined with motivational interviewing, conducted in real-time by a smoking
cessation counsellor. There was a measurement-only control condition
involving four online surveys. Smoking status was assessed at baseline, postintervention and at three- and 12-month follow-up.
15.3.3 Comparisons between different Internet interventions
Ten trials compared two or more different Internet interventions.22,32,34,40,41,44-48
Three of these studies had only short follow-up.22,40,44 In two studies all
participants
were
using
nicotine
patch
pharmacotherapy.22,34
410
A series of three studies by Munoz and colleagues evaluated adjuncts to an
online resource, the 'Guia', a National Cancer Institute evidence-based
intervention first developed for Spanish-speaking smokers. In separate
English language47 and Spanish32 studies by Munoz et al the control condition
was the provision of access to the Guia and 'Individually Timed Educational
Messages' (ITEMs). The intervention tested was the addition of an online
mood management course consisting of eight weekly lessons. Munoz et
al48 also used the 'Guia' as the control condition, but in a four-arm design that
evaluated the successive addition of ITEMs; the mood management condition
used in their earlier studies;32,47 and a 'virtual group' asynchronous bulletin
board. The study recruited English- and Spanish-speaking Internet users from
68 countries.
In two trials,22,46 the control condition provided access to a relatively static
Internet site whilst one or more intervention conditions provided more tailored
and personalised access. Rabius et al46 compared five tailored and interactive
Internet services with the targeted, minimally interactive American Cancer
Society site providing stage-based quitting advice and peer modelling. Followup surveys were conducted four and 13 months after randomization. Strecher
et al.22 assigned purchasers of a particular brand of nicotine patch to receive
either web-based, tailored behavioural smoking cessation materials or webbased non-tailored materials. This study measured smoking behaviour at
three-month follow-up.
The remaining five studies were distinctive in their designs. Etter et
al.40 compared the efficacy of two versions of an Internet-based, computer
tailored cessation programme; the control group received a shorter version
modified for use by those smoking and buying NRT over the counter, although
use of NRT was not a condition for enrolment. Follow-up was at 2.5
months. Stoddard et al.44 evaluated the impact of adding a bulletin board
facility to the 'smokefree.gov' cessation site. This study measured smoking
behaviour three months after enrolment. Strecher et al.34 identified active
psychosocial and communication components of a web-based smoking
cessation intervention and examined the impact of increasing the tailoring
411
depth on smoking cessation among nicotine patch users. Five components of
the intervention were randomised using a fractional factorial design: high
versus low depth tailored success story, outcome expectation versus efficacy
expectation messages; high versus low personalised source; and multiple
versus single exposure to the intervention component. Abstinence was
assessed after six months. McKay et al.45 compared the Quit Smoking
Network (QSN), a web-based tailored cessation programme with a multimedia
component, with Active Lives, a web-based programme providing tailored
physical activity recommendations and goal setting in order to encourage
smoking cessation. Abstinence was assessed after six months. Te Poel et
al.41 compared tailored to untailored cessation advice letters sent by email,
after participants had completed an online survey. We included this study in
the review because information was gathered via a website.
15.3.4 Risk of bias in included studies
Allocation
The majority of studies did not explicitly describe the way in which the
randomisation sequence was generated or concealed until patient enrolment.
In eleven studies, computer randomisation was used to assign participants to
intervention or control condition.17,22,32,35,40-44,47,48 Although there was little
information
about
allocation
concealment,
when
investigators
used
computerised randomisation and had minimal interaction with participants we
judged there to have been a low risk for selection bias. Munoz et al32,4748 used
the baseline questionnaire to automatically implement stratified randomisation
by gender and major depressive episode (MDE) status (no MDE history, past
MDE, current MDE) to the two conditions.
In the two studies37,38 that randomised schools to conditions there was the
potential for bias due to the way in which individual students were recruited
once their school was randomised. In both there were differences in the
baseline smoking behaviour of intervention and control participants. The two
studies also needed to take account of the non-independence of outcomes for
students clustered within schools. Mermelstein et al.
linear modelling to allow for clustering. Woodruff et al.
37
38
used hierarchical
assessed baseline
412
variable intraclass correlations and average cluster sizes. Intraclass
correlations were generally small (0.1 or less) and the magnitude of the effect
sizes was below two, so analyses were conducted at the individual level
without a school-level cluster term.
Incomplete outcome data
All studies included in this review used ITT analysis, reporting analyses based
on the total number randomised, with drop-outs and participants lost to follow
up classified as smoking. In one study33 there were no drop-outs at one-year
follow up, as all study participants attended their annual review at that point.
Five studies
15,34,36,42,43
ascertained smoking status for over 80% of
participants at follow-up. Five studies ascertained smoking status for 50-80%
of participants at follow-up.17,22,35,37-39 Eight studies had over 50% loss to
follow up.32,40,41,44-48 All studies reported similar proportions loss to follow up in
each group except in one study where survey non-response was higher
among intervention participants then among controls.37 We tested the
sensitivity of our results to the assumption that lost participants were still
smoking, by excluding those not reached at follow-up from the denominator in
the analyses, and report this below.
15.3.5 Effects of interventions
Internet intervention compared to no Internet interventions or no
interventions at all
There were six trials in adult populations.15,17,33,35,42,43 Both the Happy Endings
trials detected a significant effect of the Internet intervention on sustained
abstinence at 12 months whether it was compared to a self-help control,42 RR
2.94, (95%CI 1.49, 5.81) or tested as an adjunct to NRT43 RR 1.71, (95%CI
1.10, 2.66). Whilst the relative benefit of the intervention was similar,
prolonged abstinence at 12 months was higher, 13%, amongst control group
participants in Brendryen et al.,43 who were offered NRT, than amongst
controls who were not given pharmacotherapy (seven per cent in Brendryen
et al.).42 The authors noted that the relative effects were smaller for the
outcome of point prevalence abstinence at 12 months, rather than repeated
abstinence, because abstinence increased over time in the control group in
413
both trials. Japuntich et al.15 did not detect a significant effect of the Internet
component as an adjunct to counselling and bupropion (RR 1.27, 95%CI 0.70,
2.31) whilst Swan et al.35 did not detect an effect of an Internet component
either compared to, or as an adjunct to, telephone counselling and varenicline
(RR 0.94 95%CI 0.79, 1.13, combining web only and web plus phone arms).
All three conditions achieved similar quit rates, ranging from 27.4% to 30.6%
for 30 day abstinence at six months. Clark et al.,33 which was only a minimal
intervention, also did not detect an effect (RR 0.45, 95%CI 0.14, 1.40).
Relative effects were similar in these trials at shorter follow-up points. The
study by Swartz et al.17 with only short-term follow-up detected a significant
effect of Internet intervention compared to no intervention at all (RR 2.46,
95%CI 1.16, 5.21).
The trial by An et al.36 in a population of college students detected a
significant effect on 30-day abstinence at 30-week follow-up (RR 1.95, 95%CI
1.42, 2.69) although rates of prolonged abstinence were only six per cent and
did not differ between groups.
Patten 200639 compared a home-based Internet delivered intervention (SOS)
to a brief office intervention (BOI) for adolescent smoking cessation, and did
not detect a difference in abstinence. Rates at 24 and 36 weeks follow-up
were higher for BOI (RR 0.44, 95%CI 0.14, 1.36 at 36 weeks). Mermelstein et
al.38 detected a significant effect of the web-based adjuncts to the groupbased approaches for adolescent smoking cessation (crude RR 1.96, 95%CI
1.02, 3.77; also reported as significant, (P<0.05), using mixed model logistic
regression to account for clustering within schools). Woodruff et al.37 recruited
eligible adolescents based on a report of smoking in the past month; at
baseline some described themselves as 'former' smokers or had not smoked
in the past week. Intervention participants had lower past week abstinence
rates at baseline than controls (14% vs. 29%). At the post-assessment, they
had significantly higher abstinence rates than controls (35% vs. 22%), but by
the final 12-month follow-up, the two groups had almost identical past-week
abstinence rates (RR 0.93, 95%CI 0.60, 1.44). The interaction term
considering all four assessments was not significant. Intervention participants
414
(68%, n=52) completed a five item questionnaire assessing their satisfaction
with the programme immediately after the post-test assessment; 89% of
participants reported they would recommend the programme to another
person who smoked.
Comparison of different Internet intervention
None of the three Munoz studies detected significant benefits of adding a
mood management intervention, and pooling the comparable arms did not
make a difference at 12 months (RR 0.90, 95%CI 0.70, 1.15 ; I²=51%) or in
the shorter term. In fact Munoz et al.47 found that the more complex
intervention (GUIA + ITEMS + MM) yielded significantly lower quit rates at 12
months. All four arms of Munoz et al.48 had similar long- and short-term quit
rates, ranging from 19.1% to 22.7% at 12 months, with no evidence that the
more tailored conditions had any incremental benefit over the static website
control. Authors of this trial also assessed user satisfaction with the treatment
condition, and reported that there were significant differences in satisfaction
across conditions at all follow-up assessments, with groups 3 and 4 generally
reporting greater satisfaction at each time point.
Strecher et al.22 and Rabius et al.46 also compared tailored to static sites
but Strecher 200522 used pharmacotherapy and only reported short-term
outcomes, and Rabius et al.46 in their later trial compared multiple
sites. Rabius et al.46 detected no significant differences between a static
cessation site and any of the other Internet sites included in the evaluation.
Approximately 10% of enrolled participants assigned to the static site reported
abstinence after 13 months compared with 8% to 12% among those assigned
to any of the five different interactive sites (RR 1.12, 95%CI 0.92, 1.36 pooling
any interactive website versus the control site). At baseline, 30% of
participants reported an indicator of depression. Post-hoc analyses found that
this subgroup had significantly lower 13-month quit rates than those who did
not have signs of depression (8% vs. 12%, P <0.001). Amongst the 70% of
participants who did not report an indicator of depression at baseline, the
more interactive, tailored sites, as a whole, were associated with higher
quitting rates than the less interactive American Cancer Society site (13% vs.
415
10%, P =0.04). This exploratory analysis suggested that tailored, interactive
websites might help smokers who do not report the indicator of depression at
baseline to quit and maintain cessation. Strecher et al.22 reported higher
continuous abstinence rates in the tailored condition than the control, as an
adjunct to NRT. At 12 weeks, continuous abstinence rates were 22.8% vs.
18.1% respectively (RR 1.26, 95%CI 1.10, 1.44). Satisfaction with the
programme was also significantly higher in the tailored than in the non-tailored
programme.
McKay et al.45 detected no difference at three or six-month follow-up between
two cessation interventions, one with a focus on physical activity presenting
very few strategies for quitting smoking.
Te Poel et al.41 detected evidence of a benefit from a tailored email letter
compared to a non-tailored one, (RR 2.48, 95%CI 1.11, 5.55). All participants
in this study accessed a website once to complete a questionnaire about their
smoking.
Stoddard et al.44 compared the publicly available version of smokfree.gov,
designated as usual care condition (UC), to an identical-looking website that
included an asynchronous bulletin board (BB) and found that quit rates for
participants in both conditions were similar after three months (RR 0.95,
95%CI 0.64, 1.40). Satisfaction with the website was high and did not differ
significantly between conditions (UC: 90.2%, BB: 84.9%, P=0.08).
Strecher et al.34 compared multiple conditions in a fractional factorial design
so results are not displayed in the analyses. Participants could receive up to
three high-depth components, addressing efficacy expectations, outcome
expectations and success stories, as part of their tailored web-based
intervention. Tailoring depth was marginally related (P<0.08) to 6-month
smoking cessation in ITT analyses and was significantly related to cessation
using per-protocol analysis (excluding participants who used other cessation
aids during the follow-up period). The adjusted 6-month cessation rates
among participants receiving all the three high-depth tailored components was
416
27.7% in the ITT analysis. There was some evidence that high-depth tailored
success stories had a particular influence on participants with lower levels of
education although this interaction was not significant in the ITT analysis.
Participants reported using an average of 5.1 weeks of their supply of nicotine
patches, with 26.7% using the patch for the full 10 weeks.
Comparison of interactive Internet sites with static sites or non-Internet
controls
From
the
studies
discussed
above,
eight
studies
with
long-term
outcomes32,36,42,43,45-48 and another three with only short-term17,22,40 yielded
data that allowed pooling to assess whether sites that give smokers content
tailored to their needs and interests are more useful than a control. These
summary statistics do however need to be interpreted with caution because of
a
high
degree
of
statistical
heterogeneity
(I²=75%
for
long-term
outcomes, Figure 15.1, I²=76% for short-term outcomes, Figure 15.2).
Figure 15.1 Meta-analysis of effectiveness of tailored interactive vs. nontailored/non-Internet interventions on smoking abstinence at the longest
follow-up assessment point
417
Figure 15.2 Meta-analysis of effectiveness of tailored interactive vs. nontailored/non-Internet interventions on smoking abstinence at the shortterm follow-up assessment
15.4
Implications
for
policy,
practice
and
research
The Internet, with all its richness of options and opportunities for
communication and sharing information, has now become a regular part of
daily life for the majority of people in many countries. Therefore it is
appropriate to consider using it as a tool to increase choice and access to
smoking cessation support. Online treatment is convenient in that it can be
accessed anywhere, at any time; it also offers the option of anonymity. For
healthcare providers it has the potential for being very cost-effective if
provided as an automated service. Internet interventions for smoking
cessation can be provided in conjunction with other cessation support such as
individual or group counselling and NRT or other pharmacotherapy.
We identified 20 trials yielding data from nearly 40,000 participants. Despite
this large volume of data, this remains a relatively new field of research, with
all trials published since 2004. The studies included in this review assess
ways to help people quit smoking using a variety of Internet cessation
programmes. Trials varied in the interventions studied and in the duration of
the interventions; they also varied in the timing of follow-up assessments and
the way in which abstinence was defined.
418
There are only a small number of studies comparing the long-term effects of
Internet to a no-Internet or no intervention control. The results of these studies
have been mixed so there is as yet only limited evidence to support a longterm effect of programmes that use the Internet.
Two studies assessed whether an Internet intervention was more effective
than a self-help booklet for smoking cessation.42,43 These have shown
evidence of both short- and longer-term effect with up to one-year follow-up,
but the programme delivered to the treatment group was fully automated,
relatively intensive and proactive, and was delivered via both the Internet and
mobile phone. The three other studies reporting long-term results did not
detect a benefit, but these were relatively less intensive programmes with
fewer proactive elements to encourage programme use. One additional shortterm study providing tailored online cessation videos did detect an increase in
abstinence at three months.17
When considering trials comparing different intensities of Internet support,
there is some evidence of a short-term beneficial effect of individually tailored
Internet
programmes
compared
to
static
web-sites
or
non-tailored
programmes. For example, Etter et al found that an original more tailored
programme versus a modified and less tailored programme increased quit
rates and helped maintain short-term abstinence.40 Strecher et al found that
continuous abstinence rates at six and 12 weeks were significantly higher in
the tailored intervention than in the non-tailored intervention.22 In contrast,
however, one study found there was no significant difference in abstinence
rates between participants assigned to interactive sites and participants
assigned to static sites, even though interactive sites led to higher utilization
of the websites.46 There are also reported benefits of interactive sites with
regard to satisfaction of the users.22,37,48 One form of tailoring is to encourage
website use by sending tailored email messages. Three studies by Munoz
and colleagues used these as a component of an Internet programme, but
none of them evaluated the effect compared to a non-Internet control, and
only Munoz et al.48 used a static site as a control; this study failed to detect
any additional benefit of this particular tailoring. All three studies evaluated the
419
effect of adding a mood management component and none detected any
evidence that this was helpful, and the trend in two studies was for a reduction
in success. In a post-hoc analysis we explored pooling studies that evaluated
tailored, interactive Internet-based interventions, whatever the nature of the
control, but found there to be too much between study heterogeneity in effect
sizes.
Pharmacotherapies such as NRT and bupropion can help people who make a
quit attempt increase their chances of success.49,50 Face-to-face behavioural
support has an independent benefit, but many people are not willing to attend
or cannot access group-based programmes or multi-session counselling.
Internet programmes can be used to deliver additional behavioural support to
people using pharmacotherapy, but there is so far little evidence of an effect
of Internet intervention in addition to pharmacotherapies. Both Japuntich et
al15 and Swan et al.35 failed to detect any benefit, although Strecher et
al.22 suggested a short-term advantage of a tailored site over a static
one. Strecher et al in their later trial34 identified components of tailored sites
that could assist cessation.
Studies conducted among adolescents and young adults vary not only in the
Internet intervention used, but also in methodology and setting. All three trials
in adolescents were relatively small. Two suggested that the Internet condition
had less benefit than the comparison condition, while one reported a
marginally significant benefit sustained after the end of the programme.38 In
this study, participants in the web arm also had phone calls from a counsellor
so the independent effect of the web component is uncertain. These studies
show that Internet assistance is attractive and suggest that tailored web sites
are more popular among young people.36-38
To achieve success in smoking cessation programmes, interventions must be
accessible, efficacious and cost-effective and transportable. Only two of the
included studies in this review reported any information about costeffectiveness of their intervention.40,46 Etter et al estimated that the total cost
420
of implementing the website, for a reach of 8000 participants in computer
tailored programmes and for 600,000 visitors per year to the website, is
comparable to the cost of running a small smoking cessation clinic which
would treat about 50 smokers a month. Therefore, Internet services provide
greater potential for cost-efficiency because they can provide assistance to
many smokers at a very low cost. Internet interventions can also be delivered
alongside other more traditional smoking cessation programmes, providing
smokers who are motivated to quit smoking with different tools which
increases their overall choice.40 Rabius et al found Internet assistance for
smoking cessation to be cost-effective, since four days of programming at a
cost of less than US $2000 allowed approximately 5000 additional users for
services from the five tailored interactive service providers, comparing with
the much larger cost of serving 1000 new clients with telephone counselling
(approximately US $100,000).46
The trials enrolling adults generally relied on self-reported data on smoking
status. Biochemical validation of self-reported cessation was only attempted in
two trials where participants had face-to-face contact during the follow-up
visit.15,33 The Society for Research on Nicotine and Tobacco subcommittee on
biochemical verification in clinical trials51 considers that verification is not
necessary when a trial includes a large population with limited face-to-face
contact, and where the optimal data collection methods are through the mail,
telephone, or Internet. There is a recommendation that biochemical
verification be used in studies of smoking cessation in special populations,
including adolescents.51 Only one of the four studies in adolescents and
young people
did
not
use
biochemical
verification
of
self-reported
abstinence.37 Four included studies followed participants for less than six
months.17,22,40,44 It is hoped that reporting six-month outcomes will become
routine.52
Conducting research via the Internet provides opportunities to generate very
large sample sizes, but it is also methodologically challenging, because of
threats to internal and external validity such as selection bias or differential
drop-out.53,54 Although there was limited detail about procedures for sequence
421
generation and allocation concealment, we judged that the likelihood of
selection bias was small in studies that recruited participants online. Rates of
loss to follow-up were varied and were high in some large online studies. In
our primary analyses we followed the convention of assuming that all those
lost to follow-up continued to smoke. This can be argued to be a reasonable
approach with a volunteer sample and low attrition but if attrition rates are
high, or differential across conditions, the assumption may be wrong in some
cases and introduce bias.55 We undertook a sensitivity analysis ignoring
losses to follow-up by omitting them from the denominators, and did not find
any important differences in relative effects. A range of alternative
assumptions about those lost to follow-up could be tested, but it is important
to recognise that bias in the relative effect will only occur if there the
proportion of quitters amongst those lost to follow-up differs between
intervention and control group.
Determining the contribution of a specific website presents a difficult
challenge, since Internet users appear to access different sites when
searching for information or support to a specific topic. For example,
contamination in control groups may be difficult to prevent because of
unrestricted access to the Internet, whilst on the other side we cannot be sure
that the intervention group is using only the intended intervention.53,56
The two other recent reviews in this area drew somewhat less cautious
conclusions about the strength of the evidence for the effectiveness of
Internet interventions. One review29 pooled data from a number of studies of
both web- and computer-based interventions, and concluded that there is now
sufficient evidence to support the use of both categories of intervention for
adult smokers. Their estimate for Internet interventions based on nine studies,
using a random effects model, was RR 1.40, 95%CI 1.13, 1.72. Shahab et al
also suggest that interactive, web-based cessation can be effective in aiding
cessation,30 based on 11 studies, all but one of which is included in our review
(we were able to include longer-term data for the studies by Pike et al.57
and Rabius et al.46). We excluded the study by Prochaska et al,58 because we
were unable to confirm all data with the authors. Shahab et al.30 pooled the
422
studies in a number of subgroups; the intervention (tailored/untailored); length
of treatment; motivation to quit; and whether the intervention was fully
automated or not. They estimated interactive web-based smoking cessation
interventions to be effective compared to untailored booklet or e-mail
interventions (random effects RR 1.8, 95%CI 1.4, 2.3), but this was based on
just three trials; they also estimated that tailored interventions increase sixmonth abstinence by 17%. They also suggest that only interventions aimed at
smokers motivated to quit were effective (RR 1.3 95%CI 1.0, 1.7). We
consider that our more cautious conclusions are due to a conservative
approach to subgroup analyses, and our preference not to pool, even with a
random
effects
approach,
when
there
is
evidence
of
substantial
heterogeneity.
Further studies of the long-term effects of Internet-based cessation
interventions are clearly needed, and there are several ongoing studies in this
area. Future trials and reviews should include analyses of participants
according to socio-demographic data, in order to be able to identify the types
of smokers who seek Internet assistance in quitting smoking. Feil et al
detected that a large proportion of women was recruited in their study.53
According to our findings there is some evidence that females are more
interested in smoking cessation programmes delivered via Internet; only three
of the included trials reported that more males enrolled,32,37,47 and three had
equal number of male and female participants.33,39,43
In the future there may be an interest from patients with depression seeking
Internet assistance for quitting smoking. Although there is evidence that
depression is an important factor in smoking cessation,59 and that depression
inhibits quitting success by decreasing self-efficacy,60 only a few studies
evaluated the impact of Internet interventions among the subgroup of smokers
reporting depression. Rabius et al found no overall difference in quit rate
among smokers assigned to six experimental groups (five interactive and one
static site), but they also found that those who reported an indicator of
depression and were assigned to interactive site had lower cessation rates
than those assigned to the static site, although this difference was not
423
significant. The authors attribute these findings to the increased time
investment required from participants of interactive sites.46
There are a small number of studies which provide very limited evidence of
long-term benefits for programmes delivered only by the Internet compared to
no-Internet controls. There is some evidence that tailored Internet
interventions are more effective than non-tailored interventions. This should
therefore not for the present be a priority for the NHS.
Looking ahead, more rigorous studies comparing the long-term effects of
Internet interventions with non-Internet interventions or no intervention at all
are needed in order to determine the true long-term effectiveness of the
Internet as a tool for smoking cessation. These should, where possible,
assess outcomes using objective measures, and also assess costeffectiveness considerations. It is important that future trials also seek to
describe the likely mechanisms through which these interventions may (or
may not) be exerting their effects–they should therefore also report data on
patient satisfaction, changes in knowledge, motivation, dependency, quit
attempts and safety considerations.
Researchers should aim to assess smoking status after six months as a
minimum, so that the longer-term benefit of programmes can be determined
and meta-analyses of outcomes across studies be facilitated.
Acknowledgements: This chapter is based on a Cochrane review (Civljak M, Sheikh A,
Stead LF, Car J. Internet-based interventions for smoking cessation. Cochrane Database
Syst Rev. 2010 Sep 8;(9):CD007078) and we take pleasure in acknowledging our gratitude
to Marta Civljak and Lindsay Stead for their expert contributions to this work; our thanks are
also due to colleagues in The Cochrane Tobacco Addiction Group for their support.
Permission to reproduce this material has been applied for.
424
References
1. World Health Organization. Building blocks for Tobacco Control - A Handbook.
Serial, 2004.
2. Mackay J, Eriksen M, Shafey O. The Tobacco Atlas. 2nd Edition. American
Cancer Society, 2006.
3. Parrot S, Godfrey C. Economics of smoking cessation. BMJ 2004; 328: 947-49.
4. Scientific Committee on Tobacco and Health and HMSO. Report of the Scientific
Committee on Tobacco and Health. London: Her Majesty's Stationery Office,
1998.
5. United States Public Health Service. The Health Consequences of Smoking
Nicotine Addiction: a report of the Surgeon General. Office of the Surgeon
General 1988.
6. Schroeder SA. Conflicting dispatches from the tobacco wars. NEJM 2002; 347:
1106-09
7. West R. Assessment of dependence and motivation to stop smoking. BMJ 2004;
328: 338-39.
8. Coleman T. Cessation interventions in routine health care. BMJ 2004; 328: 63133.
9. Stead LF, Bergson G, Lancaster T. Physician advice for smoking cessation.
Cochrane Database Syst Rev 2004; 2: 10.1002/14651858.CD000165.pub3
10. Stead LF, Lancaster T. Group behaviour therapy programmes for smoking
cessation.
Cochrane
Database
Syst
Rev
2005;
2:
10.1002/14651858.CD001007.pub2
11. Lancaster T, Stead LF. Individual behavioural counselling for smoking cessation.
12. Stead LF, Lancaster T, Perera R. Telephone counselling for smoking cessation.
13. Internet
2009
in
numbers. Available
from:
http://royal.pingdom.com/2010/01/22/internet-2009-in-numbers/ (last accessed
2/1/11)
14. Modis T. The end of the internet rush. Techn Forecast Soc Change 2005; 72:
938-43.
15. Japuntich SJ, Zehner ME, Smith SS, Jorenbv DE, Valdez JA, Fiore MC, et
al.Smoking cessation via the internet: a randomized clinical trial of an internet
intervention as adjuvant treatment in a smoking cessation intervention. Nicot
Tobac Res 2006; 8: S59-67.
16. Strecher VJ, Shiffman S, West R. Moderators and mediators of a web-based
computer-tailored smoking cessation program among nicotine patch users. Nicot
Tobac Res 2006; 8: S95-101.
17. Swartz LH, Noell JW, Schroeder SW, Ary DV. A randomised control study of a
fully automated internet based smoking cessation programme. Tobac Contr
2006; 15: 7-12.
18. Graham AL, Cobb NK, Raymond L, Sill S, Young J. Effectiveness of an internetbased worksite smoking cessation intervention at 12 months. J Occ Environ
Med 2007; 49: 821-28.
19. Rodgers A, Corbett T, Bramley D, Riddell T, Wills M, Lin RB, et al. Do u smoke
after txt? Results of a randomised trial of smoking cessation using mobile phone
text messaging. Tobac Contr 2005; 14: 255-61.
20. Fox S. Health information online: eight in ten internet users have looked for
health information online, with increased interest in diet, fitness, drugs, health
insurance, experimental treatments, and particular doctors and hospitals.
Available from: http://www.pewinternet.org/Reports/2005/Health-InformationOnline/1-Summary-of-Findings.aspx (last accessed 2/1/11)
425
21. Lancaster T, Stead LF. Self-help interventions for smoking cessation. Cochrane
Database Syst Rev 2005; 3: 10.1002/14651858.CD001118.pub2
22. Strecher VJ, Shiffman S, West R. Randomized controlled trial of a web-based
computer-tailored smoking cessation program as a supplement to nicotine patch
therapy. Addiction 2005; 100: 682-88.
23. Dijkstra A, De Vries H, Roijackers J. Targeting smokers with low readiness to
change with tailored and nontailored self-help materials. Prev Med 1999; 28: 20311.
24. Bock B, Graham A, Sciamanna C, Krishanamoorthy J, Whiteley J, CarmonaBarros R, et al.Smoking cessation treatment on the internet: content, quality, and
usability. Nicot Tobac Res 2004; 6: 207-19.
25. Etter JF. A list of the most popular cessation web sites and a comparison of their
quality. Nicot Tobac Res 2006; 8: S27-34.
26. Verheijden MW, Jans MP, Hildebrandt VH, Hopman-Rock M. Rates and
determinants of repeated participation in a web-based behaviour change program
for healthy body weight and healthy life style. J Med Intern Res 2007; 9: e1.
27. Eysenbach G. Poverty, human development and the role of e Health. J Med
Intern Res 2007; 9: e34.
28. Kontos EZ, Bennett GG, Viswanath K. Barriers and facilitators to home computer
and internet use among urban novice computer users of low socioeconomic
position. J Med Intern Res 2007; 9: e 31.
29. Myung SK, McDonell DD, Kazinets G, Seo HG, Moskowitz JM. Effects of weband computer-based smoking cessation programs: meta-analysis of randomized
controlled trials. Arch Intern Med 2009; 169: 929-37.
30. Shahab L, McEwen A. Online support for smoking cessation: a systematic review
of the literature. Addiction 2009; 104: 1792-1804.
31. Civljak M, Sheikh A, Stead LF, Car J. Internet-based interventions for smoking
cessation. Cochrane Database Syst Rev 2010; CD007078.
32. Munoz RF, Lenert LL, Delucchi K, Stoddard J, Perez JE, Penilla C, et al. Toward
evidence-based Internet interventions: A Spanish/English Web site for
international smoking cessation trials. Nicot Tobac Res 2006; 8: 77-87.
33. Clark MM, Cox LS, Jett JR, Patten CA, Schroeder DR, Nirelli LM, et
al.Effectiveness of smoking cessation self-help materials in a lung cancer
screening population. Lung Cancer 2004; 44: 13-21.
34. Strecher VJ, McClure JB, Alexander GL, Chakraborty B, Nair VN, Konkel JM, et
al.Web-based smoking-cessation programs: results of a randomized trial. Am J
Prev Med 2008; 34: 373-81.
35. Swan GE, McClure JB, Jack LM, Zbikowski SM, Javitz HS, Catz SL, et
al.Behavioral counseling and varenicline treatment for smoking cessation. Am
Prev Med 2010; 38: 482-90.
36. An LC, Klatt C, Perry CL, Lein EB, Hennrikus DJ, Pallonen UE, et al.The RealU
online cessation intervention for college smokers: a randomized controlled trial.
Prev Med 2008; 47: 194-99.
37. Woodruff SI, Conway TL, Edwards CC, Elliott SP, Crittenden J. Evaluation of an
Internet virtual world chat room for adolescent smoking cessation. Addict
Behav 2007; 32: 1769-86.
38. Mermelstein R, Turner L. Web-based support as an adjunct to group-based
smoking cessation for adolescents. Nicot Tobac Res 2006; 8 Suppl 1: S69-76.
39. Patten CA, Croghan IT, Meis TM, Decker PA, Pingree S, Colligan RC, et
al.Randomized clinical trial of an Internet-based versus brief office intervention for
adolescent smoking cessation. Patient Educ Couns 2006; 64: 249-58.
40. Etter JF. Comparing the efficacy of two Internet-based, computer-tailored
smoking cessation programs: a randomized trial. J Med Intern Res 2005; 7: e2.
426
41. Te Poel F, Bolman C, Reubsaet A, de Vries H. Efficacy of single computertailored e-mail for smoking cessation: results after 6 months. Health Educ Res
2009; 24: 930-40.
42. Brendryen H, Drozd F, Kraft P. A digital smoking cessation program delivered
through internet and cell phone without nicotine replacement (Happy Ending):
Randomized controlled trial. J Med Intern Res 2008; 10: e51.
43. Brendryen H, Kraft P. Happy Ending: a randomized controlled trial of a digital
multi-media smoking cessation intervention. Addiction 2008; 103: 478-84.
44. Stoddard JL, Augustson EM, Moser RP. Effect of adding a virtual community
(bulletin board) to smokefree.gov: randomized controlled trial. J Med Intern Res
2008; 10: e53.
45. McKay HG, Danaher BG, Seeley JR, Lichtenstein E, Gau JM. Comparing two
web-based smoking cessation programs: randomized controlled trial. J Med
Intern Res 2008; 10: e40.
46. Rabius V, Pike KJ, Wiatrek D, McAlister AL. Comparing internet assistance for
smoking cessation: 13-month follow-up of a six-arm randomized controlled trial. J
Med Intern Res 2008; 10: e45.
47. Lenner L, Munoz RE, Stoddard J, Delucchi K, Bansod A, Skoczen S, et.al Design
and pilot evaluation of an internet smoking cessation program. J Am Med Inform
Assoc 2003; 10: 16-20.
48. Munoz RF, Barrere AZ, Delucchi K, Penilla C, Torres LD, Perez-Stable EJ.
International Spanish/Inglish Internet smoking cessation trial yields 20%
abstinence rates at 1 year. Nicot Tobac Res 2009; 11: 1025-34.
49. Hughes JR, Stead LF, Lancaster T. Antidepressants for smoking cessation.
Cochrane
Database
Syst
Rev
2007;
Issue
1:
10.1002/14651858.CD000031.pub3.
50. Stead LF, Perera R, Bullen C, MMant D, Lancaster T. Nicotine replacement
therapy for smoking sessation. Cochrane Database Syst Rev 2008; Issue 1:
10.1002/14651858.CD000146.pub3].
51. SRNT Subcommittee on Biochemical Verification. Outcome criteria in smoking
cessation trials proposal for a common standard. Nicot Tobac Res 2002; 4: 149159.
52. West R, Hajek P, Stead L, Stapleton J, Outcome criteria in smoking cessation
trials: proposal for a common standard. Addiction 2005; 100: 299-303.
53. Feil EG, Noell J, Lichtenstein E, Boles SM, McKay HG. Evaluation of an Internetbased smoking cessation program: Lessons learned from a pilot study. Nicot
Tobac Res2003; 5: 189-94.
54. Cobb NK, Graham AL, Bock BC, Papandonatos G, Abrams DB. Initial evaluation
of a real-world Internet smoking cessation system. Nicot Tobac Res 2005; 7: 20716.
55. Hall SM, Delucchi KL, Velicer WF, Kahler CW, Ranger-Moore J, Hedeker D, et al.
Statistical analysis of randomized trials in tobacco treatment: Longitudinal
designs with dichotomous outcome. Nicot Tobac Res 2001; 3: 193-201.
56. Eusenbach G. Issues in evaluating health websites in an Internet-based
controlled trial. J Med Internet Res 2002; 4: e17.
57. Pike KJ, Rabius V, McAlister A, Geiger A. American Cancer Society’s Quit Link:
randomized trial of Internet assistance. Nicot Tobac Res 2007; 9: 415-20.
58. Prochaska JO, Butterworth S, Redding CA, Burden V, Perrin N, Leo M, et al.
Initial efficacy of MI, TTM tailoring and HRI’s with multiply behaviours for
employee health promotion. Prev Med 2008; 46: 226-31.
59. Niaura R, Abrams DB. Smoking cessation: progress, priorities, and prospects. J
Consult Clin Psychol 2002; 70: 494-509.
60. Haukkala A, Uutela A, Vartiainen E, McCalister A, Knekt P. Depression and
smoking cessation: the role, motivation and self-efficacy. Addict Behav 2000; 25:
494-509.
427
Chapter 16
Human factors and human factors engineering
Summary
•
The study of human factors is the science of human behaviour and
performance. Human factors engineering is the application of human factors
principles and insights for the purpose of matching people’s technological,
organisational and social environments to their goals, abilities and needs.
•
Human factors issues relevant for eHealth include: users’ work practices and
workflow; the nature of the tasks to be supported by the eHealth system;
users’ capabilities and skills; training programmes; and the wider work and
organisational context in which the system will be deployed.
•
Healthcare has been slow to incorporate understanding of human factors into
eHealth projects, despite the increasing dependency on eHealth applications
for delivery of care and growing evidence of patients being put at risk.
•
Users should be involved in all stages of design, development and deployment
of eHealth applications. Feedback from users should not only be facilitated,
but must also be actively encouraged to ensure that new applications are fitfor-purpose and to minimise risks to patient safety.
•
Ease of use (“usability”) and fit with working practices are as important as the
functionality and reliability of eHealth applications such as the electronic health
record, computerised decision support and ePrescribing systems. Confusion
and frustration arising from poor usability will interfere with user acceptance,
with an adverse subsequent knock-on effect on implementation; it may also
endanger the safety of patient care.
•
Meaningful integration of human factors’ considerations is only likely to be
achievable if eHealth projects are approached as a co-production of
healthcare and IT professionals, such that users are given the opportunity to
be more fully involved throughout the design and development phase.
•
Embedding human factors engineering principles and thinking is not free;
policymakers and commissioners need to ensure that adequate time,
resources and
prioritisation are given to this so as to maximise the chances
of success of eHealth initiatives.
428
16.1 Introduction
Despite widespread acknowledgement that involvement of healthcare
professionals in the design and development of new medical informatics
systems is crucial to their successful implementation and adoption, this is
often still not given sufficient attention by system commissioners and
Information Technology (IT) professionals. Drawing on insights from human
factors’ engineering (HFE), usability research and participatory design, this
chapter seeks to provide an overview of current thinking on user involvement
in IT projects and the challenges of achieving it effectively in the context of
eHealth. We illustrate these challenges with examples of usability issues
arising in the design and implementation of a range of eHealth applications,
primarily the electronic health record (EHR) and its various associated
applications such as computerised decision support systems (CDSS) and
ePrescribing. We conclude with suggestions for how the involvement of
healthcare professionals in eHealth projects may be strengthened and made
more effective.
In the most general sense, human factors (HF) (also sometimes known as
ergonomics) is the multi-disciplinary study of how people’s performance is
influenced by their technological, organisational and social environments.1
Human factors’ engineering is the methodical application of HF principles to
the design of these environments. The goal is to ensure the safe, comfortable,
and effective use of technologies.2 To achieve this, HF set out to, in the words
of Wilson,
‘… understand people and their interactions, as well as the relationships between
these interactions, and to improve those interactions in real-life settings.3
Though the start of HF pre-dates information technology (IT) by many
decades, it is now most closely associated with the design of IT-based
systems. As IT has become more widely used, so too has the scope of HF
widened, from an initial, narrow focus on understanding the physical,
429
perceptual and cognitive aspects of human performance, to a broader agenda
that includes the social and organisational contexts in which the work that an
IT system is to support is embedded.4 The aim of HFE is to apply HF
principles to achieve a good fit between an IT system and its users’
capabilities and needs. The term “usability” is often used as a short-hand for
the measure of the quality of this fit.
The broad areas of interest and scope of HFE practice today are outlined in
Box 16.1.
Box 16.1 The scope of human factors
Physical and perceptual factors: These include the bounds of human
performance such as the accuracy of movements such as pointing, response
times to simple stimuli and capacity to discriminate between levels of
brightness, different colours and their implications for the design of the
physical interface.
Cognitive factors: These include human performance relating to the speed
of information processing and decision-making, recognition of and memory for
information, the causes of errors, learning times and styles. The aim is to
inform guidelines for design of user interface layouts, the representation of
information, the sequencing of activities and measures to reduce the
likelihood and impact of errors.
Motivational factors: These focus on people’s attitudes, and beliefs and
expectations of technologies and how these may be influenced by a person’s
status, role, profession, age, etc. One important aim is to try to arrive at an
appropriate allocation of function between user and system such that the
former is able to derive satisfaction from using the system while ensuring its
safe and effective use.
Situational factors: These describe the social and organisational context
within which the individual is expected to perform. They include how roles and
divisions of labour are managed within a group or team, the collaborative
dimensions of the work, how awareness and coordination is achieved and the
implications of new technologies and work re-design for safety and reliability
of the overall socio-technical system.
430
While technical reliability is an essential requirement for any eHealth
application, successful deployment, adoption and dependability in day-to-day
use draws on a much wider range of insights and disciplines.1
2
It is the
recognition of the socio-technical character of this challenge that has
motivated the incorporation of HFE within IT systems design and
development. HFE is of critical importance to eHealth, not only to reduce the
risks that poor design may pose to patient safety, but also to reduce the risk
that new IT systems may be rejected by healthcare professionals.1 5 Studies
confirm that when HFE is applied early and consistently throughout an IT
project, it can greatly increase the chances of higher productivity and process
improvements. It can also provide the foundations for planning an effective
deployment strategy, which may over time decrease staffing and training
costs and reduce the risks of user resistance to the new system.1
In the case of equipment such as medical devices and surgical tools, where
attention to safety is reinforced through standards and regulation,6 evidence
of the use of HFE is generally encouraging.7 Influenced by high profile cases
such as the Therac-25 accidents,8 it is now widely understood that errors in
the use of medical devices often have their origins in poorly designed user
interfaces; and that design-induced errors can lead to patient injuries and
deaths.9
User interfaces that are misleading or illogical can induce errors even by the
most skilled users.9 Good user interface design requires more than the
following of HF guidelines on use of colour and screen layout, however. It is
equally important to have a detailed understanding of how a device will be
used and the work environment to gauge the types of errors that may arise in
use and thereby to be able to design to eliminate or, at least, mitigate their
impact.6
Beyond the medical device arena, which is to say beyond a focus on physical,
perceptual and cognitive aspects of usability, the impact of HFE on the design
of IT systems has been less evident, despite the increasing reliance by
431
healthcare providers on IT for the management and delivery of healthcare.10
As IT systems become ever more deeply embedded in healthcare, the
challenges that HFE will need to address can only grow in complexity.11
Inadequately conceived objectives, poor design and lack of change
management planning continue to lead to wasted investment in new IT
systems that are not fit for purpose or to rejection by their prospective endusers;12 for example, the London Ambulance Computer Aided Dispatch
System failure,13 where patients’ lives were put at risk (see Box 16.2).
Box 16.2 The Ambulance Computer Aided Dispatch System failure
The London Ambulance Service Computer Assisted Dispatch (LASCAD)
system
was
intended
to
replace
a
manual
system
and
improve
communication, location and dispatch of vehicles to improve the timeliness of
medical treatment. The system went live on 26th October 1992, covering all of
London and stopping use of the manual backup. Although on the first two
days all functionality was provided, some response times were less than
adequate, and manual intervention to correct problems was difficult. One of
the direct results of the failure of the system included an ambulance arriving
after the patient had died and had already been taken away by the
undertaker. Estimates of the total number of fatalities caused differ from 10
upwards, even through no coronial findings included the late arrival of an
ambulance as the direct cause. The investigation blamed a wide range of
factors, including technical, managerial, human and environmental issues.
Some blame was placed upon incomplete software, and inadequate testing
(particularly the lack of adequate load testing) and optimisation of the system
the use of tools meant for prototyping and not for safety-critical systems. The
human factors aspects included that staff had little or no confidence in the
system and had not been trained in its operation. Managerial issues included
the lack of change as a result of earlier problems.
In this chapter, we provide an overview of the basic principles of usability. We
then consider different approaches to the problem of setting usability goals
432
and determining if they are being met. The failure of many healthcare IT
projects to deliver their promised benefits leads us to examine what we argue
is the central question that this failure raises: how to involve healthcare
professionals effectively in IT systems design and development processes.
16.3 Usability
16.3.1 Usability principles
At its most basic level, usability is concerned with ensuring that IT system
design is compatible with users’ physical, perceptual and cognitive
capacities.9 Beyond these basic usability goals lies an increasingly complex
and interrelated set of concerns and HF orientations such that the pursuit of
acceptable levels of usability tends to resist a simple guideline-based
approach. In reality, the usability of an IT system is a complex, multidimensional problem that is not well-bounded. Usability cannot be defined
simply by human performance parameters, but must be set in a context in
which issues such as user preferences, working practices and environment
are taken into consideration. A design response to this broader agenda is to
give users some scope to customise systems so that that they can plan and
select actions according to their individual preferences or requirements. An
example would be to allow users to choose the level of drug safety alerts in
ePrescribing. This will enable users to familiarise themselves with and cope
with so-called “spurious alerts” and associated “alert fatigue”, problems that
had not been anticipated when these systems were first introduced (see
Chapter 10).14 There may, however, be a tension between user choice and
safety. In the case of the ePrescribing example, high customisability may lead
to potentially serious prescribing mistakes and omissions (discussed below in
evaluating performance and safety) and it is for this reason that some
software developers, based on the legal advice they have received, have
restricted the degree of customisation they allow.
It will be clear from the above that design decisions that may influence
usability should therefore be subjected to a thorough evaluation process.
Usability evaluation attempts not only to identify problems, but also their
causes so that they can be eliminated. A number of different usability
433
evaluation techniques are outlined in Table 16.1 below. These techniques can
be used both ‘formatively’ (i.e. to identify usability requirements) and
‘summatively’ (i.e. to verify that usability goals have been met):
Table 16.1 Comparison of usability evaluation techniques
Method
Advantages
Usable early in the
Analytical
design process
Few resources
Heuristic
Walkthrough
required
Strongly diagnostic
Overview of whole
High potential return
Disadvantages
Narrow focus
Broad assumptions
of users’ cognitive
behaviour
Problems getting
experts
Cannot capture real
behaviour
Rich data yields in-
Observational
Ethnography
Data collection and
depth understanding
analysis can be
Capable of detecting
time-consuming
subtle issues
Natural, high validity
of results
Survey
Can be used on
Low response rates
large groups
Possible interviewer
Addresses opinions
Questionnaire
Interviews
and understanding
bias
Interviews time
consuming
Laboratory
Powerful
studies
Quantitative data
High reproducibility
High resource
demands
Artificial,
questionable validity
of results
Based on Dix et al. (2003)
43
Permission to reproduce has been applied for
434
16.3.2 Evaluating usability and safety
Improving patient safety, as in the case of England’s NHS Care Records
Service (NHS CRS), is often given as a key objective of eHealth applications
(see Chapters 3 and 18). As Reason has noted, however, ‘IT does not
eliminate error, it relocates it and can also change its form.’15 Clearly, such
unintended and hence unanticipated consequences of the introduction of IT
must be taken seriously, especially if eHealth applications become more
closely integrated with work practices, workflows and organisational goals.
Where the quality and safety of care may be adversely affected, then usability
evaluation methods must be capable of assessing the impact of a new IT
system in a clinically meaningful way. This, as, amongst others, Heathfield
and Wyatt,16 Heathfield and Buchan,17 Heathfield et al.,18 Kaplan19 and
Sharit20 have argued, remains a challenging area for eHealth.
Healthcare has developed the clinical trial to measure the efficacy and safety
of new medicines and it is a methodology that also has its place in the
evaluation of eHealth applications (see Chapter 20). However, CDSSs
provide a good example of how relying on clinical trials alone can fail to
identify how eHealth applications are actually used in practice and the
consequences of that use (see Chapters 8 and 10). For example, in a review
of clinician responses to drug safety alerts in ePrescribing, van der Sijs et al.
found that with the exception of serious alerts for overdose, safety alerts were
over-ridden in 49-96% of cases. The reasons for over-riding alerts included:21
•
Alert fatigue: this was the most important reason for overriding,
caused by too many false positive alerts affecting clinicians’ judgement.
•
Disagreements: professional disagreement with the recommendations
being made: clinicians’ faith in their own knowledge or other
information sources obtained, incorrect information, patients’ resistance
to drug change.
•
Poor presentation: alerts were too long and difficult to interpret and
the clinical consequences of over-riding them were not clear.
435
•
Lack of time: insufficient time to pay adequate attention to the
messages being generated and unnecessary workflow interruptions.
•
Knowledge gaps: lack of understanding about importance of the alert.
Similarly, a review by Eslami et al. on the appropriateness of alerts, assessed
the impact of ePrescribing on the produced, accepted and ignored, alerts from
two points of view: system weakness and user response.22 The same review
reported that most of the alerts (55-91%) were ignored by the clinicians, with
perceived “clinical irrelevance” being the main reason for over-riding alerts.
The requirement for measurement and for repeatability favours an approach
to usability evaluation that is able to control and isolate assumed key factors.
But the issue is whether the results have ecological validity–that is, do they
reflect the outcomes that will be found in real situations of use? This question
becomes especially important where the introduction of a new eHealth
application is likely (either by design or as an unanticipated adverse effect) to
lead to changes in working practices. Alberdi et al. have sought to address
this issue by using a multi-disciplinary approach combining quantitative data
from controlled clinical trials with qualitative data from ethnographic (i.e.
observational) studies of clinical decision-making.23 Applying this approach to
an evaluation of a decision support tool for mammographic screening, Alberdi
et al.’s findings suggest that the ethnographic studies of how radiologists
actually used the CDSS in practice enabled a better understanding of the
sometimes
subtle
interactions
between
human
decision-making
and
computer-generated evidence.
16.4: User involvement in eHealth projects
The challenges in setting usability goals and evaluating whether they have
been met raises fundamental questions about the nature and scope for the
meaningful and effective involvement of healthcare professionals in the
design
of
eHealth
applications,
and
throughout
the
development,
implementation and adoption process. This is an issue on which HFE has
increasingly focused, but experience to date confirms that this is not easily
436
achieved. The practical issue is whether it is sufficient just to get user input at
the beginning (i.e. formatively) and at the end (i.e. summatively) of a project.
Based on evidence such as that above, we would argue that user involvement
needs to be taken considerably further, both in terms of scale and scope.
Thanks in part to the efforts of the participatory design community,24 the case
for user involvement in IT projects is now accepted as standard software
engineering practice.25 In this time, a number of recurring themes have
emerged as evidence of the challenges of achieving this effectively in practice
while, at the same time, questioning whether current practice goes far
enough.
Firstly, as has frequently been observed by IT professionals, users often “do
not know what they want”. Not surprisingly, this problem seems to be
especially common where the aim of introducing a new system is to facilitate
major changes in work practices. This, of course, is often a goal of eHealth
projects, with IT often being used explicitly as a change agent as is the case
with the NHS Connecting for Health’s (NHS CFH) programme.26 In such
cases, while high level requirements may be easy to identify, the details are
likely to prove much more difficult to define and may be based on an
oversimplified and abstract model of the work practices involved. Actual work
processes are typically much more complex and even quite different from the
procedures documented in organisational manuals.27
28
Clarke et al., for
example, studied how hospital managers monitor the availability of beds and
how this relies on various important, but often taken-for-granted aspects of the
workplace whose significance is often overlooked when IT systems are
designed.28
29
Similarly, the lack of attention paid by IT professionals to the
real world of work–for example, to how clinical information is actually recorded
and used–has afflicted EHRs and other healthcare information management
and integration projects for many years.30 31 Healthcare professionals, in their
turn, often have limited knowledge of the technical possibilities and may find it
difficult to re-conceptualise what they do in ways that are compatible with
what is technically feasible.
437
We argue that this lack of a common language and understanding between IT
professionals and healthcare professionals is a consequence of the “silos” in
which they operate. Dealing with this calls for radical changes in the ways in
which users and IT professionals work together.32 IT professionals need to
make greater efforts to familiarise themselves with users’ work context and
practices. One way of doing this that has found increasing favour is to
undertake detailed, observational studies of users’ working practices right
from the outset of an IT project.33
34
Equally, it is essential that users have
sufficient time and opportunities to experiment with the system and learn as
the project unfolds which highlights the need for IT projects to support users
with prototyping and iteration, and the benefits of staging implementation, by,
for example, piloting a small part of the overall system. Even where changes
in working practices are not intended, the introduction of a new system can
still have unanticipated consequences, sometimes because of poor design
decisions, on other occasions because, over time, its users discover more
effective ways of using the system.35 Proactively eliciting details of these
workarounds from end-users is, we argue, important to inform further
iterations of the application.
When considering how to improve user involvement in a project, there is no
single best technique, nor are techniques mutually exclusive; rather, different
techniques may deliver the greatest benefit when used in combination. What
there is rather less agreement on, however, is which of these techniques for
user involvement are necessary or sufficient for a given project and how this
user
involvement
should
be
scheduled
and
managed.
For
project
management, the concern is typically to achieve a balance: ensuring, on the
one hand, that user input is adequate for the purposes of establishing and
(tracking, possibly changing) user requirements while, on the other,
preventing the project being thrown off schedule by a seemingly never ending
series of demands for changes.
The study by Martin et al. provides a detailed account of project management
issues in the context of an EHR project and why ensuring systematic and
effective user involvement is challenging.36 Martin et al. note, on the one
438
hand, the concerns of the project team that user participation may become
unmanageable. Diverse and possibly conflicting opinions about requirements
are almost inevitable in large-scale projects such as the implementation of the
NHS CRS, where the range of users is necessarily very broad. On the other
hand, Martin et al. point to the difficulties of securing the commitment of users
to give their time to the project. User involvement in project work is likely to be
discretionary and such work may be taken on voluntarily. Users may
experience difficulty in maintaining commitment to a project because the
benefits may seem remote and intangible. Even if commitment is strong at the
beginning, it may subsequently wane if there is a perceived lack of progress
and/or difficulties in using the new system. Unless such issues are addressed,
the inevitable result is that what is achievable in terms of user participation “in
the wild” is almost always less than what the HFE guidelines call for. Jenkings
has addressed a number of these issues in his study of another EHR
project.37
One reason why the notion of what constitutes “best-practice” for user
involvement in IT projects continues to be elusive is because the problem it is
trying to address keeps changing and becoming more complex.38 To put it
simply, as IT systems become progressively more deeply embedded within
workplaces and organisations, and are increasingly seen (rightly or wrongly)
as vehicles for innovating work practices, then uncertainty about what users’
requirements really are grows. This is not just because there may be more
and different kinds of users (“stakeholders”) with whom application developers
must deal and who may have different–and possibly conflicting–interests
(although this is, of course, an important factor). Perhaps more important,
however, is that the likelihood that requirements will be difficult to establish a
priori, but will change when the system is deployed and users get to use it “for
real”. Furthermore, innovation often has unpredictable consequences. For
eHealth applications, it is imperative that unforeseen consequences for
patient safety are quickly detected and their causes resolved. The challenge
is how to evolve applications so that an adequate fit with work practices is
maintained. As Suchman argues, ‘…system function and human work
processes must be addressed together.’33
439
As responses to these problems, new ideas are beginning to emerge that
treat user involvement more seriously. So-called “agile” methods for IT
systems design and development follow an incremental planning approach
that allows changes to be made according to evolving user requirements and
have a commitment to ongoing design throughout the software lifecycle.39 In
these ways, agile methods provide an opportunity for IT projects to be
undertaken more as a partnership or “co-production” between users and IT
professionals in which the boundaries between system design, development
and use are broken down.32
38 40
In a similar vein, and drawing on lessons
from the NHS CRS, Eason has argued that the pace of project
implementation must be geared more strongly to users’ capacity to
accommodate and contribute meaningfully, supported by a technical
approach that emphasises building upon what is already in place with discrete
and phased implementations.41
16.5 Conclusions
We have argued that HF considerations are of central importance to the
achievement of effective and safe implementation goals in healthcare
systems. There can be little doubt, for example, that the lack of attention to
HF issues hitherto in eHealth applications such as ePrescribing and the EHR
has limited their effectiveness and adoption and, quite possibly, posed a risk
to patient safety.
In this chapter, while noting the breadth of the HF agenda for IT systems
design, we have focused on questions that ensuring the usability of eHealth
applications raises for healthcare professionals’ involvement. Patient safety
demands that usability requirements and evaluation techniques be driven by
clinical agendas and are capable of providing results which are meaningful in
practice. It seems inconceivable to us that this can be achieved without close
and effective participation of healthcare professionals throughout the IT
project lifecycle. The argument that effective involvement of healthcare
professionals in design and development is imperative to ensure that eHealth
applications are fit for purpose is, we believe, irrefutable. It is also clear that
440
major problems remain to be tackled if the HFE agenda is to be integrated
effectively within IT system design and development practice.
Much progress in increasing the effectiveness of user involvement in IT
projects has already been achieved. However, in the face of increasingly
complex and multi-faceted eHealth applications–many of which aim to bring
about radical changes in complex work processes and which imply greater
integration and interdependencies between services and organisations–
healthcare and IT professionals must seek new and more productive ways of
working together if the benefits of eHealth are to be realised and patient
safety is not to be jeopardised: to put it another way, in eHealth the time has
come to stop treating users as “human factors” and, instead, to acknowledge
their contribution as “human actors”:42
‘Part of the problem resides in an implicit view of ordinary people which, if surfaced,
would seem to treat people as […] at best, simply sets of elementary processes or
“factors” that can be studied in isolation in the laboratory […] Understanding people
as “actors” in situations, with a set of skills and shared practices based upon work
experience with others, requires us to seek new ways of understanding the
relationship between people, technology, work requirements, and organisational
constraints in work settings.’42
We argue that this is only achievable if eHealth projects are approached as a
co-production of healthcare and IT professionals, such that users are given
the opportunity to be more fully involved throughout the design and
development phase. Moreover, such enhanced forms of user involvement
need to be carried through into the deployment phase so that the fit between
work practices and system functionality is maintained as these co-evolve in
use.
As, following the pattern now evident in other sectors, eHealth applications
become steadily more “organisationally embedded”, understanding the
organisational context assumes even greater importance for their successful
adoption. In the following chapter, we explore the organisation issues
441
surrounding the implementation and adoption of eHealth applications and we
then, in Chapter 18, seek to draw together these socio-techno-cultural
considerations through a detailed case study exploring the timely and
important challenge of introducing the NHS CRS into secondary care settings.
References
1. Federal Aviation Administration. Human Factors Engineering and Safety
Principles & Practices. FAA System Safety Handbook. 2000.17-1-17-16.
2. Saathoff A. Human factors considerations relevant to CPOE implementations. J
Healthc Inf Manag 2005; 19: 71-78.
3. Wilson JR. Fundamentals of ergonomics in theory and practice. Appl Ergon 2000;
31: 557-67.
4. Grudin J. The Computer Reaches Out: The historical continuity of the interface.
In: Proceedings of the ACM SIGCHI Conference on Human Factors in Computer
Systems (CHI), Seattle, Washington, 1990.
5. Higson G. Medical Device Safety: The Regulation of Medical Devices for Public
Health and Safety. Institute of Physics Publishing, 2002.
6. Kaye R, Crowley J. Guidance for Industry and FDA Premarket and Design
Control Reviewers. Medical Device Use-Safety: Incorporating Human Factors
Engineering into Risk Management. 2000.
7. Sawyer D. Do It By Design: An Introduction to Human Factors in Medical
Devices. 1996. US Department of Health and Human Services, Public Health
Service, Food and Drug Administration, Center for Devices and Radiological
Health.
8. Leveson
N.
Medical
devices:
The
Therac25.
Available
from:
http://sunnyday.mit.edu/papers/therac.pdf 1995. (last accessed 31/12/10)
9. Sawyer MA, Lim RB, Wong SY, Cirangle PT, Birkmire-Peters D, Sawyer MA et
al. Telemonitored laparoscopic cholecystectomy: a pilot study. Stud Health
Technol Inform 2000; 70: 302–08.
10. Barach P, Small S. How the NHS can improve safety and learning. BMJ 2000;
320: 1683–84.
11. Button G. A New Perspective on the Dependability of Software Systems. In:
Clarke K, Hardstone G, Rouncefield M, Sommerville I, editors. Trust in
Technology: a Socio- Technical Perspective. Dordrecht: Springer, 2006: ix–xxv.
12. Heeks R, Mundy D, Salazar A. Why Healthcare Information Systems Succeed or
Fail. Institute for development Policy and Management Working Paper Series No
9 University of Manchester, Manchester 1999.
13. Finkelstein A, Dowell J. A comedy of errors: the London Ambulance Service case
study.
Available
from:
http://people.cs.uct.ac.za/~gaz/teach/hons/papers/lascase.pdf (last accessed
31/12/10)
Manage 2003; 24: 29–38.
15. Reason J. Human Factors & Patient Safety: System Issues. Annual Conference
of the Care Record Development Board, November 2006. Available from:
www.connectingforhealth.nhs.uk/crdb/boardpapers/2006conference/workshop_i
mproving_patient_safety.ppt (last accessed 31/12/10)
16. Heathfield H, Wyatt J. Philosophies for the design and development of clinical
decision307 support systems. Methods Inf Med 1993; 32: 1–8.
17. Heathfield H, Buchan I. Current evaluations of information technology in health
care are often inadequate. BMJ 1996; 313: 1008.
442
18. Heathfield H, Pitty D, Hanka R. Evaluating information technology in health care:
barriers and challenges. BMJ 1998; 316: 1959–61.
19. Kaplan B. Evaluating informatics applications—some alternative approaches:
theory, social interactionism, and call for methodological pluralism. Int J Med
Inform 2001; 64: 39–56.
20. Sharit J. Perspectives on computer aiding in cognitive work domains: Toward
predictions of effectiveness and use. Ergonomics 2003; 46: 15.
21. van der Sijs H, Aarts J, Vulto A, Berg M. Overriding of drug safety alerts in
computerized physician order entry. J Am Med Inform Assoc 2006; 13: 138–47.
22. Eslami S, Abu-Hanna A, de Keizer NF. Evaluation of outpatient computerized
physician medication order entry systems: a systematic review. J Am Med Inform
Assoc 2007; 14: 400–06.
23. Alberdi E, Povyakalo A, Strigini L, Ayton P, Hartswood M, Procter R. Use of
computer-aided detection (CAD) tools in screening mammography: a
multidisciplinary investigation. Brit J Radiol 2005; 78: S31–S40.
24. Greenbaum J, Kyng M (eds.) Design at work: Cooperative Design of Computer
Systems. Hillsdale, NJ: Lawrence Erlbaum, 1991.
25. Sommerville I. Software Engineering. 8th ed. Harlow: Addison-Wesley, 2007.
26. NHS
Connecting
for
Health.
Available
from:
http://www.connectingforhealth.nhs.uk/ (last accessed 31/12/10)
27. Clarke K, Hughes J, Rouncefield M, Hemmings T. When a bed is not a bed:
calculation and calculability in complex organizational settings. In: Clarke K,
Hardstone G, Rouncefield M, Sommerville I, editors. Trust in Technology: a
Socio-Technical Perspective. Dordrecht: Springer, 2006: ix.
28. Pappas Y, Seale C. The opening phase of telemedicine consultations: an
analysis of interaction. Soc Sci Med 2009; 68: 1229-37.
29. Heath C, Luff P. Documents and professional practice: ‘bad’ organizational
reasons for ‘good’ clinical records. In Proceedings of CSCW’96 (Boston MA,
November), ACM Press, 1996: 354-62.
30. Hartswood M, Procter R, Rouncefield M, Slack R. Making a case in medical work:
implications for the electronic medical record. CSCW 2003; 12: 241–66.
31. Anderson S, Hardstone G, Procter R, Williams R. Supporting the Evolution of
Organizational Information Systems. In: Ackerman M, Erickson T, Halverson C,
Kellog W, editors. Evolving Information Artefacts. Springer, 2006.
32. Hartswood M, Procter R, Rouncefield M, Slack R, Voss A. Co-realization:
Evolving IT Artefacts by Design. In: Ackerman M, Erickson T, Halverson C,
Kellog W, editors. Evolving Information Artefacts. Springer, 2007.
33. Suchman L. Representations of work. Commun ACM 1995; 38: 33–34.
34. Hughes J, O’Brien J, Rodden T, Rouncefield M. Ethnography, communication
and support for design. In: Luff P, Hindmarsh J, Heath C, editors. Workplace
Studies: Recovering Work Practice and Informing System Design. Cambridge:
Cambridge University Press, 2000: 187–214.
35. Voss A, Hartswood M, Procter R, Slack R, Rouncefield M, Büscher M.
Configuring user-designer relations: Interdisciplinary perspectives. Springer,
2009.
36. Martin D, Mariani J, Rouncefield M. Practicalities of Participation: Stakeholder
involvement in an electronic patient records (EPR) project. In Voss A, Hartswood
M, Procter R, Slack R, Rouncefield M, Büscher M (Eds.) Configuring UserDesigner Relations: Interdisciplinary Perspectives, Springer, 2009.
37. Jenkings N. User-Designer Relations in Technology Production: The
Development and Evaluation of an ‘Animator’ tool to Facilitate User Involvement
in the Development of Electronic Health Records. In Voss A, Hartswood M,
Procter R, Slack R, Rouncefield M, Büscher M (Eds.) Configuring User-Designer
Relations: Interdisciplinary Perspectives, Springer, 2009.
443
38. Hartswood M, Procter R, Rouchy P, Rouncefield M, Slack R, Voss A. Working IT
Out in Medical Practice: IT Systems Design and Development as Co-Realisation.
Meth Inform Med 2003; 42: 392-97.
39. Beck K. Extreme Programming Explained: Embracing Change. Addison Wesley,
2000.
40. Catwell L, Sheikh A. Information technology (IT) system users must be allowed to
Prim Care 2009; 17: 1–4.
41. Eason K. Local sociotechnical system development in the NHS National
Programme for Information Technology. J Inform Technol 2007; 22: 257–64.
42. Bannon L. From human factors to human actors: The role of psychology and
human computer interaction studies in system design. In J. Greenbaum and M.
Kyng (eds.), Design at Work: Cooperative Design of Computer Systems.
Hillsdale, NJ: Lawrence Erlbaum, 1991.
43. Dix A, Finlay J, Aboud G, Beale R. Human-Computer Interaction. Prentice Hall.
UK, 2003.
444
Chapter 17
Importance of organisational issues in the
implementation and adoption of eHealth innovations
Summary
• The study of organisational issues as they pertain to eHealth
innovations is a multi-disciplinary field utilising bodies of knowledge
from organisational psychology, management and in particular change
management and human factors with clinical and information
technology expertise.
• There is a general consensus that organisational issues are at the root
of many of the problems with the implementation and adoption of
technological innovation in healthcare, particularly those innovations
that are likely to have a major discernible impact on care processes.
• While there is at present no overarching conceptual or theoretical
framework in relation to the implementation and adoption of eHealth
innovations, research consistently emphasises the importance of
technical, human and organisational factors, and the inter-relationships
between these.
• Early and ongoing user nvolvement, relative advantage of the
technology and early demonstrable benefits, a close fit with
organisational priorities and processes, training and support, and
effective leadership and change management seem to be particularly
important factors for success.
•
There is a need for research that moves beyond the existing largely
descriptive work to the development and testing of interventions
seeking to promote implementation and adoption.
•
This research is particularly important to undertake in contexts in which
organisational factors are likely to have a major impact on the success
or failure of the initiative–these include, for example, implementations
of technologies that are: likely to have a major clinical impact on care
delivery; “pushed” into the organisation rather than “pulled” by endusers and their respective bodies; “commercially procured” with little
445
opportunity for customisation when compared to “home-gown” products
with a greater sense of local ownership and buy-in; unreceptive
environments; and little opportunity for local “working out” of how these
systems can best be used in practice.
17.1 Introduction
There is now widespread acknowledgement that healthcare has, when
compared to other industries, been slow to adopt information technology (IT).
This in part results from the ethos of many healthcare professionals, who tend
to emphasise the importance of the more human dimensions of care,1 but
also the long history of failures associated with attempts at implementing IT
into health settings.2-4 This latter failing often relates to a failure to adequately
consider the organisational dimensions of technology, which can, depending
on the nature of the innovation and the way in which it is introduced into a
healthcare system, be substantial.5
The previous chapter focused on the need to take account of human factors’
onsiderations to ensure that the IT tools that are developed actually meet the
needs of clinicians and patients, and furthermore that those that are
developed are usable.
Building on these deliberations, we shift our focus in
this chapter to reviewing the wider social and organisational considerations
associated with IT deployment. The case study in the following chapter
(Chapter 19) brings together these important–but all too often neglected–
human and organisational dimensions in the context of a detailed reflection on
the implementation and adoption of the National Health Service Care Records
Service (NHS CRS).
We begin by discussing what we mean by organisational issues, what bodies
of knowledge contribute to their study, and their scope for use. We then
discuss the theoretical basis for addressing organisational issues when
implementing eHealth innovations, summarise the evidence-base and reflect
on its limitations. This is followed by discussion of the themes we have
identified as important in implementation to facilitate effective adoption and
446
integration, and the potential for assessing organisational readiness for
technological innovation in healthcare. Finally, we discuss the implications for
practice, policy and research.
17.2 Definitions, description and scope for use
17.2.1 Definitions
There are a number of related terms that are often used as synonyms or in a
less than clear way, which can make navigating and interpreting this body of
evidence somewhat confusing.
adoption,
deployment,
These include inter-related terms such as
diffusion,
implementation,
infusion,
integration,
normalisation and routinisation. The glossary provides detailed definitions of
these terms, but in essence they all relate to the processes by which
innovation is introduced and is then incorporated into care process such that
its use by professionals and/or patients becomes part of everyday practice.
Rather than these being straightforward linear processes, there is increasing
appreciation that the process of embedding technology into health systems is
complex and dynamic in nature involving a good deal of iteration and “working
out” as the technological, social and organisational dimensions gradually
accommodate each other (or not) over time.
17.2.2 Description
The study of organisational issues in relation to eHealth is not a clearly
defined field of study, but rather a problem-based approach centring on the
interaction between organisations or, more accurately, the people working
within these organisations and technology. The following bodies of knowledge
are potentially useful in contributing to the understanding of organisational
issues in the context of IT implementation and usage:
•
Human factors: An all-embracing term that covers: the science of
understanding the properties of human capability (human factors
science), the application of this understanding to the design and
development of innovations (human factors engineering), and the art of
ensuring successful application of human factors engineering to IT (see
Chapter 16).
447
•
Organisational/occupational/social
psychology:
A
subset
of
psychology that is concerned with the application of psychological
theories, research methods, and intervention strategies to workplace
issues, relevant topics include: personnel psychology (e.g. behaviour
and attitudes, changes in what jobs entail/working patterns and effects
on the individual); motivation and leadership; employee selection;
training and development; organisation development and guided
change; organisational behaviour; and work and family issues.6-16
•
Management
and,
in
particular,
organisational
change
management: A structured approach to change in individuals, teams,
organisations and societies that enables the transition from a current
state to a desired future state. Often focuses on increasing
organisational effectiveness and on identifying barriers and facilitators
to reaching the desired future state.17
•
Information systems: An academic discipline that is concerned with
the uses of information and information technology (IT) in organisations
and, more generally, society. This area has emerged from Systems
Theory, which assumes that the world consists of complex systems,
which have complex inter-relationships with each other and the world
at large. The defining feature here is that a system is viewed as being
more than the sum of its parts.18-22
Additionally, technological innovation in healthcare requires expertise in
technical IT considerations and clinical practice. These bodies of knowledge
contribute to the design, development and deployment of eHealth innovations.
17.2.3 The case for attending to organisational issues
An appropriate appreciation of the importance of human factors and
organisational considerations should help in ensuring that technological
innovations that are designed are not only useful and usable, but that they
also support the organisations or systems within which patients and/or
professionals operate. These innovations should therefore support care
provision as well as organisational functioning. Recognition of the fact that
change can however prove disruptive–particularly in the short- to medium448
term–means that steps should be taken to minimise potential adverse effects,
whilst at the same time, maximising the chances of successful integration with
workflow processes and organisational requirements. This is particularly
relevant in the context of eHealth applications that are likely to have an impact
on day-to-day clinical considerations such as electronic health record (EHR)
systems (see Chapter 6), rather than those which have more of an
infrastructure function such as the National Network for the NHS (N3) (see
Chapter 3). Technological innovations are often employed to enable
organisational change in healthcare–as was the explicit aim of the National
Programme for Information Technology (NPfIT) for example–in which case
attending to organisational issues is also important so as to maximise the
chances of organisations successfully achieving the desired outcomes. This is
particularly important in the context of NPfIT, where organisations have to fulfil
national demands (e.g. achievement of targets) whilst simultaneously having
to balance these with organisational demands (e.g. resources), social
demands
(e.g.
user
requirements)
and
technical
demands
(e.g.
interoperability and performance). Managing change effectively reduces the
potential for failure, but conversely a failure to give due regards to social and
cultural dimensions often proves to be a recipe for disaster.
There is potentially a wide range of theoretical considerations that can be
drawn on when conceptualising the interaction between technology, humans
and the organisations in which they function. Reviewing these in detail is
beyond the scope of this chapter, but here we introduce a few of the more
prominent and potentially most useful theories from the literature, which will,
where appropriate, be referred to when interpreting the empirical evidence:
•
Social Shaping of Technology (SST): Approaches studying the SST
highlight the importance of wider macro-environmental factors in
influencing technology and its implementation and adoption into
organisations.23 This has emerged as a response to studies focusing
on the social consequences of technology implementation, and in
doing so increasingly shifts the focus to viewing technology itself being
as shaped by social processes.23 These social processes are assumed
449
to consist of choices made by key stakeholders throughout the stages
of design, implementation and adoption.23 Different choices arising
from particular interests (which may be transitory or more stable) can
have different consequences for different stakeholders– examining
these is a central task in SST.23
•
Diffusion of innovations/Diffusion of Innovations in Health Service
Organisations: These are approaches that focus on how innovations
spread in organisations over time and how this spread can be
facilitated.24 25 Key components of these models include:
o A focus on end-user perceptions and innovation attributes
o Strategies on targeting/communicating with and engaging endusers
o Attention to both micro and macro contexts.
•
Socio-technical and sociotechnical changing: These approaches
conceptualise change as a non-linear, unpredictable and context
dependent process. They assume that both social and technical
dimensions shape each other over time in a complex and itself evolving
environment. Socio-technical systems have been defined as “any
interlinked, systems based mixture of people, technology and their
environment engaged in goal directed behaviour which, ideally, leads
to the emergence of productivity and well-being”.26 Socio-technical
approaches have emerged out of many examples of failed technology
implementations and the recognition that social factors are often
neglected when introducing technology into organisations.26-28 The aim
is to create a receptive organisation that pays attention to both
technical requirements as well as social and psychological needs of
workers in the changing organisational environment that results from
technology introduction.26 29
•
Technology Acceptance amongst Physicians (TAM): This assumes
that adoption/usage of a system is determined by the attitude towards
use and perceived usefulness ("the degree to which a person believes
that using a particular system would enhance his or her job
performance") and perceived ease of use ("the degree to which a
450
person believes that using a particular system would be free from
effort") of the application.13 30
•
Normalisation Process Model: This is a theoretical framework
describing how complex interventions in healthcare are routinely
incorporated into the day-to-day work of healthcare workers (or
“normalised”).31 The model highlights the importance of social
processes, and the organisational context in shaping outcomes.
Several factors are assumed to influence the normalisation of an
innovation:
o The effect of the intervention on interactions between people
and practices
o How interventions relate to existing knowledge and relationships
o The effect of the intervention on division of labour
o The way the intervention relates to the organisational context.
•
Sensemaking: It is assumed that here individuals in organisations
discover meanings of the status quo (frequently as a result of some
kind of change), often by transforming situations into words (expressed
in language or texts) and then displaying a resulting action as a
consequence of their interpretations.32 The underlying assumption is
that organisations are not existing entities as such, but are “talked into
action” or produced by sensemaking activities (and also the other way
around); the very way in which they are talked about defines their
existence.
•
The notion of “fit”: These models emphasise that one not only needs
to consider human, technological and work process factors in isolation,
but also the extent to which these align with each other.33 The better
the fit, the more likely the implementation is assumed to be successful
and the higher levels of adoption amongst users are likely to be. Fit can
be facilitated through management’s actions. For example, the fit
between users and technology can be improved by providing additional
training in how to use the software, whilst the fit between task and
technology can be improved by increasing the availability of computers
in the work environment. Several authors have used the notion of fit to
451
explain their findings, arguing that models relying on factors of success
fail to pay attention to dynamics between individual dimensions and the
way they affect each other.34-36 For example, Aarts and colleagues
explain an unsuccessful implementation of a computerised physician
order entry (CPOE) system in a Dutch hospital as being due to a lack
of fit between technological and environmental factors.34 They argue
that a perfectly good system is worth nothing if it is not used.
Our brief list is by no means exhaustive, but our intention is to illustrate the
range of different theoretical considerations in relation to the introduction of IT
in the healthcare context. Overall, the main tenants of these various
theoretical considerations seem to be: (1) a focus on relatively linear stages
and integration of technology over time, with some models focusing on
exploring one particular aspect of the lifecycle in detail; (2) a focus on
individual adopters; (3) a focus on complexity and unpredictability
characterising the change process; and (4) a mixture between the above with
models trying to be as inclusive as possible.
17.4 Evidence-base and its limitations
To date, much of the available primary literature concerning organisational
issues in relation to eHealth innovations is anecdotal and retrospective in
nature stemming from single organisation experiences of implementing a
specific eHealth application. These tend to be descriptive accounts, without
much in the way of attention to relevant theoretical considerations, which
makes drawing generalisable lessons from such reports difficult.37 Another
important feature of this literature is that there are as yet, for obvious reasons,
very few accounts of the experiences of implementing major programmebased IT innovations into healthcare systems of the scale and complexity of
those that are now being pursued in England and elsewhere.
We begin by highlighting key findings from systematic reviews (SRs) and then
turn to other key reports in the literature.
452
We found 13 SRs focusing on the issues of implementation and adoption of IT
in healthcare (see Appendices 4 and 5).38 38-49 We in addition identified three
SRs, which focused more generally on related questions of innovation in
healthcare settings.24
50 51
Also of relevance were a number of important
additional reports and reviews based on a combination of experience and
expert opinion.34 36 44 52-78
Overall, the evidence from the SRs draws attention to the importance of a
number of inter-related technical, human and organisational factors that can
help describe and explain potential contributing factors to success and failure
(or the perceptions of these)54 in attempts at implementing technology into
healthcare systems. We summarise below some of the recurring findings in
relation to these themes.
17.4.1 Technical features
It is now widely accepted that the majority of end-users are not technically
averse per se, but they are averse to technology that is seen as inadequate or
worse still interferes with their values, roles and aspirations.55 66
A key feature of the technology therefore, as demonstrated by the review of
101 studies of adoption by Gagnon et al,39 is that it needs to be useful and
offer relative advantages over existing ways of doing things. Particularly
important in this respect is that IT should preferably speed-up tasks, but
should at-least be as quick as the existing way of doing things. Such is the
importance of this that Bates et al, in their widely cited ‘Ten commandments’
review, refer to this as the first commandment.52 In their extension to the TAM
model, Yarbrough and Smith highlight the importance of the related perceived
ease of use, which can then shape attitudes and influence behavioural
intentions.46
Other features of technology that have found to be important determinants of
success includes the extent to which it fits and is interoperable with existing
technology in the organisation, costs, complexity and the extent to which it
has already been trialled and early problems have had an opportunity to be
453
ironed out.39 53 Given the constantly changing nature, leadership and priorities
of complex systems such as the NHS, it is also important that the technology
is able to evolve or to be adapted to support change in this way.34 56 62 72
17.4.2 Human factors
A number of human factors in relation to technology have been reviewed in
the previous chapter. Building on these, the themes that have been
highlighted as increasing the chances of successful implementation include
the IT literacy and general competencies of staff,57 personal attitudes towards
IT and also those of colleagues and patients,58 financial considerations,38 and
the extent to which the technology supports inter-professional roles and
working.42 Technologies which inadvertently undermine perceived social
standing or professional autonomy, for example, are less likely to be accepted
than those which support such functioning.59 Involvement of individuals at the
conception and early design stages, together with the opportunity for field
testing of early prototypes and open communication channels, should help to
identify key human dimensions that can help in implementing systems that are
likely to be valued and used by professionals and patients.60 61
17.4.3 Organisational factors
Larger, more complex health systems have, as predicted by Diffusion of
Innovations theory, proven particularly receptive to the introduction of
innovation.25
53
This is in part because of the human, organisational and
financial capital and “slack” that such organisations have when compared to
smaller organisations. These institutions tend however also to have more
complex management structures with greater degrees of hierarchy and as
such available evidence suggests that it is very important that senior
leadership and lead professionals or “champions” are signed-up to the vision
and have ownership of the implementation plans.34 36 44 56 60 62-70 73
74
These
champions often need to play roles as “boundary spanners”, bridging the gulfs
that often exist between and within IT staff, management and clinicians.24
They can also be very helpful in leading the redesign of workflows, providing
training and support to new users, highlighting issues, and trouble-shooting,
all of which are core considerations in the context of major IT deployments.39
454
The initial implementation is often disruptive as parallel systems are still in
operation and staff attempt to make sense of the new workflows.72 Making
time available to clinicians and administrative staff by, for example, proactively
reducing workload during this time period and/or introducing the technology
when there is not other major upheavals in the system has also been
highlighted as important.71
Strong leadership and management are also necessary to ensure that a
realistic assessment of the likely benefits and trade-offs is conveyed, the
anticipated timeframe to realise these, the avoidance of “scope creep”, that
interoperability considerations are addressed and that an appropriate
implementation approach is adopted–for example, a slow and incremental
“soft landing” or a one-off “big bang”.44 68 69 73-76 Leadership also needs to plan
for contingencies, including those that are extreme, such as the technology
failing, and ensuring that such planning pervades the system.79
Whilst much of the literature focuses on the importance of the “inner context”
of the organisation, additional wider contextual factors can also be important,
particularly, as discussed in the next chapter, in the context of governmentinitiated programmes such as the NPfIT.24
17.4.4 Ensuring “fit” between these technical, human and organisational
dimensions
As will be obvious from the discussions above, these three dimensions are
closely related. The key it seems is to ensure an adequate fit between these
considerations. This point is exemplified in the SR by Yusof et al. who, after
reviewing 55 studies using the Human, Organisation and Technology-fit
(HOT-fit) framework, concluded that “all three technology, human, and
organisational factors are equally important, in addition to the fit between
them.”
36 47
This alignment appears much easier to achieve in the case of
small-scale, organic, incrementally developed, “home grown” systems, in
contrast to the larger, more ambitious, eHealth projects that are now being
455
parachuted into complex environments according to highly pressurised
political timelines (see Chapters 3 and 18).80
There is furthermore a growing realisation that for a new technology to embed
itself in an organisation, there needs to be opportunities for the technology to
shape the organisation and also for those working in the organisation to
shape the technology–this therefore needs to be a reciprocal relationship in
which new ways of working are allowed to emerge.77 This is challenging for
those working using linear perspectives, but unless such experimentation and
re-invention is allowed, and indeed encouraged, then it may be that
technology never fulfils its potential.24 31 81 82 Although a time-consuming and
expensive process, attention to unanticipated consequences is important and
need to be studied, both in relation to those which may ultimately prove to be
advantageous and of course also those which may inadvertently increase the
risk of harm.69 78
17.4.5 A chronological perspective
The SR by Keshavjee et al. is interesting in that in addition to the technology,
people/organisational and process considerations highlighted above, it takes
a chronological perspective, focusing on the importance of pre-, during and
post-implementation considerations; this allows the interplay between these
different dimensions at different stages of the implementation journey to be
studied.43 Based on a critique of 55 reports of implementation of electronic
health record (EHR) systems, they identified how the focus of activity of
change management changes as implementation proceeds, beginning with,
for example, extensive stakeholder discussions when making key decisions
on software procurement to the creation and nurturing of user support groups
during the early post-implementation phase.43 83
Extending this system development lifecycle-based view, Gruber et al. found
that the availability or lack of end-user support during and in the aftermath of
“go live” was a particularly important determinant of success or failure.41
Catwell and Sheikh have also highlighted the need for continuous evaluation
cycles during the various stages from system conception to integration.84
456
17.5 Conclusions: Implications for practice, policy and research
Despite previous work, it is clear, as Lorenzi et al. note, that organisational
issues have not received due attention and they identify reasons why these
might be initially discounted.85 Organisational issues are experienced
subjectively and in different ways by different actors. As a result, they are
difficult to measure objectively, difficult to predict and are time consuming to
plan for. As a consequence, technical staff often responds by downplaying the
importance of organisational issues or by refusing to take responsibility for
them. A further factor is that dealing with organisational issues is often viewed
as delaying the more immediate implementation-related work.
Nonetheless, organisational issues are coming to the forefront of the eHealth
agenda due to a general consensus within the field that technological
innovation is not designed, developed nor deployed in a vacuum.
The
importance of organisational issues has important implications for practice,
policy and research.
Theories need ideally to describe, explain and predict. Findings from the
study of organisational issues have however yet to be prospectively applied
with any rigour. An ideal place to start with would be a model specific to
eHealth to test empirically (see Chapter 18). With enough empirical testing
and refining, development of best practice guidelines for implementation is
possible; for example, guidelines for involving users during the design and
implementation of large-scale systems. In contrast to the points made by
Lorenzi et al. on the unpredictability of innovations making strategies for
success too difficult to pin-down, Braude notes that ‘…predictability comes
from research. When a sufficient body of research in a field is available, it
becomes possible to predict outcomes based on prior experiences. Research
into people and organisational issues surrounding implementation of
innovations is equally scattered throughout the literature of different
disciplines with the amount of research is still relatively small.’86
457
The numerous disciplines or bodies of knowledge which contribute to the
study of organisational issues are rich in potential to facilitate implementation
and adoption of innovations in an ever increasingly complex health services
system.87 With that in mind, research employing expertise in these fields is
central to furthering knowledge on organisational adoption and best practices
for implementation. This prospective work needs to focus in particular on the
contexts in which, based on current models, there is the highest risk of
problems. Building on the themes introduced in this and the preceding
chapter, we in the following chapter focus on human and organisational
considerations in the context of a detailed case study on the introduction of
the NHS’ CRS. This chapter will also consider the importance of wider
contextual factors in impacting on the issues discussed in above.
References
1. Stevenson JE, Nilsson GC, Petersson GI, Johansson PE. NursesGÇÖ
experience of using electronic patient records in everyday practice in
acute/inpatient ward settings: A literature review. Health Informatics Journal
2010; 16(1):63-72.
2. Massaro TA. Introducing physician order entry at a major academic medical
center: 1. Impact on organizational culture and behaviour. Acad Med 1993;
68:20-25.
3. Sicotte C, Denis JL, Lehoux P. The Computer Based Patient Record: A
Strategic Issue in Process Innovation. Journal of Medical Systems 1998;
22(6):431-443.
4. Southon G, Sauer C, Dampney K. Lessons from a failed information systems
initiative: issues for complex organisations. Int J Med Inform 1999; 55(1):3346.
5. Bali RK, Wickramasinghe N. Achieving successful EPR implementation with
the penta-stage model. International Journal of Healthcare Technology and
Management 2008; 9(1):97-105.
6. Symon G, Cassell C, Dickson R. Qualitative Methods in Organizational
Research and Practice. Psychology Press, 2001.
7. Symon G, Cassell C. Neglected perspectives in work and organizational
psychology. Journal of Occupational and Organizational Psychology 2006;
79:307-314.
8. Tharenou P. The relevance of industrial and organisational psychology to
contemporary organisations: How far have we come and what needs to be
done post-2000? Australian Psychologist 2001; 36(3):200-210.
9. Cassell C, Symon G. Essential Guide to Qualitative Methods in
Organizational Research. Sage, 2004.
10. Shavinina L. The International Handbook on Innovation. Elsevier, 2003.
458
11. Silvester J, Anderson N, Patterson F. Organizational culture change: An intergroup attributional analysis. Journal of Occupational and Organizational
Psychology 1999; 72(1):1-23.
12. Korpela M, Mursu A, Abimbola H. Two Times Four Integrative Levels of
Analysis: A Framework. 2001. Proceedings of the IFIP TC8/WG8.2 Working
Conference on Realigning Research and Practice in Information Systems
Development: The Social and Organizational Perspective, Kluwer.
13. Taylor S, Todd P. Understanding information technology usage: A test of
competing models. Information Systems Research 1995; 6(2):144-176.
14. Davis FD, Bagozzi RP, Warshaw PA. User acceptance of computer
technology: a comparison of two theoretical models. Management Science
1989; 35(8):982-1003.
15. Karsten R, Schmidt D. Social Psychology for IT professionals: a proposed
model. Proceedings of the Academy of Information and Management
Sciences 2004; 8(1):37-41.
16. Bandura A. Self-efficacy: The exercise of control. New York: W.H. Freeman &
Company, 1997.
17. Cassell C, Symon G, Buehring A, Johnson P. The role and status of
qualitative methods in management research: an empirical account.
Management decision 2006; 44:290-303.
18. Moore Trauth E. Qualitative Research in IS: Issues and Trends. 2002. Idea
Group Inc (IGI).
19. Kaplan B, Maxwell A. Qualitative Research Methods for Evaluating Computer
Information Systems. Evaluating the Organizational Impact of Healthcare
Information Systems. Springer New York, 2006: 30-55.
20. Myers MD. Qualitative research in information systems. MIS Quarterly 1997;
21(2):241-242.
21. Dalcher D, Drevin L. Learning from information systems failures by using
narrative and ante-narrative methods. ACM International Conference
Proceeding Series 2003; 47.
22. Orlikowski W, Gash D. Technological frames: making sense of information
technology in organizations. ACM Transactions on Information Systems
(TOIS) 1994; 12(2):174.
23. Williams R, Edge D. The social shaping of technology. Research Policy 1996;
25(6):865-899.
24. Greenhalgh T, Robert G, Macfarlane F, Bate P, Kyriakidou O. Diffusion of
Innovations in Service Organizations: Systematic Review and
Recommendations. Milbank Q 2004; 82(4):581.
25. Rogers E. Diffusion of innovations. New York: Free Press, 1983.
26. Butson R. SocioTechnical Approach - STS. Available from:
https://russell.wiki.otago.ac.nz/SocioTechnical_Approach_-_STS
(last
accessed: 21/11/2008). 2008.
27. Mitchell VL, Nault BR. The emergence of functional knowledge in
sociotechnical systems. Haskayne School of Business, University of Calgary,
Calgary,
Alberta,
Canada.
Available
online
at:
http://www.haskayne.ucalgary.ca/haskaynefaculty/files/haskaynefaculty/emer
gence.pdf, accessed 31 October 2008. 2003.
28. Association for Information Systems. Theories used in IS research: sociotechnical
theory.
Available
from:
http://home.aisnet.org/displaycommon.cfm?an=1&subarticlenbr=242
(last
accessed: 21/11/2008). 2008.
29. Procter R, Williams R. HCI: Whose Problem Is It Anyway? 385-396. 1992.
Proceedings of the IFIP TC2/WG2.7 Working Conference on Engineering for
Human-Computer Interaction.
459
30. Davis FD. A Technology Acceptance Model for empirically testing new enduser infromation systems: theory and results. Doctoral Dissertation, Sloane
School of Management, Massachusetts Insitute of Technology. 1989.
31. May C. A rational model for assessing and evaluating complex interventions
in health care. BMC Health Serv Res 2006; 6:86.
32. Weick K, Sutcliffe K, Obstfeld D. Organizing and the Process of
Sensemaking. OrganizationalScience 2005; 16(4):409.
33. Ammenwerth E, Iller C, Mahler C. IT-adoption and the interaction of task,
technology and individuals: A fit framework and a case study. BMC Medical
Informatics and Decision Making 2006; 6:13.
34. Aarts J, Doorewaard H, Berg M. Understanding Implementation: The Case of
a Computerized Physician Order Entry System in a Large Dutch University
Medical Center. J Am Med Inform Assoc 2004; 11(3):207-216.
35. Pare G. Implementing clinical information systems: A multiple-case study
within a US hospital. Health Services Management Research 2002; 15(2):7192.
36. Yusof MM, Kuljis J, Papazafeiropoulou A, Stergioulas LK. An evaluation
framework for Health Information Systems: human, organization and
technology-fit factors (HOT-fit). International Journal of Medical Informatics
2008; 77(6):386-398.
37. Ovretveit J, Scott T, Rundall TG, Shortell SM, Brommels M. Improving quality
through effective implementation of information technology in healthcare.
International Journal for Quality in Health Care 2007;(5):259-266.
38. Boonstra A, Broekhuis M. Barriers to the acceptance of electronic medical
records by physicians From systematic review to taxonomy and interventions.
BMC Health Services Research 2010; 10:231.
39. Gagnon MP, Desmartis M, Labrecque M, Car J, Pagliari C, Pluye P et al.
Systematic Review of Factors Influencing the Adoption of Information and
Communication Technologies by Healthcare Professionals. Journal of
Medical Systems 2010;1-37.
40. Gagnon P, Pluye P, Desmartis M, Car J, Pagliari C, Labrecque M et al. A
systematic review of interventions promoting clinical information retrieval
technology (CIRT) adoption by healthcare professionals. International Journal
of Medical Informatics 79[10], 669-680. 1-10-2010.
41. Gruber D, Cummings G, LeBlanc L, Smith D. Factors Influencing Outcomes
of Clinical Information Systems Implementation: A Systematic Review.
Computers Informatics Nursing 2009; 27(3).
42. Gurses AP, Xiao Y. A Systematic Review of the Literature on Multidisciplinary
Rounds to Design Information Technology. J Am Med Inform Assoc 2005;
13(3):267-276.
43. Keshavjee K, Bosomworth J, Copen J, Lai J, Kucukyazici B, Lilani R et al.
Best practices in EMR implementation: a systematic review. AMIA Annual
Symposium Proceedings/AMIA Symposium 2006;982.
44. Ludwick DA, Doucette J. Adopting electronic medical records in primary care:
Lessons learned from health information systems implementation experience
in seven countries. International Journal of Medical Informatics 2009;
78(1):22-31.
45. Mair FS, May C, Finch T, Murray E, Anderson G, Sullivan F et al.
Understanding the implementation and integration of e-health services. J
Telemed Telecare 2007; 13(suppl_1):36-37.
46. Yarbrough AK, Smith TB. Technology acceptance among physicians: A new
take on TAM. Med Care Res Rev 2007; 64(6):650-672.
47. Yusof MM, Stergioulas L, Zugic J. Health information systems adoption:
findings from a systematic review. Stud Health Technol Inform 2007; 129(Pt
1):262-266.
460
48. Alexander G, Staggers N. A Systematic Review of the Designs of Clinical
Technology: Findings and Recommendations for Future Research. Advances
in Nursing Science 2009; 32(3).
49. Gruber D, Cummings G, LeBlanc L, Smith D. Factors Influencing Outcomes
of Clinical Information Systems Implementation: A Systematic Review.
Computers Informatics Nursing 2009; 27(3).
50. Robert G, Greenhalgh T, Macfarlane F, Peacock R. Adopting and assimilating
new non-pharmaceutical technologies into health care: a systematic review. J
Health Serv Res Policy 2010; 15(4):243-250.
51. Länsisalmi H, Kivimäki M, Aalto P, Ruoranen R. Innovation in Healthcare: A
Systematic Review of Recent Research. Nursing Science Quarterly 2006;
19(1):66-72.
52. Bates DW, Kuperman GJ, Wang S, Gandhi T, Kittler A, Volk L et al. Ten
commandments for effective clinical decision support: making the practice of
evidence-based medicine a reality. J Am Med Inform Assoc 2011; 10(6):523530.
53. England I, Stewart D, Walker S. Information technology adoption in health
care: when organisations and technology collide. Aust Health Rev 2000;
23(3):176-185.
54. Lorenzi NM, Riley RT. Organizational issues=change. International Journal of
Medical Informatics 69[2], 197-203. 1-3-2003.
55. Tonnesen AS, LeMaistre A, Tucker D. Electronic medical record
implementation barriers encountered during implementation. Proceedings
AMIA Annual Symposium 1999;624-626.
56. Doebbeling BN, Chou AF, Tierney WM. Priorities and strategies for the
implementation of integrated informatics and communications technology to
improve evidence-based practice. Journal of General Internal Medicine 2006;
21(SUPPL. 2):S50-S57.
57. Halamka J, Aranow M, Ascenzo C, Bates DW, Berry K, Debor G et al. EPrescribing Collaboration in Massachusetts: Early Experiences from Regional
Prescribing Projects. J Am Med Inform Assoc 2006; 13(3):239-244.
58. Garside P. Organisational context for quality: lessons from the fields of
organisational development and change management. Qual Health Care
1998; 7 Suppl:S8-15.
59. McDonald R, Waring J, Harrison S. At the Cutting Edge? Modernization and
Nostalgia in a Hospital Operating Theatre Department. Sociology 2006;
40(6):1097-1115.
60. Dagroso D, Williams PD, Chesney JD, Lee MM, Theoharis E, Enberg RN.
Implementation of an obstetrics EMR module: overcoming user
dissatisfaction. Journal of Healthcare Information Management 2007;
21(1):87-94.
61. Puffer MJ, Ferguson JA, Wright BC, Osborn J, Anshus AL, Cahill BP et al.
Partnering with clinical providers to enhance the efficiency of an EMR.
Journal of Healthcare Information Management 2007; 21(1):24-32.
62. Ash J, Berg M. Report of conference Track 4: socio-technical issues of HIS.
International Journal of Medical Informatics 2003; 69(2-3):305-306.
63. Austin CJ, Hornberger KD, Shmerling JE. Managing information resources: a
study of ten healthcare organizations. Journal of Healthcare Management
2000; 45(4):229-238.
64. Bossen C. Test the artefact-Develop the organization. The implementation of
an electronic medication plan. International Journal of Medical Informatics
2007; 76(1):13-21.
65. Fenton SH, Giannangelo K, Stanfill M. Essential people skills for EHR
implementation success. Journal of Ahima 2006; 77(6):60.
461
66. Greenhalgh T, Stramer K, Bratan T, Byrne E, Mohammad Y, Russell J.
Introduction of shared electronic records: multi-site case study using diffusion
of innovation theory. BMJ 2008; 337(oct23_1):a1786.
67. Lorenzi NM, Smith JB, Conner SR, Campion TR. The Success Factor Profile
for clinical computer innovation. Studies in Health Technology & Informatics
2004; 107:1077-1080.
68. Pare G, Sicotte C, Jaana M, Girouard D. Prioritizing the risk factors
influencing the success of clinical information system projects. A Delphi study
in Canada. Methods of Information in Medicine 2008; 47(3):251-259.
69. Sicotte C, Pare G, Moreault M-P, Paccioni A. A Risk Assessment of Two
Interorganizational Clinical Information Systems. J Am Med Inform Assoc
2006; 13(5):557-566.
70. Walker AR. Case study: leading change across two sites: introduction of a
new documentation system. Nursing leadership 2006;(4):34-40.
71. Sirkin HL, Keenan P, Jackson A. The hard side of change management. Harv
Bus Rev 2005; 83(10):108-18, 158.
72. Rose J, Jones M, Truex D. Scandinavian Journal of Information Systems.
Scandinavian Journal of Information Systems 2005; 17(1):133-152.
73. Duggan C. Implementation evaluation. HIM professionals share their
experiences bringing health IT online. Journal of Ahima 2006; 77(6):52-55.
74. Adler KG. How to successfully navigate your EHR implementation. Family
Practice Management 2007; 14(2):33-39.
75. Delany R. Smiling all the way. Healthcare Review Online 2004; 8(2).
76. Van Ginneken AM. The computerized patient record: balancing effort and
benefit. [Review] [104 refs]. International Journal of Medical Informatics 2002;
65(2):97-119.
77. Czarniawska B. A Golden Braid: Allport, Goffman, Weick. Organization
Studies 2006; 27(11):1661-1674.
78. Harrison MI, Koppel R, Bar-Lev S. Unintended Consequences of Information
Technologies in Health Care--An Interactive Sociotechnical Analysis. J Am
Med Inform Assoc 2009; 14(5):542-549.
79. Johnson C. Case Studies in the Failure of Healthcare Information Systems.
Available
from:
http://www.dcs.gla.ac.uk/~johnson/papers/AHRQ/case_study.pdf
(last
accessed 14/12/09). 2009.
80. Kreps D, Richardson H. IS Success and Failure—The Problem of Scale.
Political Quarterly 2007; 78(3):439-446.
81. Cornford T, Doukidis G, Forster D. Experience Outcome with a Structure,
Process and Framework for Evaluating an Information System. Omega, Int J
Mgmt Sci 1994; 22(5):491-504.
82. Eason K. Local sociotechnical system development in the NHS National
Programme for Information Technology. Journal of Information Technology
2007; 22(3):257-264.
83. Ash JS, Gorman PN, Lavelle M, Payne TH, Massaro TA, Frantz GL et al. A
cross-site qualitative study of physician order entry. J Am Med Inform Assoc
2003; 10(2):188-200.
84. Catwell L, Sheikh A. Evaluating eHealth interventions: the need for
continuous systemic evaluation. PLoS Med 2009; 6(8):e1000126.
85. Lorenzi NM, Riley RT. Organizational issues = change. Int J Med Inform
2003; 69(2-3):197-203.
86. Braude RM. People and organizational issues in health informatics. J Am
Med Inform Assoc 1997; 4(2):150-151.
87. Lorenzi NM, Riley RT, Blyth AJ, Southon G, Dixon BJ. Antecedents of the
people and organizational aspects of medical informatics: review of the
literature. J Am Med Inform Assoc 1997; 4(2):79-93.
462
Chapter 18
Case study: Lessons learnt from the literature
in relation to the design, implementation and adoption
of the NHS Care Record Service in secondary care
Summary
•
Many information technology innovations fail to realise their potential and
this unfortunately has also been true with respect to the history of eHealth
applications.
•
Major factors contributing to these failures—some of which are
spectacular—include the lack of appreciation and attention being paid to
socio-techno-cultural factors during product development and deployment.
•
There is a burgeoning change management literature, relating to the
implementation and adoption and the diffusion of innovations and
technology.
•
No one existing model in the literature seems to be comprehensive
enough to cover all aspects that need to be taken into account, when
considering the implementation and adoption of electronic health records
in the context of the National Programme for Information Technology.
•
Drawing on lessons from the existing literature, we have therefore
developed our own model and discuss this in relation to the introduction of
the NHS Care Record Service in secondary care.
•
Based on this, we make 28 recommendations that the Department of
Health may wish to consider when thinking about the future strategy of
implementing the NHS Care Record Service into English hospitals.
•
The main components of our model relate to technical, human/social,
organisational and macro-environmental dimensions – all inter-related and
situated within specific contexts and more or less prominent at particular
points in the implementation and adoption circle.
463
•
The main components of our model relate to technical, human/social,
organisational and macro-environmental dimensions–all inter-related and
situated within specific contexts and more or less prominent at particular
points in the implementation and adoption circle.
•
The Department of Health has already taken on board several key
lessons, mainly relating to implementation and adoption, but there is still a
need to address these more thoroughly in line with recent developments in
implementation strategy, the constantly evolving literature and lessons
learned from early implementations of electronic health record systems.
•
We argue that in this respect the most important components of our model
are a technology that is perceived as fit for purpose and used by endusers. Recognition of the importance of these issues may, to some extent,
have been lost due to the politically driven nature of the programme and its
large scale.
•
Given that the national implementation of NHS computer systems has
somewhat changed in strategic direction, it is important to reconsider the
evolving literature in this area and incorporate lessons into ongoing policy
deliberations.
18.1 Introduction
As discussed in Chapter 3, the National Health Service Care Record Service
(NHS CRS) represents the backbone of the National Programme for
Information Technology (NPfIT) and as such represents a potentially
fundamentally transformative and, conversely, also potentially very disruptive
eHealth innovation. Considering its centrality within the Programme, it is of
considerable importance that in the development and deployment of this
technology—which has the potential to yield great benefits to patients—
implementers are cognisant of the range of factors that are likely to have a
major impact on the acceptability and likely effectiveness of this innovation. It
is particularly important to reconsider these factors in the light of the evolving
literature base and the constantly evolving implementation strategy. In this
case study, drawing on the existing theoretical and empirical evidence, we
464
consider in detail potentially important socio-techno-cultural issues (see
Chapters 16 and 17) that may impact on the successful implementation and
adoption of NHS CRS in secondary care. In so doing, we hope to identify
possible strategies and approaches that implementers might wish to take into
account when considering the future strategic direction of the NHS CRS,
whilst at the same time contributing to the theoretical and relatively limited
empirical base for understanding IT adoption in healthcare.
We begin by briefly reviewing the nature, structure and the evolving
implementation and policy context of NHS CRS, before turning our attention
to theoretical considerations. We then introduce an inclusive theoretical model
of implementation and adoption, based on our ongoing review of the literature.
We then use this model to reflect on the current and potential future approach
to helping to ensure successful deployment of the NHS CRS, outlining areas
which, we believe, based on this model, need further attention.
18.2 Context: The NHS Care Records Service and the National
Programme for Information Technology
18.2.1 The National Programme for Information Technology
The English NPfIT is a ten year strategy (which began in 2002) to modernise
the NHS through the introduction of computer systems. In 2005, the
Department of Health established NHS Connecting for Health (NHS CFH) as
a designated body charged with overseeing the implementation of the
programme. NHS CFH has responsibility for setting standards surrounding
the implementation and systems, for providing a range of implementation
resources to individual organisations and Strategic Health Authorities (SHAs),
and also has overall responsibility for liaising with national software suppliers
and accrediting software to be implemented as part of the national
Programme. In this capacity, it holds national contracts with two Local Service
Providers (LSPs), which implement the national solutions, subcontract
software developers and deliver the software into Trusts.
465
The Programme was conceived during a time of economic prosperity and
aimed to deliver national software systems through NHS CFH as a centralised
implementation agency; the underpinning political vision was to modernise the
NHS. This implementation strategy is therefore often described as “top-down”,
with limited capability for individual Trusts to choose or develop their own
software solutions, having to adhere to a national agenda. The underlying
approach was a “one size fits all” model, the hope being that the introduction
of these standard solutions would minimise the risk of inter-operability
problems that had hitherto plagued the NHS (see Chapter 5).
18.2.2 The NHS Care Record Service
The NHS CRS is an electronic health record (EHR) that is currently being
introduced as the quintessential headline deliverable of NPfIT. This is a
complex innovation that will in due course consist of several inter-related
components, these include the following:1 2
•
National Spine, containing the basic capabilities of the system
•
National Network for the NHS (N3), a very large virtual private network
allowing electronic data exchanges between the applications across
organisations
•
Personal Demographics Service (PDS), containing demographic
patient details
•
Images in Picture Archiving and Communication Systems (PACS)
•
Summary Care Record (SCR), which is held on the National Spine and
contains a record of essential clinical information
•
Detailed Care Record (DCR), containing comprehensive clinical
information on individual patients seen and managed in secondary care
•
Secondary Uses Service (SUS) for integration of data from different
sources and then making this available for audit and research
purposes
•
Choose and book, an electronic booking service giving patients the
possibility to choose both location and time of appointments
•
An Electronic Prescription Service (EPS), an electronic service that
allows sending of prescriptions from primary care to pharmacies.
466
The LSPs delivering the NHS CRS currently include Computer Sciences
Corporation (CSC) and British Telecom (BT), who implement software
developed by Cerner (Cerner Millennium), iSoft (Lorenzo) and CSE (RiO).
The implementation of the NHS CRS in secondary care is planned as a
staged approach with software releases being implemented incrementally with
increasing capability. However, there are some differences between clusters
due to variations in software solutions. Lorenzo, the solution being
implemented into the North, Midlands and Eastern (NME) region of England,
has been developed “from scratch” as it is implemented and does not exist as
yet in its full form; Cerner Millennium is in contrast a “finished” product that
has been in routine use in the United States (US) for many years. Although
both solutions are being implemented incrementally, there are some
differences in the scale of implementations. Lorenzo has so far only been
deployed as a so-called “soft-landing” on a small scale with a limited number
of users and capabilities. Cerner Millennium, on the other hand, is
characterised by “big bang” implementations, with whole organisations
moving overnight from paper-based to electronic solutions.
The introduction of the EHRs into secondary care represents a dramatic
change from the current model of working as English clinicians working in
hospital settings are currently typically using paper-based records (with all the
associated idiosyncrasies of storing, retrieving and interpreting written
records). The introduction of the new paperless system is therefore likely to
have a significant effect on organisations, healthcare teams and individual
practitioners, as well as patients.3
Although it is possible that the introduction of the new application will, if
successfully implemented and adopted, in the longer-term result in costsavings, the initial investment is significant, estimated at anywhere between
£6-12 billion.2 Given the scale of the financial spend and the fact that so many
other aspects of the Programme depend on the success of this initiative, it is
vitally important that the implementation of NHS CRS is successful.
467
However, historically the introduction of new IT applications into healthcare
organisations has proved problematic, which is a major concern.
For
example, Sicotte et al. describe the introduction of EHRs into four US
hospitals.4 The cost of this introduction was considerable at $45 million, but
the venture failed due to the application being rejected by healthcare staff who
refused to use it as they felt the application did not fit in with existing care
processes.
This, and other examples,5 6, illustrate the importance of
considering
the
human
dimensions
of
implementation
technologies into existing organisational structures.
7 8
of
innovative
These factors do not
only relate to practicalities and technical issues, but crucially to the design
considerations and socio-cultural dimensions of organisational change.9
18.2.3 The present situation
Overall, as discussed in Chapter 3, the national Programme has made
significant progress in some respects (e.g. implementations of the Spine, N3,
and PACS have been completed), other aspects–and in particular the
implementation of NHS CRS–is considerably behind schedule.
Although there have been hospital-wide implementations of national EHR
software with limited capability, particularly in the London area, these
implementations have tended to have major problems. These have included
issues with lack of user engagement, questions being asked about the
software’s fitness-for-purpose and a general lack of support on the ground for
the national strategy. At present, particularly the sharing of records between
care settings has not been realised as yet and the more advanced clinical
functionalities (such as ePrescribing) have not been implemented as part of
the national solutions.
These developments have been taking place against a changing financial
climate, with a resulting in the tightening of NHS budgets in general and, more
specifically, concerns about the escalating costs of the Programme. This led
to considerable political debate about the need for radical changes in
468
direction. For example, in December 2009, the then Health Secretary Andy
Burnham announced cuts of £600m. The recent change in government
means that the future direction of the national strategy is therefore at present
still somewhat uncertain.10
11
This includes the role of NHS CFH, whose
responsibility has been integrated with the Department of Health’s Informatics
Directorate. The new coalition government has, despite announcing budget
cuts and an estimated £20 billion in NHS efficiency savings, not published a
new IT strategy for the NHS as yet, but announced a general NHS strategy in
July 2010.10 This included a plan to establish new regulatory structures to
monitor progress and quality, increased control of local GPs coupled with a
decreased input from commissioners and the government, and plans to
abolish SHAs.
Nevertheless,
despite
uncertainties
surrounding
the
detailed
future
implementation strategy, the implementation of EHRs in secondary care is
likely to continue as significant investment has already been made. It is likely
that this new strategy will mean a move away from centralised implementation
of national systems towards increased local decisions with respect to both the
procurement and implementation strategy.
18.3
Socio-techno-environmental
dimensions
of
organisational
change—theoretical considerations
As discussed in the previous chapter, several authors have attempted to
devise models of technology implementation in healthcare settings including
those based on critical success/risk factors (with an ingredients-to-successtype of approach) and temporal stages.7
12-14
Most appropriately, the
introduction of technology into healthcare is, however, viewed as a non-linear,
unpredictable and context dependent process of change where the
technology and organisation shape each other (the so-called “socio-technical”
view) during the different stages of design, implementation and adoption.12 1517
Central to this is the notion of “organising”.18 This includes the assumption
that a complex system is characterised by a constant process of change, in
which related structures re-order themselves continuously.19 In this context,
an organisation is conceptualised as an accumulation of events and actions
469
by different actors, taking the focus away from an organisation as a structure
towards an organisation as a process.
Nevertheless, and despite the messy nature of the process surrounding EHR
implementations, a range of factors are repeatedly found to be important.
These factors have been conceptualised to include the following:20-22
•
Innovation attributes (usability, user-friendliness, flexibility, integration
with existing work practices)
•
A focus on end-user perceptions (attitudes, perceived benefits and
concerns)
•
Strategies on targeting/communicating with and engaging of end-users
(training and support, effective communication between different
stakeholders,
effective
change
management
and
leadership,
evaluation)
•
A receptive context and planning (e.g. clear setting of goals/vision and
preparing the organisation for change, incentives)
•
Attention to both macro and micro contexts.
In order to reflect the process and complexity of implementation and adoption,
the present case study avoids using the available models in a concrete way
and incorporates components of existing models to be as inclusive as
possible.
The following sections will summarise the main lessons from the evolving
literature, outlining which factors have been found to influence the successful
implementation of IT systems in healthcare in general and apply these to the
NHS CRS. In so doing, we will attempt to give an indication of the main
recurring
technical,
social/human,
organisational
and
wider
macro/environmental dimensions that are likely to play a role and make
recommendations relating to the future implementation strategy of the NHS
CRS.14 17 23-28
470
The figure below illustrates a model of dimensions and factors identified
throughout our emerging understanding of the literature and by observing the
evolving implementation landscape. It has, however, to be kept in mind that
the division into headings is subjective and that some factors may fit under
two or more categories. Also, the complexity of the process and variations in
context mean that the factors identified are not prescriptive for implementation
success.
TIME - STAGES OF DESIGN, IMPLEMENTATION, ADOPTION, EMBEDDED USE
Wider macro-environmental dimension4
Organisational dimension³
Social/human
dimension²
Future state
Local context of implementation
Technical
dimension¹
¹Includes usability, system performance, integration and interoperability, stability and reliability, adaptability and flexibility, cost, accessibility and
adaptability of hardware
² Includes attitudes and concerns, resistance and workarounds, expectations, benefits/values and motivations, engagement and user input in design,
training and support, champions, integration with existing work processes
³ Includes getting the organisation ready for change, planning, leadership and management, realistic expectations, management and user ownership,
the implementation team, complexity and management in the NPfIT, teamwork and communication, learning and evaluation
4 Includes other healthcare organizations, industry, policy, professional groups, independent bodies and the wider economic environment
Figure 18.1 A refined model emerging from our emerging understanding
of the literature
Using the above model, we aim to build on existing work through undertaking
a detailed and up-to-date case study of the implementation of the NHS CRS
into hospital-based care.
In so doing, we aim to reflect, using existing
complementary frameworks, on the utility of the above model.
Our aim
throughout is to identify, where possible, additional strategic steps that might
471
usefully be taken to maximise the likelihood that the NHS CRS will
successfully be implemented into existing organisational structures and, most
importantly, be used by hospital staff.
18.4 Technical dimension (the technology…)
Implementation of an EHR involves learning to use a new technology and in
some instances replacing an existing one, which may result in negative user
attitudes.29 Choosing the right software is important as it needs to fit in with
organisational and individual user requirements so that it does not need to be
replaced again in the near future.12 30
Factors under the technical dimension relate to the design of the technology
itself, which is commonly considered to play an important part in impacting on
adoption behaviour of users as technical issues can lead to significant levels
of frustration.31
32
Equally, user satisfaction and adoption can also be
positively influenced by the design of the technology.33-35 The way humans
and technology interact is commonly referred to as ergonomics or human
factors.
18.4.1 Usability
Usability problems can pose significant barriers to adoption of IT in
healthcare, as systems that are difficult to use are often rejected by users
(see Chapter 16).12 27 31 36-39 Conversely, if a system is perceived to be userfriendly and easy to learn it is likely to be adopted more readily.17 40
Usability issues are commonly conceptualised to be due to a poor alignment
between work processes and software specifications.33
41
This is often
exacerbated by the complexity of EHRs, which makes them difficult to
navigate and, in particular, to obtain an overview of a patient or particular care
activities.29 42 Usability is also often impacted on by the number of interfaces
in EHR applications, which can significantly slow down work.15 27 36 43-45 The
difficulty for designers therefore lies in balancing the complexity of information
needed for day-to-day healthcare activities without increasing the overall
number of screens.35
46 47
In doing so, they also need to bring together the
472
large amount of information required by healthcare staff within a limited
amount of screen space and to ensure that the interface in any one screen is
not too complex (i.e. that the items on the screen are not too close together).14
35
Usability can be improved by making the software design as intuitive
possible.48-52 This can, for example, be achieved through an increased use of
graphics or “visual hierarchies”, but also through standardised symbols, alerts
and reminders.35
53
Hardware usability can be improved by portable
technology (e.g. tablet computers).12
Van Ginneken describes what clinicians are looking for in an EHR in relation
to usability.45 It is, for example, important that the design of the record
provides an overview (which means finding quickly for what one is looking
for), has a predictable structure (e.g. through use of headings and consistent
buttons, use of tables and graphs) and is clinically relevant (i.e. meeting the
need to be able to find clinically relevant data quickly).
Issues with usability also often occur if the EHR system does not fit in with
existing work practices of users (discussed in more detail below). For
example, Jaspers and colleagues conducted an evaluation on the usability of
a new and an old EHR system in the Netherlands in secondary care through
standardised questionnaires given to clinicians.46 They found that adoption of
EHRs is influenced to a large degree by perceived usability. The more in line
the system was with existing work processes and the more intuitive it was the
more satisfied were the clinicians.
In this respect it is also important to consider that different users have
different work practices, which has implications for individual usability. Rose
and colleagues used a combination of task analysis and focus groups with
clinicians and some nurses to inform the usability of an EHR in US primary
care.35 They discovered that users found the EHR difficult to use and
navigation was viewed as too complicated and not in line with existing workflows. Users also felt that EHR access needed to be quicker. Usability
473
problems resulted in feelings of frustration and in some cases users had
developed workarounds that were perceived to have adverse effects on
patient safety. The authors recommend therefore that designers need to
design the system paying attention to different types and different needs of
different user groups (e.g. experienced/inexperienced users and different
professional groups). Testing systems with users to identify issues has also
been recommended to improve usability and avoid frustration.35 54
The literature suggests that the best way of achieving application usability is
through a close collaboration between the designers of application and endusers.3
55
This may take the form of continuous testing of prototypes in
different groups of end-users and re-design if necessary.56-58 Hartswood et al.
have applied an ethnographic method called “co-realisation” to help achieve
user-informed IT design.59 They argue that a facilitator is important in this
context in developing a partnership between users and designers of the IT
application.
NHS CFH has already made efforts of increasing user
involvement during deployment, which will be discussed in more detail below.
There have thus far however been only limited efforts focused on
incorporating user input into application design. Such engagement needs to
occur with a range of user groups.
Issue 1: NHS CRS software is often not perceived as usable by end-users.60
Recommendation 1: Start with a usable system, paying attention to human
factors’ considerations. Alternatively, start with a system that can be “made
usable” with input from users and prompt incorporation of suggested changes.
A close working relationship between designers and users is essential to
achieve this.
18.4.2 System performance
Frequently, EHR users need to rely on service providers to tell them how the
system should perform and what it can and cannot do in line with technical
specifications.38 This is complicated by the immaturity of many existing EHR
systems, including the NHS CRS.27 32 39
474
When considering the necessary integration of the system into the complex
and fast-moving nature of clinical work, users can be frustrated and unhappy
for a variety of performance-related reasons. The most common one is slow
speed of the application.14 15 23 27 28 39 53 61 62 In this context, it is important that
using the application is not significantly slower than the old system, whether
paper-based or electronic.63 This may include system downtime and long login times, which are frequently reported to delay users.29
39
Slowness of the
system has been reported by many NHS CRS software users, which is of
concern. Whilst logging on with smartcards and pin number to verify identity
is a good way of addressing concerns surrounding confidentiality, the
resulting length of the logging-in process may compromise valuable time in
the case of emergencies. There may here be a potential for using biometric
technologies, but this has obvious additional cost implications and privacy
considerations. Conversely, systems need to be in place to ensure that after
use, healthcare staff will be logged-out after a certain amount of time to
minimise risk of inadvertently breaching confidentiality.55
If the system is not perceived as improving performance, it is likely to be
rejected. Generally, the new system should therefore help to make work more
efficient rather than slowing it down.36 For example, Granlien and colleagues
report on the implementation of an electronic medication record in Danish
hospitals.64 Three years after implementation it was still only used by the
minority of healthcare staff despite increased efforts to facilitate adoption. The
authors conclude that the system was reported to be slow, not user-friendly
and that there was a perceived lack of training on how to use the system.
However, it also has to be kept in mind that an initial (temporary) slow down in
work processes is to be expected. This should ideally ease off as users get
used to the new system as performance also often depends on the skill level
of users.65 For example, Quinzio and colleagues report on the results of a
survey about user acceptance of an anaesthesia information management
system at a German hospital after the system had been used for a period of
five years.66 Although no time savings were reported, the authors suggest that
475
this may have been due to other factors such as users’ lack of typing skills.
They suggest typing classes to increase speed of data entry.
Issue 2: NHS CRS software is perceived to slow down the work of endusers.60
Recommendation 2: There is a need to communicate that using an
electronic system will never be as fast as using paper. In the best case
scenario, it will take the same amount of time. Whilst it should be expected
that systems will slow down work processes initially, there also needs to be
some consideration as to what level of slow performance is acceptable and
ways to address unnecessary levels of bureaucracy.
18.4.3 Integration and interoperability
A new EHR should integrate relatively easily with existing IT systems (see
Chapter 5).23
28 42 49 67
The development of standards for interfacing and
integration with existing systems has been found to facilitate implementation
across studies.12
26 36 51 52 67-70
Conversely, if there are problems with
integration, this can slow down clinical work and result in increased user
frustration.14 15 27 37 65 66 71 For example, Miranda and colleagues report that a
lack of standards inhibited sharing of information across specialties.54 Also,
Sicotte and colleagues, reporting on two longitudinal case studies of EHR
system implementations in Canada, found that one case used a data
warehouse, which meant that new standards and new interfaces had to be
developed to connect systems.14 This led to delays in the implementation
timeline and additional costs. It also impacted on user perceptions of the
system.
However, it has also to be kept in mind that whilst interoperability means that
systems have to a certain degree to be standardised, these have to be flexible
enough to fit in with the complex individual work practices in healthcare.38 44 45
68
The discussion as to exactly how much standardisation is needed in
different contexts is ongoing.
476
Stead and colleagues outline three different types of integration projects with
increasing levels of complexity. These range from a single database, to
integration of similar sources and systems, and integration of sources and
systems that are by nature very different.72 The problem is that individual
commercial systems are often not designed to integrate with other systems.
Integration and interoperability between systems is particularly important and
complex in large-scale IT implementations across different care settings, as in
such scenarios the number of interfaces required is higher than in smaller and
more contained settings (such as for example GP settings).73
The NHS CRS has been set up as an integrated solution by centrally
procuring selected software applications. However, recent changes in the
direction of the Programme mean that the market is likely to open up so that
hospitals can chose other software suppliers. Therefore, issues surrounding
integration and interoperability will become increasingly important. Significant
progress has already been made by NHS CFH in developing standards for
interfacing of systems (the dm+d standard), which vendors supplying systems
have to adhere to. Standards have also already been adopted to authenticate
the identity of users and to ensure the secure transfer of information across
applications (electronic Government Interoperability Framework) and to
outline the technical requirements regarding the specification of the
application (OBS—Output Based Specifications).74 75
Issue 3: The market of NHS CRS software is likely to open up with an
increasing number of vendors.
Recommendation 3: There is a need to build on NHS CFH’s efforts to set
standards for interoperability. A central authority is extremely important for
achieving this.
477
18.4.4 Stability and reliability
Some studies also refer to the importance of system stability. The more stable
a system is, the higher the likelihood of successful implementation and
adoption.28 32 Conversely, a lack of system reliability (e.g. crashes) can result
in negative user attitudes.29 In this context, Bossen describes the
implementation of an electronic medication plan as part of a national EHR
implementation in three Danish hospitals.28 Data on adoption were collected
through observations and interviews with key stakeholders. During the initial
test period, several issues emerged. Log-in times were viewed as too long,
computers froze after a certain period of inactivity and had to be “rebooted”,
and the network was not stable enough.
NHS CRS software itself seems to be relatively stable, but existing networks
and infrastructures to support the software are often not fit for purpose and
different from the environment in which the software is tested. This, in
combination with constant system upgrades, tends to give users the
impression that the system “crashes”.
Issue 4: Existing infrastructures do not support the software adequately.
Recommendation 4: Focus on building adequate infrastructures before
introducing a system. Testing should be done in an environment that mirrors
existing infrastructures.
It however needs to be acknowledged that no system can ever be fail-safe. It
will therefore be necessary to have systems in place and disseminate a plan
of action of what to do in such situations. This will mean devising alternative
forms of accessing and storing data in collaboration with application designers
and may take the form of reverting to paper processes. However, this requires
that staff are trained on what to do in such situations.76
478
Issue 5: No system is failsafe
Recommendation 5: There is a need to have back-up systems in place and
plan for such situations in advance.
18.4.5 Adaptability and flexibility
A common barrier to adoption is that IT systems in healthcare are often not
flexible enough to fit in with the nature of clinical responsibilities and local
needs.28 41 71 77 Conversely, space for adaptation of newly introduced systems
can facilitate adoption and increase user acceptance as the system needs to
in with existing work practices of users (discussed in more detail below).15 23 34
35
Some authors have in this context referred to the “design reality gap”, which
points to the differences in intended use of a technology by designers, as
opposed to actual use (in context by healthcare staff).78 This gap needs to be
closed for the new EHR to be successfully integrated. For example, Aarts and
colleagues describe an in-depth qualitative study of the implementation of a
computerised physician order entry (CPOE) system into a Dutch hospital.15
Here, the purchased US system needed to be (and was) adapted to the Dutch
context through translation of terms and specification of different clinical
pathways. This is also relevant in relation to the NHS CRS as Cerner
Millennium is a US system. Adaptability to the UK context is therefore crucial
for successful implementation.
However, this need for adaptability and flexibility also needs to be balanced
with standardisation and interoperability considerations, especially in relation
to national systems such as the NHS CRS.40 79 80 Designers therefore need to
find a balance between locally customised and shared parts of the system.81
The potential for re-invention is another important described attribute.
Healthcare staff needs to be able to adapt the use of NHS CRS to their
particular profession’s requirements. This may be done via design features
and requires active efforts of involving staff. Supporting the usefulness of this
construct, Ash et al. interviewed health professionals at sites where CPOE
479
was successfully implemented and identified the ability to adapt the
application to local needs as a facilitator for adoption.82
Issue 6: There are trade-offs between customisation and standardisation
Recommendation 6: There needs to be a common agreement on the shared
element of the NHS CRS, whilst there is also a need to allow flexibility for
customisation and adaptation.
18.4.6 Cost
Several studies indicate that costs of the system and the implementation
process can be a significant barrier to adoption.14 27 34 37 38 40 45 47 48 51 54 62 65 77
83 84
For example, a recent systematic literature review to identify barriers to
EHR adoption amongst primary care physicians has suggested that financial
barriers are frequently cited and include high perceived initial investment cost,
high ongoing costs, uncertainty surrounding return of investment, and lack of
financial resources.29
Cost as a barrier to implementation has also been identified specifically in
relation to the NPfIT, although Trusts have to date in theory received the NHS
CRS system “for free” as it is centrally procured.85 Nevertheless, ongoing
costs (e.g. for training and maintenance) and a lack of quantifiable, and
particularly long-term net, returns from these investments are still a concern in
this context.23
34 51 71 84 86
There is also currently no business case for EHR
implementation and its cost-effectiveness is based on speculation in the longterm (as outlined above).38
Issue 7: There is a cost involved in implementing the NHS CRS for both local
organisations and the country as a whole.
Recommendation 7: There is a need to realise that implementing the NHS
CRS should not be based primarily on cost savings as these may take a long
time to or possibly may never materialise.
480
18.4.7 Accessibility and adaptability of hardware
Another crucial factor is providing an adequate workspace for users of the
new application,57 as the lack of access to computers has previously been
identified as a barrier for use.39 56 87 This may, for example involve healthcare
professionals waiting for computer terminals to become available, which again
slows down work.
This means thinking through the positioning of computers throughout the
setting and may involve access testing for simulated emergency situations.
Once users reach a terminal, it is important to consider that a variety of
different individuals may need to use the same computer terminal. This
means that settings at the workstation therefore need to be easily
changeable.70 Similarly, a commonly cited barrier to EHR adoption is a lack of
hardware to support the new system.29
In the context of the NHS CRS, issues surrounding accessibility and
adaptability of hardware are likely to become more important over time as
local systems become more developed. It is, however, important to consider
these issues when planning for EHR implementations, particularly when
considering that the majority of NHS buildings have simply not been built with
computers in mind.
18.5 Social/human dimension (those that interact with the technology…)
In order for successful implementation and adoption to occur, a strategic
targeting of end-users (adopters) of the innovation is crucial. The NHS CRS is
likely to be used by a variety of healthcare professionals as well as patients in
a range of settings. Groups of adopters may, however, differ in terms of
particular needs and or existing skills which may in turn influence individual
adoption. A potential problem is that end-users of the NHS CRS have thus far
not been mapped out and can therefore not be systematically targeted and
motivated to use the application. There is also a lack of research investigating
how adoption of eHealth technologies can successfully promoted.88
481
Once prospective end-users have been mapped out it is important to consider
in which capacity they will utilise the NHS CRS. This is best informed by
users themselves in order to develop a sufficiently rounded understanding.89
A potential problem is, however, that there is still some confusion of how
exactly the SCR will be used—this uncertainty within NHS CFH has a knockon effect on professionals and patients—and there is a pressing need to
address this.2 There is also still confusion over who exactly will use DCR
applications and which organisations will be enabled to share information.
The NHS Care Records Service Registration Authority is a designated
authority for registering staff to use the NHS CRS and may have a role to play
in this context.
Most existing studies in the existing literature have focused on doctors and
some on nurses. There is clearly a need to investigate factors that are
relevant to other users such as administrative staff and allied healthcare
professionals as their work processes and resulting needs are likely to vary
from those of clinicians. For example, a study by Lium and colleagues found
that secretarial staff found adapting to a new EHR system easier than clinical
users.42 This may be due to this group of healthcare staff being used to
working with computers. There is also another stakeholder group, which
although of prime importance, has often been neglected: the patients. It is
acknowledged that they too are important, but patient views are not the focus
of the current review.
Individual differences between healthcare staff also exist in ways of working,
patterns of communication, backgrounds and place in the hierarchy.12
24
These factors are likely to result in each user having his/her own perspective
of the system in the context of his/her own work and at a particular time.41
Issue 8: There are many users and many different contexts of use of the NHS
CRS.
482
Recommendation 8: It is important to map these out and communicate
consistent messages as to what exactly the functionality will consist of so that
appropriate plans can be made.
18.5.1 Attitudes and concerns
It is generally acknowledged that for change initiatives to be successful
change needs to have a receptive context that is characterised by positive
attitudes of stakeholders.90 In line with this, attitudes towards the EHR
amongst users are important in determining adoption decisions. Accordingly,
many studies have found that many EHR implementations fail due to a lack of
user acceptance and satisfaction.7 16 17 27 35 40 45 47 50 51 53 65 66 70 71 77 91
Although significant efforts have been made to assess patient and user
attitudes towards the Programme, some have argued that these were not
conducted with representative samples.80 Nevertheless, there now appears to
be a widespread unfavourable attitude towards the Programme, especially
amongst doctors.92 Clinicians tend to support the underlying aims of shared
EHRs, but often not the implementation strategy; of concern is that support for
the Programme in this group is decreasing.93 Concerns mostly voiced
amongst healthcare staff in relation to EHRs–and the NHS CRS in particular–
include those surrounding confidentiality and security.
83-85 94
12 14 29 32 45 51 62 65 67 69 77
Some of these concerns have to an extent been addressed by
initiatives such as the NHS Care Record Guarantee, which is setting
standards to protect confidentiality.75 It has however been pointed out that if
work is disrupted through security measures, this can decrease user
motivation to carry out a certain task.95 Role-based access features (that
define the level of access to patient information according to profession) can
help address these issues, but this may delay implementation and can result
in reduced user satisfaction due to the limited amount of information provided
by the system to each individual user, and lengthy log-in times.14 28
483
Other concerns amongst healthcare professionals relate to a fear: of
increased workload; that breakdown in computers may lead to threats to
patient safety, about how the system will influence individual roles and
responsibilities (e.g. professional control over patient information); that the
system will become outdated relatively quickly; and that the system will
negatively impact on doctor-patient relationships (e.g. less conversation, less
eye contact).14 17 29 29 31 38 41 45 49 51 65 66 71 83 85
One case study described how concerns surrounding the impact on doctorpatient relationship were diffused by offering users the possibility of portable
computers, whilst confidentiality concerns were addressed by explaining the
system and how it protects confidentiality.49 Another study found that doctors
felt apprehensive towards modernisation including technology in the NHS and
tended to feel that their professional identity and deep rooted values were to
some degree threatened by modernisation.96 In relation to the NHS CRS, it
has therefore been argued that there is a need to address existing concerns
and map out how CRS can contribute to safety and quality of care with robust
evidence in order to address existing concerns.22
Demographic and individual factors such as level of experience with
technology, salary status, personal values and norms also play an important
role in determining attitudes to technology.17
77
There also seem to be
differences in attitudes between younger and older users with seniors
showing lower levels of acceptance.42 One survey indicated that clinicians felt
that the real use of computers would come as the old generation of users
would retire and the younger generation who are more used to using them
would become more prominent.85 Similarly, in a Dutch study, Van Der Meijden
and colleagues investigated the attitudes of future users of an EHR for stroke
patients before the system was implemented.50 They found more negative
attitudes amongst inexperienced computer users and a general expectation
that the EHR would improve the perceived negative aspects of paper records.
Usability, speed and reliability were high priorities in this context as discussed
above.
484
However, despite this, it has to be kept in mind that change is unpredictable
and attitudes amongst players/parties can change over time.97
Issue 9: There are widespread negative attitudes towards the Programme,
with particular concerns surrounding the issues of confidentiality, increased
workload and a negative impact on care provision.
Recommendation 9: There is a need to face these concerns “head-on”.
Mapping of stakeholders is likely to facilitate this as concerns are often
profession-specific.
18.5 2 Resistance and workarounds
Negative attitudes and the feeling that change is externally imposed on users
can lead to resistance to use a system or result in workarounds. Some have
argued that resistance to technology itself is rare; rather, the problem is more
often resistance to the management introducing the technology. 95
Especially, high levels of resistance among clinical users can threaten the
implementation as this staff groups tends to have a high degree of
professional autonomy.14
28
31
98
For example, Doolin describes the
implementation of an information system (that monitored the performance of
doctors) into a hospital in New Zealand. Here, doctors resisted using the
system which ultimately resulted into changing the system itself to encourage
clinical use.99
However, Ferneley and Sobreperez argue that resistance to use has
traditionally been viewed negatively, but is not necessarily so.100 It may, for
example, be a response to an inadequate system and can in certain situations
be adaptive (e.g. in compensating for inadequate technology).80
101
If
resistance occurs then “workarounds” can emerge as a response on the
individual level. Workarounds are a result of the demands/rules that a certain
organisation imposes through a system, which cannot be accommodated with
everyday practice.
485
Ferneley and Sobreperez outline three types of workarounds that can result
from resistance.100 The first type is labelled “hindrance workarounds”. These
tend to occur when using a system is perceived as so cumbersome that users
either enter partial data, wrong data or enter data retrospectively. This type of
workaround may also result in avoidance to use a system. The second type is
labelled “harmless workarounds”. Here, users tend to use the system in a
different way to that which was intended, but the outputs of the system are not
affected (e.g. conducting work in different sequence, sub-dividing tasks). The
third type comprises of “essential workarounds”. These are workarounds that
are viewed as important by users (e.g. retrospective data entry in an accident
and emergency (A&E) care setting).
An example relating to workarounds in healthcare IT comes from a study
conducted by Vogelsmeier and colleagues evaluating the introduction of an
electronic medication administration record. Here, the authors argue that
workarounds used by nurses can be divided into two main categories.102 The
first category included workarounds as a result of a “work flow block”
introduced by the system (e.g. a warning asking the user to reconsider a
certain action) and the second category included the failure to integrate
organisational processes with the technology (e.g. if the system was too
slow). If the block in the system was perceived as unjustified or the system
was perceived as too slow, then staff tended to revert to using paper systems.
However, the authors conclude that workarounds such as these may
compromise the effectiveness and integrity of the electronic system as here
staff deal with an emerging problem without trying to address the underlying
causes.
Similarly, in a recent literature review, Stevenson and colleagues found that
nurses tended to view EHR documenting practices as taking time away from
patient care and close interaction with patients.39 Therefore, notes were often
taken on paper at the bedside and then later transferred onto the EHR
system.
486
Some have argued that in relation to the national Programme, technology
may be resisted as it is not viewed as central to, and/or disruptive to
healthcare professional work.80 98 Nevertheless, it has to be kept in mind that
resistance is typically expressed from a certain viewpoint, typically by those
who want others to comply with the change.103
Issue 10: Introduction of the NHS CRS may be resisted by users.
Recommendation 10: There is a need to seek to understand why this is the
case, as it may be due to managerial views of users finding it difficult to cope
with change, but it may also be a result of inadequate technology.
18.5.3 Expectations
Potential assumed benefits of EHRs include improved patient safety and work
processes, time savings and improvement of communication between
different settings.40
45
Users do tend to have pre-conceived ideas and
expectations about the potential benefits of the technology and how the
software and hardware will look.104 These can influence adoption behaviour
negatively, especially if expectations of users and the product delivered do not
align.27 41 54 105 106 Greenhalgh and colleagues state that this tends to be most
often the case with “off the shelf” systems, which often do not meet
expectations of users.80 The result is an increased risk of implementation
failure.80 101 107
Karsten and Laine have explored expectations through interviews with
healthcare staff in a Finnish hospital.41 They have identified user expectations
relating to the technology itself, expectations of how it influences work
processes, expectations
of
how it
influences
the
organisation and
expectations in relation to usability and training. Negative attitudes and
expectations are often also exacerbated by previous negative experiences
with technology and the many examples of failed IT implementations in the
healthcare sector.14 27 38 40 47 65
487
It is, however, important to keep in mind that different stakeholders tend to
have different expectations. In the context of the national Programme, there
are numerous groups of potential users and stakeholders of EHRs, making it
difficult all of these expectations to be met.80 100 108 This is further complicated
by the large scale of the NPfIT and a “strategic optimism” on the part of policy
makers.97 This results in aims that cannot be achieved and perhaps too high
levels of public expectations, neglecting the complex nature of such
programmes. Others have supported this notion by outlining the often very
high expectations of managers and implementers, who assume that safety
and quality of care can be improved through increased availability of
information,109 that clinical information will be entered at the point of care,110
and that clinicians to support implementation of NPfIT.93 Many have viewed
the expectations from management as unrealistic and too ambitious, which
may contribute to implementation failure.101
107 108
For example, there still
seems to be a lack of belief amongst users that EHRs will result in
improvements to patient care in line with a lack of quantitative data to support
this.29
Issue 11: Expectations of what the NHS CRS will achieve are often not
realistic.
Recommendation 11: There is a need to find out what expectations
stakeholders have and assess to what extent these can realistically be met.
“Strategic optimism” is suitable for getting an initiative “off the ground”, but
will, if sustained and not matched with reality, lead to disillusioned
stakeholders.
18.5.4 Benefits, value and motivations
In order to address expectations and make these realistic, it is important to
consider and explicitly state the potential value to users and implementing
organisations, the potential trade-offs and anticipated effects on work
processes. In this context, Klein and Sorra argue that the success of an
innovation in an organisation is determined by organisational climate and the
488
degree of fit between the innovation and user values. To strengthen
“innovation-value fit” the organisation should involve users in the decision to
introduce an innovation and help employees identify the need for an
innovation by educating them about it. This should involve making the benefits
arising from the innovation explicit.111 It is assumed that if new systems
address an identified user need and fit in with existing work practices, this
creates a “user pull” and facilitates implementation.110 In doing so,
management should not make assumptions that people understand the
vision.91 Otherwise, there is a danger that the system will be rejected by
users.14 17 23 27 31 32 34 37 42 44 49 52 71 Ownership needs to be developed allowing
users to identify individual benefits to them and their patients (not only to the
organisation or healthcare in general).104 110 Therefore, one may wish to state
separate organisational and individual benefits as well as gains from different
aspects of the system.40
42 49
This is because different gains may be
motivators for different groups of stakeholders. Mapping potential benefits and
anticipated changes to work processes for each group of stakeholders and
before implementation can help to anticipate some problems.30
70
For
example, Nikula reports on a qualitative study at two Swedish hospitals with
EHRs and found that managers viewed the EHR from an organisational
perspective (as facilitating organisational processes) whilst clinicians viewed it
as facilitating clinical processes.112 Similarly, the focus of clinicians is mainly
on usability issues, whilst the focus of management tends to be on business
process issues.47 Here, the vision for the future was mainly present in
managers but not clinicians and the author argues that it is crucial that
additional efforts are spent on communicating the vision more effectively to
clinicians.
Van Ginneken considers the alignment of relative efforts and benefits in more
detail.45 The author argues that when considering the implementation of
electronic record systems, efforts of users will only be invested if the resulting
benefits are motivation enough to invest. Therefore, clinicians are more likely
to adopt if they can see the benefit of using an EHR for themselves. In this
context, it is important that the relationship of benefits and efforts is relatively
equal for all stakeholders. For example, benefits of the system are most
489
obvious if clinicians are inputting structured data directly, but often clinicians
are reluctant to do so as it increases their workload and does not yield visible
benefits in the patient encounter itself.29 The benefits of structured data entry
are primarily organisational and governmental (e.g. cost saving). Several
other studies have supported the notion that if desired benefits cannot be
observed and the system is perceived to be of limited value to users, then this
can increase user resistance and inhibit adoption.15
53 113
Conversely, other
studies also support the notion that an identified individual need for the
system can facilitate adoption.30 49 65
In relation to the NHS CRS, there seems to be a lack of tangible benefits for
staff on the ground, which is of concern.62 114 Aligning the NHS CRS with user
and organisational values is clearly difficult, due to the national and top-down
implementation strategy. Although NHS CFH has invested a considerable
amount of effort into benefits realisation, these efforts were mainly based on
business process issues, lacking a focus on individual users. There has also
so far been a lack of attention paid to measuring risks and trade-offs.
An identified need for a system can help to increase staff motivation to use it.
115
Motivations for using the new system therefore need to be identified and
systematically targeted whilst concerns about the new system and barriers to
using it need to be voiced, openly discussed and addressed as early as
possible in order to facilitate user ownership.
116
12 17 25 34 42 49 51 54 61 70 80 94 113 114
If requests or feedback can not be incorporated, users need to be informed
and reasons why need to be communicated in order to prevent frustration.42 54
This also needs to involve outlining the benefits of using the system for
different user groups.12
42 61
For example, Miranda and colleagues describe
lessons learned during implementation and expansion of a nursing
information system in secondary care.54 Actively involving all different users in
the pre-implementation phase in order to help to define what advantages the
system would have for different user groups was viewed as crucial in order to
create a vision and to help define system requirements so that it would fulfil
users’ needs. Similarly, Smaltz and colleagues describe a case study of the
implementation planning of a US secondary care core clinical system
490
(including CPOE, results management etc.).117 They argue that the essence
of a good vision is that it integrates a plan with different goals defined for
different clinical areas and stakeholders. This was achieved through an
assessment of the hospital’s current practices before implementation in order
to help define the vision (e.g. through work process mapping) and to identify
potential areas for improvement. The importance of vision will be discussed in
more detail below.
Issue 12: There appears to be a lack of tangible benefits for individual users
of the NHS CRS.
Recommendation 12: There is a need to assess how both small scale and
large scale benefits can be aligned. There is also a need for benefits
realisation teams to measure both benefits and risks to individual users,
organisations and nationally.
18.5.5 Engagement and user input in design
As attitudes and a lack of perceived benefits can lead to resistance, user
involvement
and
engagement
at
every
stage
of
the
design
and
implementation process is crucial to increase ownership and prevent
alienation.12 14 21 25 28 33 34 36 37 40 45 47 67 68 112 114 116 118-120
This is particularly important when users feel that on them and the decision to
implement has been made without their consultation as could be the case with
NHS CRS software (resulting in a lack of psychological ownership).8 23 25 44 48
51 113
For example, Mair and colleagues have conducted a systematic review
of 66 studies investigating engagement in and resistance to using eHealth
systems amongst healthcare professionals.58 They identified barriers and
facilitators to engagement and resistance and argue that implementation and
integration issues are key to resistance and lack of professional engagement.
Miranda and colleagues report that in this context user feedback may lead to
an increased feeling of user involvement rather than a feeling that they were
presented with a finished product.54 Therefore, efforts need to focus on user
491
involvement and implementation needs to be guided by consensus wherever
possible.117 Here, it is also important that changes to the system are made as
quickly as possible so that users can see that their input is incorporated.36 116
In support of the importance of active user engagement in EHR
implementations, case studies of successful implementations often describe a
substantial amount of user involvement. For example, Dagroso and
colleagues describe the implementation of an obstetrics system in US
secondary care.116 During the roll-out they uncovered problems with the
design and functionality of the system. As a response to complaints from
users, the implementation team reviewed the performance of the system,
heavily involving users in the implementation process (e.g. on-site trainers
were used to promote the system and collect user feedback). A new improved
system was launched tailored to the needs of users which increased user
acceptance and resulted in successful implementation.
Design of the system should to be based on and tailored to individual needs
and this can only be achieved through close collaboration between designers,
IT-staff and end-users.17 32 36 39 40 42 42 46 50 51 53 66 70 79 83 94 101 104 118 118 119 The
underlying assumption here is that users know their environment and the
context of use of the system best, whilst developers of systems are often
detached from the use of systems.109 121 For example, Erdley and colleagues
describe how a new information system in US prenatal care was developed
with no collaboration between designers and nurses (the users).78 The result
was a ‘technically correct but unwieldy application’.
In this context, a lack of customisability resulting in a system that does not
meet the individual needs of users is a commonly cited barrier to EHR
adoption.22
29
This may be addressed by implementing a combination of
standardised core solutions and customised elements, corresponding to
particular users and environments (so-called configurational technology).122
This can help designed solutions to be adapted to the particular context of use
and help to facilitate ownership.123 However, some have stated that it is
492
exactly the limited local customisation that is one of the biggest risks to the
success of the national Programme.93
Tailoring a new EHR system to individual needs also involves an analysis of
the cognitive aspects of use, the findings from which then have to be
incorporated into design.33
46
If the design is too complex then the cognitive
demands on users increase, errors are more likely to happen and usability
goes down.35 The difficulty here is that cognitive needs of healthcare
professionals are likely to differ and that this needs to be incorporated into
design. For example, Jonson and Turley undertook a cognitive task analysis
to explore the needs of clinicians and nurses.43 Participants were asked to
look at different cases on the EHR and were instructed to think aloud. They
found that nurses tended to view the patient in observational terms (e.g. in
relation to monitoring and documenting changes), whilst clinicians tended to
view the patient in causal terms and in relation to decision-making (e.g. in
relation to diagnosis, treatment). The authors conclude that clinicians and
nurses have different practice models and that IT design needs to take these
differences into account by designing different interfaces tailored to the needs
of the professions.
However, other healthcare professionals’ input in system design has so far
somewhat been neglected. It has been argued that this has resulted in
systems not being aligned with, for example, nursing needs.124
125
Moen
describes EHR implementations from a nursing perspective.26 Nurses handle
large amounts of information and present a large group of EHR users. It is
therefore important to assess existing nursing work practices before design
and facilitate nursing input so that the system can meet nursing needs.26
Nursing leadership is important in this context in order to identify barriers to
implementation specific to nursing practice. Similar issues arise with other
professions. No study included in the review has addressed any other
members of the healthcare team such as administrative staff and
pharmacists.
493
Some have also cautioned that user input in design pre-supposes that users
know what they want and that this does not change over time. This is often
difficult, especially when being presented solely with an idea of an electronic
solution, without understanding how this may be applied in practice (see
Chapter 16). It has
therefore been argued that ideally design should be
evolving and users should be able to try a system making subsequent
changes to its design.126 This then also allows the system to respond to
changing contextual requirements. However, the different requirements of
different users may complicate this idea.
In relation to the NPfIT a more participatory approach to development and
implementation has repeatedly been advocated.120 Catwell and Sheikh outline
that this should be characterised by early user involvement and development
of a shared vision, formative evaluations and testing of prototypes to assess
whether the system is perceived as usable, potential redesign so that the
system fits with user’s needs, summative evaluation and benefits identification
once implementation is completed, with incorporation of issues that have
been identified along the way.
NHS CFH has already made significant efforts to facilitate clinical
engagement in the context of the national Programme.21 They have, for
example, established a Communications and Engagement Team with the
responsibility to facilitate communication between key stakeholders including
NHS CFH and the NHS and other stakeholders such as the commercial
sector (e.g. LSPs) or the media. This is divided into public engagement
(aiming at increasing patient involvement) and clinical engagement (aiming at
increasing clinical engagement). NHS CFH has also developed user informed
design guidance as part of the NHS Common User Interface to address users’
needs.22 Despite these efforts, many have argued that there is still a lack of
clinical engagement in the NPfIT and NHS CFH’s efforts have been criticised
as lacking input from other user groups, which leads to the initiative being
viewed as being imposed by the government by many.21
22 93
There is of
course an important tension here because a national implementation of
systems by definition means limited local customisation.22 110
494
Issue 13: Engagement and user input in design is limited in national
implementations such as the NHS CRS.
Recommendation 13: The key is that different user groups are not only
consulted, but feel that they are listened to. The new strategic direction of the
Programme with increased local input and choice may help to facilitate this.
18.5.6 Training and support
Finding the time to allocate time for training in the busy healthcare
environment can be difficult, but intensive training and support is important for
the successful implementation of a new IT system as it can facilitate adoption.
17 23 28 29 32 36-38 40-42 44 49 51 53 61 62 64 83 85 94 113 115 118 119 127
Systems that are
perceived as easy to learn tend to be adopted more readily and “well-trained”
users tend to be more satisfied with a new system.17 66 Conversely, if training
is not adequate this can decrease confidence and inhibit adoption.31 65 Some
have even gone as far as to state that the lack of training users is one of the
most important factors to contribute to large scale IT implementation failure.101
Most studies stated that in hindsight training efforts should have received
more attention and therefore it is of primary importance to allocate sufficient
time for training in advance.40
42 51 54
Planning training will also involve
considering how it will fit in with clinical responsibilities. It is therefore
important that the organisation supports training efforts by helping clinical
users to free time e.g. by reducing their workload.45 Training needs to begin
well before the new system is implemented and continue when it is
established so that arising problems can be dealt with effectively and do not
compromise care.12 25 69 118 The issue here is also that some functions are not
practised and therefore often forgotten.36 For example, this may also involve
training staff in using paper processes in case of system failure.76 Training
and support should therefore be ongoing as users discover new functions and
as software changes. Maximising use and gaining proficiency in using a new
system may take years and should start well before actual implementation.12
13 32 35 36 38 42 66 70 128
495
Extensive user support during the initial “go-live” period is most important.25 51
61 127
live.
51
In addition, ongoing help has to be available when needed after goThis means quick turnaround and response times to arising problems.39
One successful implementation project had training coaches available who
were reachable by pagers, which facilitated availability and timeliness of
support.119
Ideally, special attention should be paid to target ward managers, senior staff
and local leaders with thorough training so that they can help to train other
staff.42
70
In one case study discussing the introduction of a documentation
system at a hospital in Canada, for example, some nurses were trained as
change managers and acted as coaches in the initial period after go-live.13
They provided support, but also collected feedback from users, which was
then incorporated in the redesign of the system. The authors conclude that
coaches were successful as they were peers of the users (nurses) and
understood existing work processes. However, it needs to be kept in mind that
training by peers means that one learns only what the peer shows and how
he/she uses the system.42
A variety of training approaches can be used including web-based formats,
classroom-based formats, reading material and one-to-one training formats.
Studies indicate that training sessions with the possibility to try out the system
and one-to-one training sessions are most effective.12 31 41 65 115 116 119 128 This
can be done on the ward itself or through simulated clinical scenarios.32 49 70 91
Short sessions of on the ward training seem most effective, as this integrates
best with busy clinical work schedules.36 66 115 Due to the nature of healthcare
professionals’ work, a large amount of paper learning resources may not be
appropriate.70
However, it has to be kept in mind that not all training formats and training
contents are appropriate for all types of users. Users are likely to draw on the
system in different ways depending on their role. Training therefore needs to
be tailored accordingly in the light of existing levels of experience, existing
496
skills, existing needs and emerging problems.12 16 36 50 66 115 In this context, an
initial assessment of current skills before training commences and associated
tailoring of training efforts in line with levels of existing computer skills as well
as assessments of the context of use is important.29 32 35 51 52 61
Again, training is complex in a national implementation and due to the large
scale of the programme, limited resources, and tight implementation
deadlines, in-depth training and intensive support are often not possible.
Therefore, the responsibility for training in NHS CRS software has been
devolved to local organisations. To support training of users, NHS CFH had
set up a national “Education, Training and Development Programme”, working
with organisations and suppliers to help with effective training by developing
standards and providing guidance surrounding delivery and evaluation.129
However, some have cautioned that early implementers of NPfIT software
tend to be enthusiasts and followers are likely to need more support. This has
to be planned for in terms of resources and can easily be under-estimated.93
Due to some NHS CRS software still being in development (e.g. Lorenzo),
dummy training giving users the chance to play around with the system, is
complicated and has often not been achieved in relation to the NHS CRS. A
Training Messaging Service (TraMS) with a 'Like Live' training environment
was planned to be introduced and to enable NHS staff to practice on dummy
patients for the NHS CRS, but the majority of users do not seem to be aware
of this or have access to it.130
Issue 14: Central guidance in relation to training has to be limited and
organisations need to devise their own approaches best suited to their staff
and needs. Availability of software to play around with and tight
implementation deadlines may compromise the quality of training.
Recommendation 14: Organisations need to have increased input into
implementation timelines so that they can prepare well for go-live. Dummy
software reflecting the system that will ultimately be used needs to be made
available.
497
18.5.7 Champions
The value of key individuals in facilitating the adoption of information systems
in healthcare has repeatedly been highlighted in facilitating adoption and in
helping to embed systems in the organisation.131 Key individuals may include
IT and clinical leaders, local champions, respected peers, opinion leaders, so
called “super users” and “boundary spanners” (i.e. those that have dual
roles).12 17 23 25 27-31 45 49-51 65 69 70 118 123 132 The role of these key individuals may
involve winning over others, feeding back emerging problems and
recommendations, facilitating user engagement, and communicating the
vision.17 31 118 The underlying assumption is that if these key individuals start
“modelling change” then others are likely to follow.103
As organic spread of benefits seems most effective, key individuals are ideally
peers of the user group.42 Therefore, it has been argued that support of senior
clinical staff is crucial in getting other clinical staff on board.115
Due to the complexity of the healthcare environment, however, one may need
to consider involving departmental leaders from different clinical specialties.132
For example, Bossen describes how in a Danish EHR implementation,
several groups were formed to facilitate cooperation between different
stakeholders.28 Influence and range of influence of key individuals was found
to be particularly important in this context. Similarly, further illustrating the
variety of roles these individuals can play, Sicotte and colleagues describe
two cases of EHR implementations in Canada. In one case, the impact of a
selected project champion remained local, whilst in the other case there was a
local clinical champion who was heavily involved in system design at each
site.14 Conversely, if champions themselves become negative, then this can
compromise implementation and inhibit adoption through impacting on
attitudes of others.15
31 83
Greenhalgh and colleagues refer to “negative
champions” in this context.31
In
support
of
the
important
role
of
key
individuals
in
facilitating
implementation, Nagle reports on a Canadian government initiative to
498
promote EHR use and to promote healthcare professional engagement.124 In
this context, bringing different groups of healthcare professionals together and
establishing national and international peer networks with a view to discuss
lessons learned was viewed as crucial amongst key stakeholders.
Furthermore, highlighting the importance of key individuals in relation to
training, De Mul and colleagues, describe the successful implementation of a
Dutch intensive care unit system.118 They had “super-users” who helped
others with learning how to use the system. It went live initially but due to a
decrease in performance the ward returned to using paper systems. After
more testing, refining the system, increased training efforts and incorporating
user feedback, the system was implemented without further problems. Other
studies have confirmed that an increased use of power users that can help
out when needed can facilitate implementation.32
However, in the context of the national Programme, despite NHS CFH’s
efforts to appoint and involve clinical champions on a national level, some
have argued that their influence has not been sufficiently harnessed as they
have constantly changed during the Programme.93 In addition, national
champions can only influence peers to a limited extent as local champions
seem more important to facilitate adoption.98 Also, some non-clinical staff
groups have so far been neglected such as, for example, administrative staff.
Issue 15: National champions only have limited influence locally and there is
an increasing threat that champions become negative champions if NHS CRS
software doe not align with their expectations or remains stagnant in terms of
development.
Recommendation 15: Despite the usefulness of national clinical leads, it
seems that local clinical leads are most successful due to their increased
contact to the target group. There is an urgent need to target other user
groups and carefully assess the impact and address concerns of negative
champions.
499
18.5.8 Integration with existing work processes
New IT systems are often viewed as an opportunity to review existing work
practices and increase organisational efficiency, getting users to adopt new
ways of working desired by the organisation. It is important that the new
system during this development is effectively integrated with existing work
processes. Otherwise, the technology is likely to get rejected by users as
healthcare staff need to retain patient safety as the top priority.14 22 26 27 33 35 36
38 42 47 51 65 68 80 133
If staff find it hard to integrate a new system into their day-
to-day work practices they may adopt either “partial use” (i.e. only use the
parts that are useful to them) or “workarounds” (avoid using the system, as
discussed above).81 100 110 Some have argued that this is particularly likely in a
situation where a new system is perceived as being imposed on users as is
the case of the NPfIT.110
A new system is likely to result in new practices and can also lead to changes
to existing professional responsibilities.28 30 47 104 110 For example, in one study
clinicians felt that the introduction of the EHR resulted in them having more of
a data entry role than was thought appropriate.47 This again illustrates the
important role of user input in design as well as adaptation and customisation,
in order to integrate systems with existing work practices and facilitate
implementation.80 Even if effective planning and work process mapping occur,
one will never foresee all changes in work practices that will be brought about
by a new system in different settings. There is therefore always a need to try
out the system in practice and then keep re-configuring it based on local
experiences.81
Not having to make radical changes to work processes can, on the other
hand, be a facilitator for adoption and implementation.40 Stewart and Williams
have referred to the notion of “domestication” in this context.122 Here it is
assumed that users adapt to and integrate the new technology with existing
practices (this may involve getting around faults through workarounds), and
eventually the technology becomes part of the everyday work routine.
500
Integration can also be facilitated by a thorough analysis of existing individual
work practices before the system is introduced and a recognition that these
vary across individuals and settings.12 17 25 33 35 44
45 50 67 71 118 132 134
System
designs then need to be configured accordingly, so that they fit in with existing
routines and the roles of different users.31
36
43
This requires the
implementation team and system suppliers to understand the clinical context
of use a top priority and to allocate sufficient time so healthcare staff can
modify existing practices.40
132
Healthcare staff, on the other hand, need to
recognise that new routines may have to be developed and existing work
practices may need to change.36 This also means a recognition that new
practices should not be mirrored on paper systems as this may limit potential
effectiveness of the new EHR.54 In line with this, Lium and colleagues argue
that users are often still trapped in “paper-thinking” and that this may inhibit
adoption (paper is still used and viewed as valuable for certain tasks).
According to these authors, paper routines need to be dropped and new
routines n

NHS CFHEP 001 Extension: Final report

Transcription

Similar documents

Erb`s Palsy - Home Stretching Programme (1 of 2) - ICID

March 2014 - NHS Highland

Give Carolinas (Summer 2014) - Carolinas HealthCare Foundation

Department of Health Manila 2012

Health inequalities review of physical activity

Hard to reach communities appendices

TSTool User`s Manual - Colorado`s Decision Support Systems