Resilient Healthcare Going Forward
Transcription
Resilient Healthcare Going Forward
Resilient Health Care: going forward… 8th May 2015 Dr Al Ross (PhD; C.ErgHF; C.Psychol) Lecturer in Behavioural Science Glasgow Dental School FAQ • What does resilience thinking imply for Y? • How does resilience relate to X? • How can a resilience model inform my practice with regards to Z? 2 The Safety I model • Systems are inherently safe and well designed – People and technology are unreliable • Errors and breakdowns cause adverse events – Technical rational approaches can be applied to identify specific failures – Decomposition allows for understanding Human error unsatisfying explanation To Err is Human....? • Er, well…. in one sense maybe, but… • But “Human Error” as we use it in context in healthcare is fundamentally socially constructed • It comes with baggage, stigma, blame that are all learned Google “cat fail”: 109mil results…. • Cats have no fear of failure • They simply explore the limits of their physical environment until they succeed • They become very adept very quickly 1 • • • • • To adapt is human To test To get feedback To try again To overcome Error and failure are learned and constrain us from then on Michael Hodder Kerry Hodder in 2007 • ‘[…] I am angry at Network Rail,' she adds. 'They are guilty of killing 31 people including my husband. Michael has been vilified as the driver who drove through a red signal, but he was a victim as much as anyone else. • 'The bosses […] had known for years that the signal outside Paddington Station was dangerous, yet they did nothing about it. He was led into a rat-trap and there was no way out. Obscured signal Unusual design Environment • The red aspect of SN 109 was particularly badly obscured by the overhead electrification equipment; it was last of all the gantry 8 signal aspects to become clearly visible to the driver of a Class 165 approaching from Paddington Bridge Training misaligned • Trainer: “I was not there to teach ... the routes. I was totally to teach ... how to drive a Turbo – The training manager was unaware of this • No proper route plans; drivers relied on handwritten sketch plans they passed amongst themselves Learning and co-ordination • Eight trains passed signal SN109 at red in the preceding six years; nothing had been done • Railtrack employee with formal responsibility for action-tracking had been told his responsibility ended once someone accepted an action, and did not extend to checking that they had acted upon it 2 • • • • • Proximal is not enough Where there's smoke, there's fire. And fuel. And oxygen. And an ignition source. And a system that put them all in the same place In retrospect… • Misaligned system (design, capacity, learning) • Brittle system; relied on one thing and one thing only for it’s safety “SINGLE PATH TO SUCCESS” • Front line staff somehow, with reducing training, and with tight time pressure, had to pick out the correct action. Every time. And if they couldn’t adapt, an accident would be the inevitable consequence. Ring any bells re. healthcare? • Design problems, poor learning and feedback, unhelpful regulation • Limited training, time pressure • Reliance on front line staff adapting • Human error as an after-the-fact explanation You are all Michael Hodder… Supposed to be: Lipitor 10mg PO 1 QD Read as: Zyrtec 10mg Filled Rx: Zyrtec 10mg 22 3 Structure is vital • The chemicals in the human body are easily obtained • Mixing them up will not create a person • Structure is what life emerges from Root cause analysis Produces linear structures Linear trajectories A way to structure? Ross et al System resilience CTW 2014;16(1): 91-102. Consideration of complexity: comorbidity; psychological; social Organizational learning takes place PATIENT Patient information is gathered and assessed; initial decisions are taken; problems anticipated CLINICAL AIMS Organizational factors can form gaps or bridges in resilient systems THE WORK SYSTEM Staffing, procedures, equipment CARE PLANS SPECIALIST AND NONSPECIALIST DECISIONS Decision feedback and support Clinical aims evaluated and revised Care decisions balance organizational factors and clinical aims; multi-professional co-ordination necessary TREATMENTS AND PROCESSES OUTCOMES REVIEW AND MONITORING Process monitoring; patient outcomes reviewed Outcomes emerge from complex patient and organizational factors 4 Post hoc investigation is biased • (Genesis 3:12-13) The man said, ‘The woman whom you gave to be with me, she gave me fruit from the tree, and I ate.’ Then the LORD God said to the woman, ‘What is this that you have done?’ The woman said, ‘The serpent tricked me, and I ate.’ 30 Outcome bias • Did not reach the patient • Reached the patient; no harm • Monitoring required • Temporary harm and intervention • Initial or prolonged hospitalization • Permanent harm • Intervention to sustain life • Death • We investigate red deeper • We come up with deeper systems causes • This is a property of the attributional bias not the world 31 • Davies et al 2003: post event attributions A way to address post-hoc bias? 5 Incident data are not scientific • The chances of errors and violations causing harm are not the same as the chances of finding something somebody did wrong once harm has been identified P (non-compliance/fail)= big P (fail/ non-compliance) = small • The evidence shows that a significant proportion of errors (probably the majority) do not result in harm for the patient, either because they are detected and mitigated […] A way to quantify effects? 6 Operating procedures are tools • Don’t mistake the map for the territory • SOPs do not align all situations • Use will diverge as they diffuse • They bring demands –They constrain adaptation 39 The world • Analogue • Fluid • Dynamic • Fuzzy • Qualitative • Complex 40 The protocol • • • • • • Digital Static Fixed Discrete Quantitative Linear 41 Tools to help us, not faithful representations Protocols as tools • Wallace and Ross in Beyond Human Error (2006 p219) “Instead [...] of rules with the hidden implication ‘do this or you will be fired’, they should perhaps be offered more in a ‘these are some of the methods we have developed of doing these particular tasks here, and we have found them useful” An unspecified world • In (the hospital) … hardly anyone knows all the extant rules, much less exactly which situations they apply to for whom and with what sanctions • […] all categories of personnel are adept at breaking rules […] when it suits convenience. … hence the area of action covered directly by clearly enunciated rules is really very Strauss A et al. (1963) small Page 44 Protocol modification WHO applied: variations in practice • Whiteboards in theatre • Hanging a clipboard from an IV stand that accompanies the patient from the anaesthetic room into theatre • Laminated checklists to act as a prompt • Incorporating the checklist into care pathway documentation • Inputting the checklist into IT systems such as Galaxy • Add ‘check completed’ sticker to patient notes • Adding a column to the Theatre Register for the surgeon to sign showing the checklist was used for the patient Protocol divergence • • • • The rules are just a starting point Guidelines get interpreted locally Culture is important and changes over time Improvisation is inevitable – Compliance is dehumanising – Adjustment and modification characterise humans at work A way to evaluate tools? From Kevin MD (online) I admitted a gallbladder over the phone at 2am. I gave the nurse admitting orders which included one for a broad spectrum antibiotic… When I saw the patient in the morning […] 49 From Kevin MD (online) It had never been given; the floor nurse held it so that it wasn’t administered until 60 minutes before […] OR time, just like the algorithm dictates, despite the fact it had been ordered nearly 8 hours prior to the case, not for peri-op prophylaxis, but for treatment of an established pathology 50 From Kevin MD (online) […] here you have a situation where health care providers are so terrified of violating […] that they end up withholding necessary treatment. It’s just astounding.. Further implications… • Research – Model generates questions • Process mapping – Higher level processes can be interrelated • Training – Aligning demand and capacity – Studying performance adjustments A way to design training? Demand Alignment Capacity Success Adaptations Adjustments Failure • "Taking a model too seriously is really just another way of not taking it seriously at all.“ –Prof Andrew Gelman; Columbia NY Delighted to chat more… janet.anderson@kcl.ac.uk alastair.ross@gla.ac.uk