analysis of individual health insurance data pertaining to pap

Transcription

Scientific institute of Public Health
Departement of Epidemiolopgy
Analysis of individual health insurance
data pertaining to pap smears,
colposcopies, biopsies and surgery on
the uterine cervix
(Belgium, 1996-2000)
Marc ARBYN1,2, Herman VAN OYEN
IPH/EPI REPORTS Nr. 2004 – 021
Epidemiology, October 2004; Brussels (Belgium)
Scientific Institute of Public Heatlh, SIPH/EPI REPORTS N 2004 – 021
Depotnumber: D/2004/2505/42
Analysis of individual health insurance data
pertaining to pap smears, colposcopies, biopsies
and surgery on the uterine cervix
(Belgium, 1996-2000)
Autors:
Marc ARBYN 1,2
Herman VAN OYEN
1
1.European Network for Cervical Cancer Screening
2.Scientific Institute of Public Health, Unit
Epidemiology
Scientific Institute of Public Heatlh
J. Wytsmanstr. 14
1050 Brussels
Belgium
Tel: + 32 2 642 50 21
Fax: +32 2 642 54 10
e-mail: m.arbyn@iph.fgov.be
http://www.iph.fgov.be/epidemio/
SIPH/EPI REPORTS Nr. 2004 – 021
©2004 by Scientific Institute of Public Health, Brussels (Belgium)
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any
form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of
the publisher.
Analysis of individual health insurance data pertaining to
Pap smears, colposcopies, biopsies and surgery on the
uterine cervix
(Belgium, 1996-2000)
1. Table of contents
1.
2.
3.
4.
5.
6.
7.
Table of contents....................................................................................................3
Executive summary................................................................................................5
Résumé et commentaires .......................................................................................7
Samenvatting en commentaar ..............................................................................10
Introduction..........................................................................................................13
Material and methods...........................................................................................15
Results..................................................................................................................17
7.1. Papanicolaou smears of the uterine cervix....................................................17
7.1.1.
National level .......................................................................................17
7.1.2.
Regional level ......................................................................................26
7.1.3.
Provincial level ....................................................................................30
7.1.4.
District level.........................................................................................35
7.2. Profession of Pap smear takers .....................................................................43
7.2.1.
Regional level ......................................................................................43
7.2.2.
7.2.3.
District level.........................................................................................45
7.3. Colposcopies & cervical biopsies .................................................................48
7.4. Hysterectomies..............................................................................................51
7.4.1.
National level .......................................................................................51
7.4.2.
Regional level ......................................................................................55
7.4.3.
7.4.4.
District level.........................................................................................59
7.5. Cytological screening in hysterectomised women........................................62
7.5.1.
Use of Pap smears one year after hysterectomy ..................................62
7.5.2.
Use of Pap smears in the 3-year period following hysterectomy ........65
8. Discussion ............................................................................................................67
8.1. Quality of the dataset ....................................................................................67
8.2. Comparison with results of the National Health Interview Survey (HIS) ....67
8.3. Comparison with statistics from by the National Health Insurance
Institute .........................................................................................................70
8.3.1.
Collection and interpretation of cervical smears .................................70
8.3.2.
Colposcopies ........................................................................................75
8.3.3.
Hysterectomies.....................................................................................75
8.4. Prevalence of cytological cervical lesions ....................................................77
8.5. Conclusions...................................................................................................78
8.6. Propositions for further research...................................................................80
9. Acknowledgements..............................................................................................80
10.
Glossarium .......................................................................................................81
11.
Abbreviations...................................................................................................82
12.
References........................................................................................................82
13.
Annexes............................................................................................................85
IntermutCVX96_00e.doc
6/10/2005
3
13.1.
13.2.
13.3.
13.4.
13.5.
Pap smear use at District level ......................................................................85
Profession of smear takers, by district ..........................................................85
Colposcopies and cervical biopsies, by district ............................................85
Hystectomies at District level .......................................................................85
Data files and statistical syntaxis files ........................................................102
6/10/2005
4
2. Executive summary
Back ground
In 2003, the Council of Europe, recommended to offer organised cytological cervical
cancer screening to all women of the target population in all EU member states. An
optimal coverage is the main determinant of success of a screening programme.
The screening coverage in Belgium could until recently only be assessed by surveys.
The current report aims to assess the real cytological screening coverage and also the
consumption of medical acts for screening and follow-up or treatment of screenpositives using a data file obtained from the Inter-Mutualistisch Agentschap
(IMA)/Agence Inter-Mutualiste (AIM), which pools information from all health
insurance agencies in Belgium.
The underlying study will be used to define strategies for a future cervical cancerscreening programme.
Material and methods
On demand of the Scientific Institute of Public Health, a data file containing more
than 13 million individual patient records was compiled by the Inter-Mutualistic
Agency. It contained information of all medical acts related to cervical screening, and
diagnostic or therapeutic interventions on the uterine cervix (Pap smear collection and
interpretation, colposcopies, cervical biopsies and their interpretation, surgery on the
cervix) from the years 1996 to 2000.
Results
Screening coverage
The cervical cancer screening coverage, defined as the proportion of the target
population of women between 25 and 64 years old, that had a Pap smear taken in the
last 3 years, measured in 2000, was 59%. The range of variation in screening
coverage at the level of the Regions was small: 57% in the Flemish Region, 58% in
Brussels and 61% in the Walloon Region. The increase in coverage measured over
the period 1996-2000 is limited and respectively 2%, 5% and 3% in the Flemish,
Brussels and Walloon Region. This limited increase in the Flemish Region is
remarkable, given the fact that only in this region a screening campaign was
organized.
Differences at lower geographical levels are important: at provincial level in 2000, it
is situated between 52% (Luxembourg) and 69% (Walloon-Brabant); whereas at
district level the highest coverage is found in Huy (71%) and the lowest in Arlon
(only 40%).
At the national level, in the youngest 5-year age group of the target population,
women of 25-29 years old, the coverage is 64%. It increases until 67% among
women in the age range 30-39 years. From the age of 40 the coverage decreases
gradually until 56% in the 50-54 year age group. Thereafter the coverage declines
6/10/2005
5
more rapidly. In the Brussels and Walloon Region this decline by age group is less
pronounced.
The coverage computed from the individual health insurance data file is substantially
lower than the estimates derived from the national health interview surveys. For
instance at national level, for the years 2000-2001, this difference was 11%. This
discrepancy is probably due to reporting biases, which - according to the international
literature – are inherent to interview surveys.
Screening interval and target age range
The modal screening interval is one year. A time span of 3 years is observed in only
5% of women with 2 or more smears in studied time period. Ten percent of all
interpreted Pap smears are taken from women younger than 25 years and 7% in
women older than 64 years.
Profession of smear takers
The proportion of smears taken by general practitioners is limited and varies
substantially by Region: in 2000, in the Flemish Region one fifth was taken by GPs,
whereas only 8% and 3% in Brussels and the Walloon Region respectively. The trend
between 1996-2000 was decreasing.
Pap smear use
The amount of used smears is theoretically sufficient to cover more than 100% of the
target population over a time span of 3 years. In 2000, the ratio of the number of Pap
smear interpretations over the mid-year target population concerned was 1.1. This
ratio should be higher if we were able to adjust the size of the target population by
excluding women whose cervix is removed. Nevertheless, only 59% is covered with
one or more smears.
In absolute figures: 3 million Pap smears were interpreted in the period 1998-2000
which were taken from only 1.6 million women. 1.1 million women did not get a Pap
smear. The excess use of cervical cytological examinations was 88%, which means
that each screened women received on average 1.88 smears over a 3-year period. The
excess smear use was high in all parts of Belgium: it was less high in the Flemish
Region (86%), and was highest in the Capital Region (99%), and intermediate in the
Walloon Region (90%). Of course, a part of this “excess use” is used for follow-up of
women with previous cervical abnormality. Assuming that 10 % of the overuse is
spent for follow-up, we can estimate that yearly about 400,000 Paps are taken that do
not contribute to screening coverage or follow-up. This corresponds with an
estimated amount of € 8.2 million per year reimbursed by the National Health
Insurance Institute for taking and reading Pap smears with limited utilitya. This
amount is further increased with € 3.9 million for screening beyond the target age
range.
Colposcopy use
An impressive amount of colposcopies are performed in Belgium. At the national
level, on average, one colposcopic examination is done for every 3 Pap smears. In the
Walloon Region the ratio of the number of Pap smears over the number of
colposcopies is even less than two, whereas for the Flemish Region this ratio is five.
In certain Walloon districts the number of examined Pap smears equals the number of
a
Costs for consultation by a gynaecologist or GP were not taken into account.
6/10/2005
6
performed colposcopies. The biopsy/colposcopy ratio is low (on average 5%) which
is due to the very high frequency of colposcopy in normal women.
It is clear that colposcopy is not used as indicated in national or international
guidelines. Colposcopic exploration is indicated in case of a first observation of HSIL
or glandular abnormality or after a second observation of ASC-US or LSIL or after a
positive HPV triage result in ASCUS-cases and in follow-up after treated lesions.
General conclusions
The cervical cancer screening coverage in Belgium, defined as the proportion of
women in the 25-64 year age range that received a Pap smear in the last 3 years, is
only 59%. Nevertheless, the amount of smears used is theoretically sufficient to
cover the whole target population.
Estimation of the cervical cancer screening coverage, derived from interview surveys
should be interpreted with caution, given the inherent risk of overestimation.
Most often smears are taken by gynaecologists; in the Walloon Region, it is almost an
exclusivity.
Colposcopy is too often used simultaneously with the Pap smear, whereas its main
clinical indication is the exploration of women with previous cytological abnormality.
Structural reduction of overuse and re-investment in coverage and quality
improvement can potentially result in more life-years saved, without increase in
public funding.
Measures foreseen in the European Council Recommendation on Cancer screening
should be binding and universal. Regionally isolated actions such as those undertaken
in the Flemish screening programme, did not yield the desired result, because they
were based on voluntary participation without financial compensation for involved
health professionals.
Health authorities of the Federal and Community Governments and representatives of
the scientific societies should meet as soon as possible in order to define a rational,
evidence-based and cost-effective screening policy.
3. Résumé et commentaires
Historique
En 2003, le Conseil de l’Europe a recommandé d’offrir un dépistage cytologique
organisé du cancer du col de l’utérus à toutes les femmes de la population-cible dans
tous les Etats-membres de l’UE. Une couverture optimale est le principale facteur de
succès d’un programme de dépistage.
Jusqu’à récemment, la couverture du dépistage en Belgique pouvait uniquement être
évalué par des enquêtes. Le rapport actuel a pour but d’évaluer la couverture réelle du
dépistage cytologique ainsi que l’utilisation d’actes médicaux pour le dépistage et le
6/10/2005
7
suivi ou traitement des cas positifs du dépistage en utilisant un fichier de données
provenant de l’Agence Intermutualiste (AIM), qui rassemble l’information de toutes
les mutuelles en Belgique.
L’étude sous-jacente sera utilisée pour définir les futures stratégies de dépistage du
cancer du col de l’utérus.
Matériel et méthodes
A la demande de l’Institut Scientifique de Santé Publique, un fichier de données
individuelles de plus de 13 millions d’enregistrements a été compilé par l’Agence
Intermutualiste. Ce fichier contient l’information de tous les actes médicaux se
rapportant au dépistage du cancer du col de l’utérus ainsi qu’aux interventions
diagnostiques ou thérapeutiques du col de l’utérus (collecte et interprétation du frottis
vaginal, colposcopies, biopsies cervicales et leur interprétation, chirurgie du col de
l’utérus) de 1996 à 2000.
Résultats
Couverture du dépistage
La couverture du dépistage du cancer du col de l’utérus, mesurée en 2000, était de
59%. Elle est définie en tant que la proportion de la population-cible (femmes entre
25 et 64 ans) ayant eu un frottis du col de l’utérus durant les trois dernières années.
La différence, de couverture de dépistage au niveau des régions, est petite les
couvertures respectives observées étant de : 57% dans la Région Flamande, 58% à
Bruxelles et 61% dans la Région Wallonne. L’augmentation de la couverture
mesurée pour la période 1996-2000 est limitée respectivement à 2%, 5% et 3% dans
la Région Flamande, pour Bruxelles et pour la Région Wallonne. Cette augmentation
limitée dans la Région Flamande est étonnante étant donné le fait que dans cette
région seulement une campagne de dépistage était organisée.
Les différences observées à un niveau géographique plus bas sont importantes : au
niveau provincial en 2000 elles se situent entre 52% (Luxembourg) et 69% (BrabantWallon), tandis qu’au niveau des arrondissements la couverture la plus élevée a été
trouvée à Huy (71%) et la plus basse à Arlon (40% seulement).
Au niveau national, dans le groupe le plus jeune, groupé par tranche de 5 ans, de la
population-cible, c. à d. les femmes de 25 à 29 ans, la couverture est de 64%. La
couverture augmente jusqu’à 67% parmi les femmes de la tranche d’âge de 30 à 39
ans. A partir de l’âge de 40 ans, la couverture diminue graduellement jusqu’à 56%
chez les femmes dans le groupe d’âge de 50 à 54 ans. Ensuite la couverture diminue
plus rapidement. Pour Bruxelles et la Région Wallonne, cette diminution en fonction
de l’âge est moins prononcée.
La couverture évaluée grâce aux fichiers de données des mutuelles est
significativement plus basse que les estimations déduites des enquêtes nationales par
interviews sur la santé. Au niveau national, pour la période 2000-2001 cette
différence est de 11%.
Cette divergence est probablement due à des biais de rapportage, qui selon la
littérature internationale, sont inhérentes aux enquêtes sous forme d’entretiens.
6/10/2005
8
Intervalle de dépistage et tranche d’âge cible
L’intervalle modal de dépistage est de un an. Un laps de temps de trois ans est
observé chez seulement 5% des femmes avec 2 ou plusieurs frottis pendant la période
de l’étude. 10% de tous les frottis exécutés sont prélevés chez des femmes de moins
de 25 ans et 7% chez des femmes âgées de plus de 64 ans.
Collection de frottis par type de médecin
Le pourcentage de frottis préparés par les médecins généralistes est limité et varie
substantiellement par Région. En 2000, un cinquième des frottis était fait par les
médecins généralistes dans la Région Flamande, tandis que seulement 8% et 3%
respectivement à Bruxelles et dans la Région Wallonne. La tendance était
descendante dans la période 1996-2000.
L’utilisation des frottis
Le nombre de frottis utilisé est théoriquement suffisant pour couvrir plus de 100% de
la population-cible sur un laps de temps de trois ans. En 2000, le rapport entre le
nombre d’interprétations de frottis sur l’effectif moyenne de la population-cible, était
de 1.1. Ce rapport devrait être plus élevé si nous étions à même d’ajuster l’effectif de
la population-cible en excluant les femmes qui ont subi une hystérectomie totale.
Néanmoins, seulement 59% sont couvertes avec un ou plusieurs de frottis.
En chiffres absolus: 3 millions de frottis étaient interprétés durant la période 19982000 qui étaient prélevés chez seulement 1.6 millions de femmes de 25-64 ans. 1.1
millions de femmes de cet âge n’ont pas eu de frottis. La surconsommation
d’examens cytologiques du col de l’utérus était de 88%, ce qui signifie que chaque
femme examinée a reçu en moyen 1.88 frottis sur une période de 3 ans. L’excès de
l’utilisation de frottis était élevé dans toutes les régions de la Belgique : il était moins
élevé dans la Région Flamande (86%), et le plus élevé dans la Région Capitale de
Bruxelles (99%), et intermédiaire dans la Région Wallonne (90%).
Une partie de cet excès de frottis prestés est utilisée pour le suivi de lésions cervicales
et/ou le suivi après le traitement. En supposant que 10% de cette surconsommation
est dépensé pour le suivi, on peut estimer qu’environ 400.000 frottis par an sont
exécutés sans contribuer à la couverture de dépistage ou au suivi. En chiffres cela
correspond à un montant estimé de 8.2 millions € que l’INAMI rembourse
annuellement pour les prélèvements et l’interprétation des frottis avec utilité non
prouvée. A ce montant il faut encore ajouter 3.9 millions € pour le dépistage au-delà
de la tranche d’âge cible.
Utilisation de la colposcopie
Un nombre impressionnant de colposcopies sont exécutées en Belgique. Au niveau
national, en moyenne un examen colposcopique est réalisé pour trois frottis prestés.
Dans la Région Wallonne, le rapport du nombre de frottis sur le nombre de
colposcopies est même inférieur à deux, tandis que dans la Région Flamande ce
rapport est de cinq.
Dans certains arrondissements wallons le nombre de frottis examinés est égal au
nombre de colposcopies exécutées. La proportion biopsie/colposcopie est basse (en
moyenne 5%) ce qui est dû à la très grande fréquence d’examens colposcopiques chez
des femmes normales.
Il est clair que la colposcopie n’est pas utilisée comme l’indiquent les guides
nationaux et internationaux. L’exploration colposcopique est indiquée dans le cas
d’une première observation de HSIL ou anomalie glandulaire ou après une deuxième
6/10/2005
9
observation de ASC-US ou LSIL ou après un résultat positif de HPV triage dans des
cas de ASCUS et en suivi après des lésions traitées.
Conclusions générales
La couverture du dépistage du cancer du col d’utérus en Belgique, définie comme la
proportion de femmes dans la tranche d’âge entre 25 et 64 ans ayant reçu un frottis les
trois dernières années, n’est que de 59%. Pourtant, le nombre de frottis prélevés est
théoriquement suffisant pour couvrir toute la population-cible.
L’estimation de la couverture de dépistage du cancer du col de l’utérus venant
d’enquêtes par interviews, devrait être interprétée avec prudence, étant donné le
risque inhérent de surestimation.
Le prélèvement est surtout une activité de gynécologues, ce qui est quasiment une
exclusivité en Wallonie.
La colposcopie est trop souvent exécutée simultanément au frottis, alors que
l’indication clinique principale de la colposcopie est réservée aux femmes ayant eu
des anomalies cytologiques.
Une réduction structurelle de la sur-utilisation et une redistribution pourrait améliorer
la couverture et en même temps la qualité ce qui pourrait potentiellement résulter en
une augmentation des années de vie sauvées, sans augmentation du financement
public.
Les mesures prévues dans les Recommandations sur le Dépistage du Cancer du
Conseil d’Europe devraient être exécutoires et universelles. Des actions isolées et
régionales comme celles entreprises dans le programme flamand de dépistage n’ont
pas fourni le résultat désiré, parce qu’elles étaient basées sur une participation
volontaire sans compensation financière pour les médecins impliqués.
Les ministères de la santé publique du gouvernement fédéral et des gouvernements
des communautés ainsi que les représentants des sociétés scientifiques devraient se
réunir au plus vite pour définir une politique de dépistage rationnelle, efficient et
basée sur de l’évidence scientifique.
4. Samenvatting en commentaar
Achtergrond
In 2003 heeft de Europese Raad alle lidstaten van de Unie aanbevolen om
georganiseerde screening voor baarmoedershalskanker aan te bieden aan alle vrouwen
uit de doelgroep. Een optimale dekking of couverture van de doelbevolking is de
sleutel tot een succesvolle opsporingscampagne.
Tot voor kort kon de dekkingsgraad voor screening in België enkel geëvalueerd
worden aan de hand van enquêtes. Het huidige rapport heeft tot doel de reële
6/10/2005
10
couverture te evalueren alsook de hoeveelheid van medische prestaties te meten
bedoeld voor opsporing, follow-up of behandeling van letsels van de baarmoederhals.
Hiervoor werd een databestand gebruikt samengesteld door het Inter-Mutualistisch
Agentschap (IMA).
Deze studie zal gebruikt worden om strategieën te bepalen voor een toekomstig
cervixkankeropsporingsprogramma.
Materiaal en methoden
Op aanvraag van het Wetenschappelijk Instituut Volksgezondheid werd door het
Inter-Mutualistisch Agentschap, een databestand samengesteld met meer dan 13
miljoen patiëntendossiers. Dit bestand bevatte informatie van alle medische prestaties
geleverd tussen 1996 en 2000, die verband houden met baarmoederhalskankerscreening, diagnostische en therapeutische interventies op de baarmoederhals, zoals
afname en interpretatie van cervixuitstrijkjes, colposcopieën, biopsieën van de
baarmoederhals en de interpretatie ervan, alsook chirurgische ingrepen op de
baarmoeder(hals).
Resultaten
Dekking van de screening.
In 2000 was de screeningcouverture of dekkingsgraad 59%. Onder dekkingsgraad
verstaan we het percentage van de doelbevolking van vrouwen tussen 25 en 64 jaar,
bij wie een cytologisch onderzoek van een cervixuitstrijk heeft plaatsgevonden
minder dan drie jaar geleden. Het verschil in couverture tussen de gewesten was
gering: 57% in het Vlaamse Gewest, 58 % in Brussel en 61% in het Waalse Gewest.
De stijging van de couverture, gemeten over de periode 1996-2000, was beperkt en
bedroeg respectievelijk 2%, 5% en 3% voor het Vlaamse Gewest, voor Brussel en
voor het Waalse Gewest. Deze beperkte toename in het Vlaamse Gewest is
opmerkelijk aangezien enkel in dit gewest een opsporingscampagne georganiseerd
werd over het ganse grondgebied.
De verschillen op lagere geografische niveaus zijn aanzienlijk: op provinciaal vlak
schommelde de couverture, bereikt in het jaar 2000, tussen 52% (Luxemburg) en 69%
(Waals-Brabant), terwijl op niveau van de arrondissementen de hoogste dekking werd
aangetroffen in Huy (71%) en de laagste in Aarlen (slechts 40%).
In de jongste 5-jaar leeftijdsgroep, vrouwen van 25-29 jaar oud, was de dekkingsgraad
64% op nationaal niveau. Deze verhoogde tot 67% bij vrouwen tussen 30 en 39 jaar.
Vanaf de leeftijd van 40 jaar verminderde de dekking geleidelijk tot 56% in de
leeftijdsgroep van 50-54 jarigen. Vervolgens vermindert de dekking sneller. In
Brussel en het Waalse Gewest is deze vermindering in de oudere leeftijdsgroep
minder uitgesproken.
De dekking berekend op basis van gegevens uit het IMA bestand is aanzienlijk lager
dan de schattingen afgeleid uit de nationale gezondheidsenquête. Op nationaal vlak
bijvoorbeeld is dit verschil, voor de periode 2000-2001, 11%. Dit verschil is
waarschijnlijk te wijten aan rapporteringsbiases, die volgens de internationale
literatuur inherent zijn aan enquêtes aan de hand van interviews.
6/10/2005
11
Screeningsinterval en leeftijdsdoelgroepen
Het modaal screeningsinterval is één jaar. Een tijdsspanne van 3 jaar wordt
waargenomen bij 5% van de vrouwen met 2 of meer uitstrijkjes in de studieperiode.
Tien procent van alle geïnterpreteerde uitstrijkjes werden genomen bij vrouwen
jonger dan 25 jaar en 7% bij vrouwen ouder dan 64 jaar.
Afname van uitstrijkjes per type van arts
Het aandeel van huisartsen in de afname van cervixuitstrijkjes is beperkt en varieert
sterk tussen de gewesten. In 2000, werd één vijfde van de uitstrijkjes in het Vlaams
Gewest afgenomen door huisartsen, en respectievelijk 8% en slechts 3% in Brussel en
het Waals Gewest. Er was een dalende trend over de periode 1996-2000.
Intensiteit van het gebruik van uitstrijkjes
Het aantal gebruikte uitstrijkjes is in theorie voldoende om meer dan 100% van de
doelbevolking te dekken over een tijdspanne van drie jaar. In 2000 was de
verhouding tussen het aantal geïnterpreteerde uitstrijkjes t.o v de gemiddelde grootte
van de doelbevolking 1.1. Deze verhouding zou hoger moeten zijn indien we de
grootte van de doelbevolking konden corrigeren door exclusie van vrouwen die een
volledige hysterectomie hebben ondergaan. Desondanks is slechts 59% gedekt met
één of meer uitstrijkjes.
In absolute cijfers: in de periode 1998-2000 werden 3 miljoen uitstrijkjes
geïnterpreteerd, die genomen werden bij 1.6 miljoen vrouwen. 1.1 miljoen vrouwen
werden niet onderzocht. De overconsumptie van cytologische baarmoederhalsonderzoeken was 88%, hetgeen betekent dat iedere gescreende vrouw 1.88 uitstrijkjes kreeg
in een periode van 3 jaar. Het overmatig gebruik van uitstrijkjes was hoog in alle
delen van België: het was iets minder hoog in het Vlaamse Gewest (86%), het hoogst
in het Hoofdstedelijk Gewest (99%), en intermediair in het Waalse Gewest (90%).
Het spreekt vanzelf dat een deel van deze “overconsumptie” aangewend wordt voor
de follow-up van vrouwen met eerder vastgestelde baarmoederhalsletsels. In de
veronderstelling dat 10% van de “overconsumptie” aangewend wordt voor opvolging
van vroegere abnormaliteiten, kunnen we inschatten dat jaarlijks 400,000 uitstrijkjes
afgenomen worden die niet bijdragen tot de screeningscouverture of follow-up. Dit
komt overeen met een geschat bedrag van 8,2 miljoen € per jaar die terugbetaald
worden door het RIZIV voor afname en interpretatie van uitstrijkjes met beperkt nut.
Dit bedrag wordt verder opgedreven met 3,9 miljoen € voor opsporing buiten de
doelgroep.
Gebruik van colposcopie
Een indrukwekkend aantal colposcopieën worden in België uitgevoerd. Op nationaal
vlak wordt er gemiddeld één colposcopie uitgevoerd op drie uitstrijkjes. In het
Waalse Gewest is de verhouding van het aantal uitstrijkjes tot het aantal
colposcopieën zelfs minder dan twee, terwijl in het Vlaamse Gewest deze verhouding
vijf is. In bepaalde Waalse arrondissementen is het aantal onderzochte uitstrijkjes
gelijk aan het aantal uitgevoerde colposcopieën.
De verhouding aantallen
biopsieën/aantallen colposcopieën is laag (gemiddeld 5%) hetgeen te wijten is aan het
zeer hoge aantal colposcopieën bij normale vrouwen.
Het is duidelijk dat de colposcopie niet gebruikt wordt volgens de indicaties in
nationale of internationale richtlijnen.
Colposcopisch onderzoek is aangewezen in gevallen van een eerste observatie van
HSIL of glandulaire letsels of na een tweede observatie van ASC-US of LSIL of na
6/10/2005
12
een positief HPV triage resultaat en in follow-up na behandelde letsels [Arbyn, 1997 ;
Arbyn 2004a; Arbyn 2004b; BSCC, 2004; Wright, 2002].
Algemene conclusies
De couverture betreffende baarmoederhalskankeropsporing in België, gedefinieerd als
het percentage van vrouwen in de leeftijdsgroep van 25 tot 64 jaar die een uitstrijkje
hebben laten nemen gedurende de laatste drie jaar, is slechts 59%. Nochtans is het
aantal onderzochte uitstrijkjes in theorie voldoende om de ganse doelbevolking te
dekken.
Schattingen over de couverture afgeleid van enquêtes moeten met voorzichtigheid
geïnterpreteerd worden, gezien het inherente risico op overschatting.
Afname van cervix uitstrijkjes is voornamelijk een activiteit van gynaecologen, in
Wallonië is dit bijna een exclusiviteit.
De colposcopie wordt te vaak uitgevoerd, in vele gevallen wellicht tezelfdertijd als
het routine uitstrijkje terwijl de belangrijkste klinische indicatie is nader onderzoek bij
vrouwen met een vroegere cytologische afwijking.
Een structurele vermindering van de overconsumptie en herinvestering in
couvertureverhoging en kwaliteitsverbetering kan er potentieel toe bijdragen dat er
meer levensjaren gered worden, zonder de nood aan meer publieke middelen.
Maatregelen voorzien in de aanbevelingen voor kankeropsporing van de Europese
Raad zouden bindend en universeel moeten zijn. Regionaal geïsoleerde acties zoals
deze ondernomen in het kader van de Vlaamse opsporingscampagnes, hebben niet tot
het gewenste resultaat geleid omdat ze gebaseerd waren op vrijwillige deelname
zonder financiële compensatie voor het betrokken beroepsgroepen.
De federale overheidsdiensten verantwoordelijk voor volksgezondheid en de
ministeries van de gemeenschappen zouden zo snel mogelijk moeten samenkomen
alsook de vertegenwoordigers van de wetenschappelijke artsenverenigingen om een
rationele en efficiënte strategie voor de preventie van baarmoederhalskanker te
bepalen gebaseerd op wetenschappelijke evidentie.
5. Introduction
Currently, in Belgium, still 600 to 700 women get cervical cancer each year. In fact
the number of cervical cancer cases at the national level is not precisely known since
only for the Flemish Region a functional cancer register exists. According to that
register 411 Flemish women got the diagnosis of cancer of cervix in 2000 [Vlaams
Kankerregistratienetwerk, 2004].
Also the rate of mortality from cervical is not known precisely due to death cause
certification problems. It often happens that for women dying from uterine cancer the
6/10/2005
13
exact topography, cervix uteri or corpus uteri, is not identified. According to a recent
trend analysis, where a tentative solution for this certification problem was proposed,
it was estimated that in the nineties between 300 and 350 died from cervical cancer in
Belgium each year [Arbyn, 2002]a.
By well-organised cytological screening, the incidence of cervical cancer can be
reduced to a small residual level [IARC expert team, 2005]. This is the key
recommendation of the Council of Europe to all EU member states (Council of the
EU, 2003]. It was already shown to be true in certain Nordic countries [Laara, 1987]
and more recently also in the United Kingdom [Peto, 2004] and Norway [Nygard,
2002]. In Belgium, screening is essentially opportunistically organised, resulting in a
high level of over screening, and a heterogenous quality. In spite of the large amount
of Pap smears consumed yearly, the impact was lower than observed in the Nordic
countries [Arbyn, 1999; 2000; 2002]. Again the real impact of the Belgian screening
on incidence and mortality cannot be assessed exactly by lack of performant health
information systems.
One of the main determinants of success of a screening programme is an optimal
attendance of the target population. Until recently, screening coverage could only be
measured from interview surveys [Arbyn, 1997; Arbyn 2000; Demarest, 2002].
In Belgium the whole population is covered by an obligatory health insurance system.
Administrative data registered by health insurance agencies from their members
constitute a virtually complete source of information on health services delivered to
patients. In 2003, the Scientific Institute of Public Health received a data file of more
than 13 million records pertaining to medical acts related to cervical cancer screening
and other diagnostic and therapeutic interventions on the uterine cervix. The data file
was compiled by the Inter-Mutualistic Agency from all seven health insurance
agencies. Analysis of this data file enabled us to assess to real cervical cancer
screening status and the intensity of the use of Pap smear testing, colposcopies,
cervical biopsies, local surgery on the cervix and hysterectomies.
The purpose of this report is to describe the current situation with respect to cervical
cancer screening in Belgium, at national, regional, provincial and district
(arrondissement) level. The conclusions will used to propose new strategies for
organised screening in whole of Belgium, in concertation with the Communities.
This report was elaborated within the framework of a convention between the General
Directorate of Primary Health Care of the Federal Public Service of Health, Food
Chain Security and Environment and the Scientific Institute of Public Health. This
budget completed funding from Europe Against Cancer.
a
Readers interested in the evolution of the cervical cancer mortality trend are invited to consult the
website pages of the unit cancer screening at the Department of Epidemiology of the Scientific Institute
of Public Health, available at; www.iph.fgov/epidemio/epien/prog2.htm.
6/10/2005
14
6. Material and methods
In 2002, a demand was addressed by the Scientific Institute of Public Health to the
IMA/AIM (Intermutualistisch Agentschap/Agence Inter-Mutualiste)a for individual
patient data concerning medical acts involving the uterine cervix for the last five
available years. The IMA/AIM collects data on all medical acts covered by the health
insurance, which is obligatory in Belgium. End 2003, a dataset was received
containing more than 13 million records from medical acts, which took place between
1996 and 2000.
The data set contained the following variables: anonymous individual ID code, age,
date of the act, residence of the woman and type of the medical act. Several data
details were truncated to reduce the risk of obtaining unique data in the cells of cross
tables. The age was converted into five-year age group at the exception of the group
0-14 and the 75+ age groups, which were each grouped into one category. The
calendar date of the act was restricted to the year of performance. The residence was
converted into the district (arrondissement). The code, district=99, was reserved for
Belgian insured women, not residing in Belgium. Sometimes, district code=0 was
used in case of rare combinations to avoid identification of women. The medical
codes were fused into groups as explained in Table 1. All surgical interventions
indicated as group 8, involved total hysterectomies or removal of the cervix. In this
report this group of interventions are indicated as total hysterectomies. No
information on subtotal hysterectomy was obtained.
We assumed that all acts involving the age group 0-14 years were performed on girls
aged 10 to 14 years. Similarly, all acts involving women of 75 years and older were
assumed to be performed in women aged 75-79.
The individual patient data were aggregated over 5-year age groups, calendar years
and district and linked with corresponding population data published by the National
Institute of Statistics. This allowed computing proportions, rates or ratios such as the
cervical cancer screening coverage, the excess Pap smear use, hysterectomy incidence
rates and the ratio of the number of acts / number of concerned women.
Table 1. Codes of medical acts included in the data set.
Group
1
2
3
4
5
6
7
8
Codes
114030-114041
149612-149623
588350-588361
431955-431966
432110-432121
588276-588280
432294-432305
431270-431281
431314-431325
431336-431340
431351-431362
431491-431502
432154-432165
Description
Collection of cellular material from the uterine cervix, by a GP
Collection of cellular material from the uterine cervix, by a specialist
Microscopic interpretation of a Pap smear
Colposcopy
Biopsy from the uterine cervix
Pathological examination of a biopsy
Conisation
Total abdominal hysterectomy
Total vaginal hysterectomy
Radical hysterecctomy (Wertheim)
Total hysterectomy & pelvic lymphadenectomy
Cervix amputation
Removal of residual cervix
a
“Inter-Mutualistic Agency” is used as English term for the Dutch term “Inter-Mutualistisch
Agenschap, abbreviated as IMA or for the French term “Agence Inter-Muturaliste”, abbreviated as
AIM. In this report we will use the Dutch abbreviation “IMA”.
6/10/2005
15
For an overview of computed indicators we refer to the glossarium in chapter 10.
When coverages over multi-year intervals were computed, women were assumed to
belong to the age group and district at the last medical act concerned.
Statistical methods
The statistical package Stata (version 7 and 8) [Stata Corp, Stata Corporation, Texas,
US] was used for processing and statistical analysis. Analysis consisted in monovariate and multivariate tabulations, histograms, scatter and line plots. We performed
some ordinary least square linear regressions using population frequency weights and
Poisson regressions, with the logarithm of the population as offset. The strength of
correlations was assessed using Pearson’s or Spearman’s rank correlation coefficients.
Four different geographical levels were considered for analysis: the whole of
Belgium, the 3 Regions (Flemish, Walloon Region and Brussels); the 11 provinces
and the 43 districts or arrondissments. Geographical contrasts of cervical cancer
screening coverage and hysterectomy incidence were mapped using choropleth maps.
Indirect standardisation was sometimes used to control for differences in age
composition.
Given the large populations involved, it was generally inappropriate to report 95%
confidence intervals, since upper and lower bound are nearly equal.
For computation of cumulative incidence until a given age k, the following formula
was applied: Cumulative Incidence = 1 – Πk(1-eai . ∆T), where Π stands for cumulative
product, ai for age specific incidence and ∆T for the amplitude of the age categories
[Kleinbaum, 1982].
6/10/2005
16
7. Results
The received data file contained more than 13 million records. The distribution by
type of medical act is shown in Table 2.
Table 2. Number of individual acts pertaining to Pap smears, colposcopies, biopsies and surgical
interventions on the uterine cervix performed in Belgium between 1996 and 2000.
Act
Collection of Pap smear by GP
Collection of Pap smear by specialist
Total hysterectomy or cervix amputation
Colposcopy
Biopsy from cervix
Conisation of cervix
Pathological examination of a biopsy
Interpretation of Pap smear
Total
Frequency
815,782
4,435,416
84,564
2,009,956
106,138
24,445
72,757
5,734,201
13,283,259
Percent
6.14
33.39
0.64
15.13
0.8
0.18
0.55
43.17
100
In this report, we will address subsequently the cytological interpretation of Pap
smears (subchapter 7.1); the type of physician who takes Pap smears (subchapter 7.2);
colposcopy and taking cervical biopsies (subchapter 7.3); hysterectomy (7.4) and
finally the interpretation of Pap smears in women after a recent hysterectomy
(subchapter 7.5).
7.1. Papanicolaou smears of the uterine cervix
7.1.1.
National level
In total, about 5.7 million smears were interpreted in the period 1996-2000 or a yearly
average of over 1.1 million (see Table 3). The total number of smears rose almost
linearly (R2=0.95) with an increase of 45 000 additional examinations per year. This
corresponds to an average yearly increase of 4.4%.
Table 3. Total number of Pap smears interpreted by year (Belgium, 1996-2000).
Year
1996
1997
1998
1999
2000
Total
Frequency
1,042,821
1,095,337
1,171,347
1,201,152
1,223,544
5,734,201
Percent
18.19
19.10
20.43
20.95
21.34
100.00
Cumulative
percent
18.19
37.29
57.72
78.66
100.00
The distribution per 5-year age group is shown in Table 4. The proportion of smears
from women in the 25 to 64 year target age group was 82.3%. Respectively 10.0%
and 6.8% of Pap smears were taken from women who were younger or older.
6/10/2005
17
Table 4. Total number of Pap smears by 5-year age group (Belgium, 1996-2000).
Age group
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
≥ 75
Total
Frequency Percent
4,260
0.07
123,802
2.16
442,454
7.72
692,356 12.07
790,245 13.78
780,372 13.61
694,605 12.11
625,377 10.91
537,123
9.37
357,656
6.24
294,414
5.13
203,103
3.54
114,270
1.99
74,164
1.29
5,734,201 100.00
Cumulative
percent
0.07
2.23
9.95
22.02
35.8
49.41
61.53
72.43
81.8
88.04
93.17
96.71
98.71
100.00
The time interval between two successive smears for women having at least two
smears between 1996 and 2000 is shown in Table 5 and in Figure 1.
A detailed picture of the real interval was not possible since the dates of smears were
truncated at the calendar year. Nevertheless, a clear modus at the 1-year interval is
observed (in 61% of the cases). In 12% there was less than 1 year between successive
smears.
Table 5. Distribution of the time interval between two successive smears (rounded at one year),
(Belgium, 1996-2000).a
Interval
(years)
0
1
2
3
4
Missing
Total
Frequency
426,348
2,115,644
726,865
173,061
39,661
2,251,615
5,733,194
(missing included)
Cumulative
Percent
frequency
7.44
7.44
36.90
44.34
12.68
57.02
3.02
60.03
0.69
60.73
39.27
100.00
100.00
(missing excluded)
Cumulative
Percent
frequency
12.25
12.25
60.77
73.01
20.88
93.89
4.97
98.86
1.14
100.00
100.00
a
One fictitious woman, age=0 & arrondissement=0, with 1007 smears was excluded. This was
probably due to non-unique anonymistion of the identifier.
6/10/2005
18
Interval between successive cervical smears
Belgium, 1996-2000
.6
Fraction
.4
.2
0
0
1
2
Interval (years)
3
4
Figure 1. Histogram of the distribution of average time interval between two successive smears,
(Belgium, 1996-2000).
In total 2,251,615 women had at least one smear. For their first smear in the database
no interval can be calculated. The year that their last smear took place is relevant to
complete the information on the screening interval and is shown in Table 6.
Table 6. Distribution of the years where the last smear of a woman took place (Belgium, 1996-2000). a
Year Frequency
Percent
1996
123,250
5.47
1997
173,278
7.70
1998
281,210 12.49
1999
547,053 24.30
2000
1,126,824 50.05
Total
2,251,615 100.00
Cumulative
percent
5.47
13.17
25.66
49.95
100.00
The distribution of the total number of smears per women within the period 19962000 is displayed in Table 7.
a
One fictitious woman, age=0 & arrondissement=0, with 1007 smears was excluded. This was
probably due to non-unique anonymisation of the identifier.
6/10/2005
19
Table 7. Distribution of the total number of smears taken from the same women (Belgium, 19962000). a
Number of
smears per
women
Frequency
1
712,810
2
549,112
3
426,621
4
306,331
5
183,311
6
44,324
7
14,926
8
6,628
9
3,699
10
1,971
11
951
12
361
13
193
14
96
15
77
16
60
17
34
18
22
19
13
20
75
Total
2,251,615
Percent
31.66
24.39
18.95
13.60
8.14
1.97
0.66
0.29
0.16
0.09
0.04
0.02
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
100.00
Cumulative
percent
31.66
56.05
74.99
88.6
96.74
98.71
99.37
99.66
99.83
99.92
99.96
99.97
99.98
99.99
99.99
99.99
100.00
100.00
100.00
100.00
6/10/2005
20
7.1.1.1 Use of cervical cytology at yearly intervals
The evolution of the one-year screening coverage and the consumption of Pap smears
over time is illustrated in Table 3 and in Figure 2. The coverage (full line curve, red)
increased with 3.5% in the first three years: from 29.9% to 33.4 %. Over the last two
years the increase was only 0.6%. The #smears/#women ratio (green curve,
interrupted line) was 2 to 3% higher than the screening coverage.
Table 8. Consumption of Pap smears and one-year screening coverage between 1996 and 2000 for
women between 25 and 64 years old.
Number of
Number of women Number of
smears
women screened
25-64 years
2,694,738
873,359
807,018
2,703,970
915,851
846,669
2,710,432
975,223
906,196
2,717,013
996,947
924,987
2,723,354
1,010,768
930,323
Year
1996
1997
1998
1999
2000
1-year #smears/#women
coverage
ratio
29.9%
0.32
31.3%
0.34
33.4%
0.36
34.0%
0.37
34.2%
0.37
Consumption of Pap smears; Coverage
Belgium, 1996-2000, women 25-64 year
# smears/# women ratio
Coverage, 1-year interval
.4
.3
.2
.1
0
1996
1998
Year
2000
Figure 2. Evolution of the one-year screening coverage and of the #smears/#women ratio between
1996 and 2000.
The variation of the screening coverage considering a one-year interval by five-year
age group is shown as a red curve for four years in Figure 3. The coverage increases
rapidly to reach a maximum of more than 35 % in the age group 30-34 years. From
this age the coverage declines slowly until around 30% in the age group 50-54 years.
The coverage decreases rapidly after the age of 55. The four graphs look very similar.
Nevertheless, the screening coverage changed somehow over time. This evolution
becomes obvious by observing Figure 4. Between 1996 and 1998, the coverage
increased with 2 to 4 % in all age groups older than 20 and younger than 70. The
further increase between 1998 and 2000 was minor.
6/10/2005
21
Cytological screening for cervical cancer
Belgium, 1996
Belgium, 1997
.4
.4
.3
.3
.2
.2
.1
.1
0
0
10
20
30
40
age
50
60
10
70
Belgium, 1998
20
30
40
age
50
60
70
Belgium, 1999
.4
.4
.3
.3
.2
.2
.1
.1
0
0
10
20
30
40
age
50
60
70
20
10
30
40
age
50
60
70
Figure 3. Variation of the screening coverage and the # smears / # women ratio considered over a 1year interval according to 5-year age group and by year.
1996
2000
1998
1-year coverage
.4
.3
.2
.1
0
10
20
30
40
50
Age group
60
70
Figure 4. Change in the relation between one-year coverage and age group over the years 1996, 1998
and 2000.
6/10/2005
22
7.1.1.2 Use of cervical cytology at 3-year intervals
The consumption of Pap smears and the coverage built up over three-year periods is
illustrated in Table 9. The number of smears used was sufficient to cover more than
100% of the target population since the ratio of the number of smears taken in women
between 25 and 64 years surpasses the average number of women of that age living in
Belgium (#smears/#women ratio > 1). Nevertheless the 3-year coverage was only 56
to 59%. Between 84 and 88% of smears do not contribute to the population coverage,
since they are used to screen women who are already screened within the previous 3year period.
Table 9. Consumption of Pap smears, three-year screening coverage and excess use of between 1996
and 2000 for women between 25 and 64 years old.
Period
1996-1998
1997-1999
1998-2000
Mean female
population
(25-64 years)
2,703,047
2,710,473
2,716,933
Number of Number of
smears women screened
3-year
coverage
takena
<3years ago
2,778,474
1,511,864
55.9%
2,901,180
1,563,895
57.7%
2,995,380
1,591,255
58.6%
#smears/
#women
ratio
1.03
1.07
1.10
Excess use
83.8%
85.5%
88.2%
The 3-year screening coverage in 2000 reached a maximum of 67% in women of 3034 years. It decreased gradually until 56% in the 50-54 year age group. Thereafter
the coverage declined more rapidly. In the age group 20 to 54 years, more smears
were taken than what was necessary to cover 100% of the population (green curve
surpasses the horizontal line through unity). The age relation for the years 1998 and
1999 were similar (data not shown).
Belgium, 2000
1.4
1.2
1
.8
.6
.4
.2
0
10
20
30
40
Age
50
60
70
Figure 5. Variation of the screening coverage and the # smears / # women ratio considered over a 3year interval according to 5-year age group (Belgium, 2000).
a
Only smears taken in the age group 25-64 years are considered.
6/10/2005
23
The excess Pap smear use as a relation of age is plotted in Figure 6. The excess is
higher than 80% from the age of 25 to 69 and reaches a maximum of 98% in the age
50-54.
Extra smears/screened woman
1.5
1
.5
0
10
20
30
40
Age group
50
60
70
Figure 6. Excess Pap smear use as a function of age group (Belgium, 2000).
7.1.1.3 Use of cervical cytology at a 5-year interval
The consumption of Pap smears and the coverage built up over five-year periods is
illustrated in Table 9. In total 4.8 million smears were interpreted between 1996 and
2000 from 1.8 million women in the age group 25-64. The number of smears used
was sufficient to cover 179% of the target population. Nevertheless the 5-year
coverage was only 67%. Screened women receive on average 1.66 additional smears
they do not need if a 5-year screening policy should be adapted.
Table 10. . Consumption of Pap smears, five-year screening coverage and excess use of between
1996 and 2000 for women between 25 and 64 years old.
Period
1996-2000
Mean female
population
(25-64 years)
2,709,901
Number of
smears taken
4,841,250
Number of women
screened <5years
5-year
ago
coverage
1,822,749
67.3%
#smears/
#women
ratio
Excess use
1.79
165.6%
The 5-year screening coverage in 2000 reached a maximum of 75 % in women of 3039 years (see Figure 7). It decreased gradually until 67 % in the 50-54 year age
group. Thereafter the coverage declined more rapidly. In the age groups 30-39, the
amount of smears used could cover the population twice.
6/10/2005
24
Belgium, 2000
2.5
2
1.5
1
.5
0
10
20
30
40
Age
50
60
70
Figure 7. Variation of the screening coverage and the # smears / # women ratio considered over a 5year interval according to 5-year age group (Belgium, 2000).
6/10/2005
25
7.1.2.
Regional level
7.1.2.1 Use of cervical cytology at yearly intervals
The evolution of 1-year screening coverage and Pap smear use over the period 1996
to 2000, for 25-64 year old women is shown in Figure 8 for the three Regions. The
differences between the Regions were small. In 2000, the 1-year coverage was
highest in the Walloon Region (35.8%), followed by the Capital Region Brussels
(35.6%). The coverage was lowest in the Flemish Region (32.9%).
Table 11. Consumption of Pap smears and one-year screening coverage between 1996 and 2000 for
women between 25 and 64 years old, by Region.
Number of women Number of Number of
1-year
#smears/
25-64 years
smears women screened coverage #women ratio
1,577,602
489,210
455,104
28.8%
0.31
1,583,165
516,449
479,727
30.3%
0.33
1,585,616
552,626
516,370
32.6%
0.35
1,588,142
560,163
524,611
33.0%
0.35
1,589,783
569,876
523,227
32.9%
0.36
Brussels
252,392
82,962
73,932
29.3%
0.33
253,543
86,647
77,905
30.7%
0.34
254,738
93,054
84,278
33.1%
0.37
255,319
96,019
87,430
34.2%
0.38
256,912
101,319
91,500
35.6%
0.39
Walloon
864,744
298,847
275,875
31.9%
0.35
Region
867,262
310,440
286,963
33.1%
0.36
870,078
327,247
303,489
34.9%
0.38
873,552
338,604
311,016
35.6%
0.39
876,659
337,560
313,795
35.8%
0.39
Year
1996
1997
1998
1999
2000
1996
1997
1998
1999
2000
1996
1997
1998
1999
2000
Region
Flemish
Region
Consumption of Pap smears; Coverage
Belgium by Region, 1996-2000, women 25-64 year
Flemish Region
Brussels
.4
.3
.2
.1
0
1996
Walloon Region
1998
2000
.4
.3
.2
.1
0
1996
1998
2000
Year
1996 and 200, by Region.
6/10/2005
26
There was a nearly linear yearly increase in coverage over the 5-year period (R2=0.85
with a relative yearly increase (RR) of 1.011 for the Flemish Region; R2=0.99 and
RR= 1.016 for Brussels; R2=0.97 and RR=1.013 for the Walloon Region). In 2000 in
the Flemish Region, no further increase of the coverage was observed, whereas the
#smears/#women ratio still continued to increase. In the Walloon Region, on the
other hand, the coverage still increased in 2000 whereas relatively less smears were
used in comparison with 1999. The #smears/#women ratio was 2 to 4 % higher than
the 1-year coverage.
Figure 9 illustrates the relation between age and Pap smear consumption and the 1year coverage for the year 2000. A maximum coverage is reached in the 30-34 year
age group, respectively at 39%, 37% and 40% in the Flemish, Capital and Walloon
Region. Remarkable is that the coverage remains at a plateau until the age of 54 years
in Brussels and that it only decrease slowly until this age in the Walloon Region.
However, in the Flemish Region, the coverage drops more quickly. The coverage in
age group 50-54 year is respectively 31, 32% and 37% in the Flemish, Capital and
Walloon Region. From the 55-59 year age group the coverage declines quickly in the
Flemish and Walloon region. From this age, the coverage is highest in Brussels.
Belgium by Region, 2000
Flemish Region
Brussels
.4
.2
0
10
Walloon Region
20
30
40
50
60
70
.4
.2
0
10
20
30
40
50
60
70
Age group
Figure 9. Variation of the screening coverage and the # smears / # women ratio considered over a 1year interval according to 5-year age group, by Region (Belgium, 2000).
6/10/2005
27
Data concerning the consumption of Pap smears and the screening coverage
considered over a 3-year interval are documented in Table 12 for the three Regions.
The coverage was highest in the Walloon Region (60.9% in 2000), which was
respectively 2.3% and 2.5% higher than in the Capital and Flemish Region. The
increase over the whole period was 2.4% for the Flemish Region, 4.6% for Brussels
and 2.5% for the Walloon Region. At the exception of the first period in the Flemish
Region, the #smear/#women ratio was always higher than one. The excess use of Pap
smears among screened women was always above 80%. It was highest in the Capital
Region (between 1.97 and 1.99) and was lowest in the Flemish Region (between 1.80
and 1.88).
Table 12. Consumption of Pap smears, three-year screening coverage and overuse between 1996 and
2000 for women between 25 and 64 years old, by Region.
Number of Number of
smears women screened 3-year
coverage
taken
<3years ago
1,569,687
870,520
55.0%
1,639,685
900,538
56.8%
1,691,932
911,761
57.4%
#smears /
#women
ratio
0.99
1.03
1.07
Excess
use
80.3%
82.1%
85.6%
Period
Region
1996-1998 Flemish
1997-1999 Region
1998-2000
Mean female
population
(25-64 years)
1,582,128
1,585,642
1,587,847
1996-1998 Brussels
1997-1999
1998-2000
253,557
254,534
255,656
264,611
277,884
292,582
134,397
141,204
147,381
53.0%
55.5%
57.6%
1.04
1.09
1.14
96.9%
96.8%
98.5%
1996-1998 Walloon
1997-1999 Region
1998-2000
867,361
870,296
873,429
944,176
983,611
1,010,866
506,947
522,153
532,113
58.4%
60.0%
60.9%
1.09
1.13
1.16
86.2%
88.4%
90.0%
The differentiation of the 3-screening coverage and Pap smear consumption by age is
illustrated in Figure 10 for the last period. A maximum coverage was reached at the
age of 30-34 years. This was 67% in the Flemish Region, 61% in Brussels and 68%
in the Walloon Region. The decline in coverage by age after 35 years was slow until
the 50-54 year age group in the last 2 Regions. In the 50-54 year age group the 3coverge was respectively 53%, 58% and 62% for the Flemish, Capital and Walloon
Region. After the age of 50 the coverage was highest in Brussels.
In the Flemish Region and in Brussels the #smears/#women ratio was higher than 1
between the age 25 to 59. In the Walloon Region it was above unity in the 25-54 year
age group.
The excess smear use as a function of age is plotted in Figure 11. The excess is
maximal in the age groups 50-59. The excess use is highest in Brussels and lowest in
the Flemish Region.
6/10/2005
28
Belgium, 2000, by Region
Flemish Region
Brussels
1.4
1.2
1
.8
.6
.4
.2
0
10
Walloon Region
20
30
40
50
60
70
1.4
1.2
1
.8
.6
.4
.2
0
10
20
30
40
50
60
70
Age
Figure 10. Variation of the screening coverage and the # smears / # women ratio considered over a 3year interval according to 5-year age group (Belgium, by Region, 2000).
Excess Pap smear use
Belgium, by Region, 2000
Number of smears/screened woman
Flemish Region
Walloon Region
Brussels
1.5
1
.5
0
10
20
30
40
50
Age group
60
70
Figure 11. Excess Pap smear use as a function of age group (Belgium, by Region, 2000).
6/10/2005
29
7.1.3.
Provincial level
We limit the analysis at the provincial level at screening over 3-year periods.
Statistics concerning screening at one-year intervals can be produced on demand.
The screening intensity and the evolution of the three-year coverage are documented
in Table 13. The linear increase in coverage is the slope of a linear regression
equation obtained by fitting an ordinary least square linear regression using year as a
numerical variable and including an interaction term (Province*year). The last year
of the respective 3-year period was used to identify the “year”-variable. The mean
female population computed over the 3-years was used as a frequency weight. The
slope has to be interpreted as following: the fitted coverage in Antwerp in 1998 was
55.0%, one year later, in 1999, it was 1.3% higher or 56.3%, and two years later, in
2000, it was 2.6% higher than in 1998 or 57.6%. The coverage was increasing in all
provinces, but this increase varied from only 0.1% per year in Limburg to 2.3% per
year in Brussels. The observed and fitted coverage is plotted for each province in
Figure 12.
6/10/2005
30
Table 13. Consumption of Pap smears, three-year screening coverage, excess use and increase in
linear trend of the coverage between 1996 and 2000 for women between 25 and 64 years old, by
Province.
Period
Province
1996-1998 Antwerp
1997-1999
1998-2000
Number of
Mean female Number of
women
#smears
Linear
screened
population
smears
3-year /#wome
increase
(25-64 years)
taken
<3years ago coverage n ratio Overuse (coverage)
437,972
436,889
241,029
55.0%
1.00
1.81
1.3%
438,520
451,965
248,275
56.6%
1.03
1.82
438,639
462,971
252,749
57.6%
1.06
1.83
1996-1998 Brussels
1997-1999
1998-2000
253,557
254,534
255,656
264,611
277,884
292,582
134,397
141,204
147,381
53.0%
55.5%
57.6%
1.04
1.09
1.14
1.97
1.97
1.99
2.3%
1996-1998 Flemish1997-1999 Brabant
1998-2000
275,384
276,180
276,672
306,823
321,365
327,939
163,819
170,105
171,586
59.5%
61.6%
62.0%
1.11
1.16
1.19
1.87
1.89
1.91
1.3%
1996-1998 Walloon1997-1999 Brabant
1998-2000
91,758
92,576
93,409
125,063
128,896
132,189
61,799
63,431
64,565
67.3%
68.5%
69.1%
1.36
1.39
1.42
2.02
2.03
2.05
0.9%
1996-1998 West1997-1999 Flanders
1998-2000
293,049
293,225
293,326
266,550
277,744
284,736
152,356
157,643
158,640
52.0%
53.8%
54.1%
0.91
0.95
0.97
1.75
1.76
1.79
1.0%
1996-1998 East1997-1999 Flanders
1998-2000
1996-1998 Hainaut
1997-1999
1998-2000
363,355
364,128
364,580
335,767
336,218
336,696
355,875
380,592
406,133
342,317
355,772
368,590
195,814
205,566
209,580
186,627
193,035
197,710
53.9%
56.5%
57.5%
55.6%
57.4%
58.7%
0.98
1.05
1.11
1.02
1.06
1.09
1.82
1.85
1.94
1.83
1.84
1.86
1.8%
1996-1998 Liège
1997-1999
1998-2000
266,827
267,378
267,915
303,995
318,538
323,542
163,598
167,881
169,914
61.3%
62.8%
63.4%
1.14
1.19
1.21
1.86
1.90
1.90
1.1%
1996-1998 Limburg
1997-1999
1998-2000
212,371
213,586
214,631
203,550
208,019
210,153
117,502
118,949
119,206
55.3%
55.7%
55.5%
0.96
0.97
0.98
1.73
1.75
1.76
0.1%
1996-1998 Luxem1997-1999 bourg
1998-2000
60,077
60,538
61,033
52,583
55,869
57,688
29,551
30,762
31,557
49.2%
50.8%
51.7%
0.88
0.92
0.95
1.78
1.82
1.83
1.3%
1996-1998 Namur
1997-1999
1998-2000
112,932
113,587
114,377
120,218
124,536
128,857
65,372
67,044
68,367
57.9%
59.0%
59.8%
1.06
1.10
1.13
1.84
1.86
1.88
0.9%
6/10/2005
1.6%
31
Fitted values
Antwerpen
West-Flanders
East-Flanders
Hainaut
Liège
Limburg
Luxembourg
Namur
Brussels
.7
.65
.6
.55
3-year coverage
.5
.7
.65
.6
.55
.5
.7
.65
.6
.55
.5
1998
1999
2000
1998
1999
2000
1998
1999
2000
year
Walloon-Brabant
Flemish-Brabant
.7
.65
.6
.55
.5
1998
1999
2000
1998
1999
2000
Figure 12. Three year-coverage for cervical cancer screening among 25-64 year old women evaluated
in the years 1998, 1999 and 2000, by Province: observed values (green circles) and fitted linear
regression line.
The 3-year screening coverage in the target population reached in 2000, and ranked
from highest to lowest is shown in Table 14. The range between the highest and
lowest ranking province was 17.4%. The screening coverage was highest in WalloonBrabant (69%) and lowest in Luxembourg (52%). At the exception of FlemishBrabant and Luxemburg, coverage was higher in the Walloon provinces in
comparison with the Flemish provinces.
6/10/2005
32
Table 14. Three-year coverage of cervical cancer screening in 2000 among 25-64 year old women, by
province, ranked from highest to lowest.
Province
Walloon-Brabant
Liège
Flemish-Brabant
Namur
Hainaut
Brussels
Antwerp
East-Flanders
Limburg
West-Flanders
Luxembourg
3-year coverage in 2000
69.1%
63.4%
62.0%
59.8%
58.7%
57.6%
57.6%
57.5%
55.5%
54.1%
51.7%
Figure 13 shows the relation between coverage and #smears/#women ratio on the one
hand and age group on the other hand. We can remark that the high coverage reached
at the age group 30-34 is maintained over a longer age range in Brussels and the
Walloon provinces than in the Flemish provinces where the coverage tends to
decrease with increasing age after having reached the maximum.
The gap between the green curve and red curve in Figure 13 illustrates the amount of
over-screening in all the provinces.
The excess smear use varied between 1.75 (West-Flanders, 1998) and 2.05 (Brussels,
2000) (see Table 13). This means that each covered women received on average
between 2.75 and 3.05 smears over a period of 3 years period.
6/10/2005
33
Antwerpen, 2000
Brussels, 2000
Flemish-Brabant, 2000
Walloon-Brabant, 2000
10 20 30 40 50 60 70
10 20 30 40 50 60 70
West-Flanders, 2000
East-Flanders, 2000
Hainaut, 2000
Liège, 2000
10 20 30 40 50 60 70
10 20 30 40 50 60 70
Limburg, 2000
Luxembourg, 2000
1.4
1.2
1
.8
.6
.4
.2
0
1.4
1.2
1
.8
.6
.4
.2
0
1.4
1.2
1
.8
.6
.4
.2
0
1.4
1.2
1
.8
.6
.4
.2
0
1.4
1.2
1
.8
.6
.4
.2
0
10 20 30 40 50 60 70
Namur, 2000
1.4
1.2
1
.8
.6
.4
.2
0
10 20 30 40 50 60 70
Age
Figure 13. Three-year interval Pap smear coverage, smears/women ratio, by 5-year age group and by
province (Belgium, 2000).
6/10/2005
34
7.1.4.
District level
Below, we present data on cervical screening at the lowest geographical level (the
arrondissement or district). The total number of Pap smears and their relative
frequency over the 5-year period is shown in Table 15.
Table 15. Pap smears interpreted in the period 1996-2000, number and relative frequency, by district.
Use of cervical cytology at a 1-year interval
Absolute
Relative
number of frequency
District
smears
(%)
Living abroad
11580
0.2
Unknown
3921
0.07
Antwerpen
504660
8.8
Mechelen
160232
2.79
Turnhout
204377
3.56
Brussels
586105
10.22
Vilvoorde
367528
6.41
Leuven
246320
4.3
Nivelles
255923
4.46
Brugge
132481
2.31
Disksmuide
18269
0.32
Ieper
49812
0.87
Kortrijk
132314
2.31
Oostende
72399
1.26
Roeselare
58587
1.02
Tielt
40167
0.7
Veurne
27901
0.49
Aalst
137373
2.4
Dendermonde
96755
1.69
Eeklo
35403
0.62
Gent
283896
4.95
Oudenaarde
56438
0.98
St-Niklaas
126931
2.21
District
Ath
Charleroi
Mons
Mouscron
Soignies
Thuin
Tournai
Huy
Liège
Verviers
Waremmes
Hasselt
Maaseik
Tongeren
Arlon
Bastogne
Marche-en-Famenne
Neufchâteau
Virton
Dinant
Namur
Phillipeville
Total
Absolute
Relative
number of frequency
smears
(%)
42915
0.75
234237
4.08
145926
2.54
25667
0.45
109031
1.9
83721
1.46
78367
1.37
69862
1.22
379438
6.62
150865
2.63
44767
0.78
200457
3.5
102688
1.79
94103
1.64
18693
0.33
19421
0.34
28142
0.49
22918
0.4
21732
0.38
51704
0.9
164277
2.86
35898
0.63
5734201
100
The coverage of Pap smear use, considered over a one-year interval is shown for the
five specified years in Table 16. The districts with highest coverages were Nivelles,
Huy and Waremmes (>35%). The districts of Arlon and Moucron showed always the
lowest coverages (<25%). The table also displays the slope of the district-specific
regression trend line. The average yearly increase was highest (>2%) in Aalst,
Neufchâteau and Oudenaarde. Only one district, Maaseik showed a mildly negative
trend (-0.1%).
6/10/2005
35
Table 16. The 1-year coverage by year and by district and yearly increase over the period 1996-2000
(restricted to women between 25-64 years old).
District
Antwerpen
Mechelen
Turnhout
Brussels
Vilvoorde
Leuven
Nivelles
Brugge
Disksmuide
Ieper
Kortrijk
Oostende
Roeselare
Tielt
Veurne
Aalst
Dendermonde
Eeklo
Gent
Oudenaarde
St-Niklaas
Ath
Charleroi
Mons
Mouscron
Soignies
Thuin
Tournai
Huy
Liège
Verviers
Waremmes
Hasselt
Maaseik
Tongeren
Arlon
Bastogne
Marche-en-Famenne
Neufchâteau
Virton
Dinant
Namur
Phillipeville
1996
29.5%
28.9%
29.0%
29.3%
34.0%
28.7%
39.4%
28.8%
22.3%
27.2%
26.6%
27.1%
23.1%
29.2%
28.1%
25.4%
28.2%
23.7%
28.4%
24.6%
30.1%
28.6%
29.1%
30.4%
20.8%
33.2%
30.8%
28.6%
39.6%
33.1%
33.4%
35.8%
30.6%
29.9%
26.5%
19.9%
26.8%
32.7%
22.1%
26.1%
29.3%
31.7%
33.3%
1997
30.6%
30.5%
31.1%
30.7%
36.2%
31.3%
40.8%
29.0%
23.2%
28.9%
30.9%
28.2%
24.7%
26.6%
28.8%
27.2%
28.6%
25.8%
30.2%
26.5%
33.3%
29.8%
30.0%
30.8%
20.2%
34.9%
32.2%
30.6%
40.9%
34.7%
33.9%
37.2%
30.2%
28.0%
27.5%
21.2%
29.8%
33.8%
23.5%
27.8%
30.5%
32.6%
34.1%
1998
32.2%
33.2%
32.4%
33.1%
39.3%
32.7%
42.7%
30.1%
27.5%
33.2%
28.8%
33.6%
26.2%
28.8%
33.6%
32.3%
31.9%
29.7%
34.6%
32.0%
34.5%
33.8%
31.7%
33.1%
23.9%
36.8%
33.6%
33.6%
42.6%
36.0%
34.9%
39.8%
31.8%
28.4%
29.7%
22.0%
31.2%
35.4%
27.2%
28.4%
32.0%
34.1%
36.1%
1999
33.1%
33.4%
30.7%
34.2%
39.8%
33.2%
43.0%
31.7%
26.5%
32.0%
29.4%
30.8%
29.7%
32.5%
31.5%
33.0%
32.9%
28.8%
35.3%
32.4%
34.9%
34.2%
32.7%
33.6%
24.6%
37.1%
34.0%
34.6%
42.3%
37.2%
35.8%
40.2%
31.9%
30.1%
30.4%
23.2%
32.1%
35.6%
29.1%
30.1%
32.2%
34.8%
35.8%
6/10/2005
Yearly
2000
increase
34.5%
1.24%
32.9%
1.09%
30.4%
0.24%
35.6%
1.62%
39.5%
1.46%
32.9%
1.02%
43.5%
1.03%
30.4%
0.58%
26.4%
1.16%
31.3%
1.13%
31.8%
0.89%
31.4%
1.11%
27.4%
1.37%
30.2%
1.37%
31.9%
1.01%
32.7%
2.04%
32.8%
1.36%
27.8%
1.12%
35.1%
1.84%
31.7%
2.01%
34.7%
1.07%
33.9%
1.49%
33.1%
1.07%
34.7%
1.13%
24.7%
1.24%
37.8%
1.13%
34.9%
0.99%
35.5%
1.77%
41.8%
0.59%
36.7%
0.96%
35.3%
0.56%
40.0%
1.13%
31.2%
0.31%
28.2%
-0.13%
29.7%
0.93%
22.1%
0.64%
32.4%
1.34%
35.9%
0.81%
29.5%
2.04%
29.2%
0.84%
33.0%
0.91%
35.3%
0.94%
35.9%
0.69%
36
The ranking of districts by increasing 1-year coverage hardly changed over time as
indicated the high Spearman correlation coefficient Table 17.
Table 17. Spearman’s rank correlation coefficient comparing the rank in 1-year coverage of the
respective districts over successive year.
Spearman
Years
coefficient
1996-1997
0.89
1997-1998
0.85
1998-1999
0.91
1999-2000
0.96
The observed trend of the coverage and the #smears/#women considered over oneyear intervals is shown for each district in Figure 41. The relation between both
indicators and five-year age groups is illustrated in Figure 42 for the year 2000.
The 3-year coverage for the periods 1996-98, 1997-99 and 1998-2000 in 25-64 year
old women; the difference between the last and the first period; the #smears/#women
ratio and the overuse of Pap smears in the last period are documented in Table 18.
Arlon, Mouscron and Diskmuide always were the less screened districts (<50%) and
Vilvoorde, Waremme, Huy and Nivelles always were the best ones (>65%). The
contrast between best and worst screened district was stable over time. In 1998, the
gap between the best and worst screened district was 32.5%, in 1999 it was 32.3% and
in 2000 31.9%. As already observed for the 1-year coverage the ranking from worst
to best screened district is rather stable. The Spearman rank correlation coefficient for
the first and second period was 0.97 and for the second and the third period it was
0.98. In 2 of the 43 districts the difference between the last and the first period was
negative. The coverage decreased with 1% in Maaseik and 0.1% in Turnhout. In all
other districts the trend was positive. In 3 districts the increase was less than 1%:
Hasselt, Kortijk and Huy. In 8 districts the increase exceeded 4%: Mouscron, Aalst,
Gent, Tournai, Brussels, Oudenaarde, Neufchâteau and Roeselare. Roeselare showed
the highest increase (5.%).
6/10/2005
37
Table 18. Three-year coverage for cytological cervical cancer screening in 3 periods indicated by the
last year of the period: difference in coverage between the last and the first period; the number of
smears/# of women and overuse in the last period among 25-64 years old women, by District.
District
Arlon
Mouscron
Disksmuide
Virton
Eeklo
Roeselare
Neufchâteau
Kortrijk
Maaseik
Tongeren
Tielt
Brugge
Ieper
Dendermonde
Bastogne
Aalst
Oostende
Veurne
Turnhout
Oudenaarde
Hasselt
Antwerpen
Brussels
Dinant
Mons
Charleroi
St-Niklaas
Ath
Leuven
Mechelen
Verviers
Tournai
Namur
Gent
Thuin
Marche-en-Famenne
Phillipeville
Liège
Soignies
Vilvoorde
Waremmes
Nivelles
Huy
# smears /
3- year coverage
Difference # women Overuse
1998
1999
2000 2000-1998
2000
2000
37,8% 38,8% 39,3%
1,4%
0,71
1,81
43,2% 45,6% 47,4%
4,1%
0,77
1,63
46,8% 48,9% 49,7%
3,0%
0,85
1,71
49,2% 50,3% 51,1%
1,9%
0,93
1,83
49,3% 52,2% 52,2%
3,0%
0,94
1,80
46,9% 51,4% 52,4%
5,6%
0,88
1,69
47,8% 50,5% 52,6%
4,8%
0,93
1,77
52,3% 53,4% 53,0%
0,7%
0,96
1,82
54,4% 54,4% 53,8%
-0,6%
0,92
1,71
52,4% 53,6% 53,8%
1,4%
0,98
1,82
51,7% 53,7% 54,6%
3,0%
0,98
1,79
53,5% 54,5% 54,7%
1,2%
0,98
1,79
52,3% 54,4% 54,7%
2,4%
1,03
1,88
52,7% 53,9% 55,0%
2,3%
1,06
1,93
52,2% 54,8% 55,0%
2,8%
1,03
1,87
51,2% 54,0% 55,5%
4,3%
1,08
1,95
53,8% 55,6% 56,2%
2,4%
1,02
1,82
55,1% 55,9% 56,5%
1,4%
1,02
1,80
56,6% 57,6% 56,6%
-0,1%
0,99
1,75
52,1% 55,6% 56,9%
4,8%
1,07
1,88
57,3% 57,5% 57,4%
0,0%
1,01
1,77
54,2% 55,8% 57,6%
3,4%
1,08
1,87
53,0% 55,5% 57,6%
4,6%
1,14
1,99
56,1% 57,0% 57,7%
1,6%
1,06
1,83
54,5% 56,2% 57,8%
3,3%
1,12
1,94
55,4% 57,0% 58,2%
2,9%
1,06
1,82
57,0% 58,7% 58,4%
1,4%
1,13
1,94
55,3% 57,7% 58,5%
3,3%
1,10
1,88
56,5% 58,5% 58,8%
2,3%
1,06
1,80
55,3% 57,7% 59,1%
3,7%
1,07
1,81
57,9% 58,7% 59,3%
1,4%
1,13
1,90
55,0% 57,7% 59,5%
4,5%
1,12
1,89
57,7% 58,9% 59,8%
2,1%
1,14
1,91
55,5% 58,6% 60,0%
4,5%
1,18
1,97
58,1% 59,3% 60,3%
2,2%
1,12
1,85
60,9% 61,7% 62,3%
1,3%
1,16
1,87
61,5% 62,6% 62,9%
1,3%
1,16
1,84
60,9% 62,7% 63,4%
2,6%
1,20
1,90
61,1% 62,8% 63,8%
2,6%
1,22
1,92
61,9% 64,1% 64,6%
2,7%
1,29
1,99
65,2% 67,1% 67,7%
2,5%
1,30
1,92
67,4% 68,5% 69,1%
1,8%
1,42
2,05
70,4% 71,1% 71,1%
0,7%
1,37
1,93
6/10/2005
38
Over-use
Coverage change (%)
Fitted over-use
Fitted coverage change (%)
6
5
4
3
2
1
0
-1
.8
.6
1
#smears/#women ratio
1.2
1.4
Figure 14. Relation between cervical cancer screening coverage (green), overuse of Pap smears (red)
and the ratio of #smears/#women ratio, considered over a 3-year interval. Each point represents a
district in 2000.
Over-use
Coverage change (%)
Fitted over-use
Fitted coverage change (%)
6
5
4
3
2
1
0
-1
.6
.8
1
1.2
1.4
Figure 14 illustrates the increase in 3-year screening coverage (in green) and overuse
of Pap smears (in red) for each district in 2000 as a function of the #smears/#women
ratio. It must be remarked that change in coverage is plotted on a percentage scale
and overuse on a unity scale. The regression line through the green triangles is nearly
flat (slope= -0.40% (95% CI: -3.58 - 2.76%). The slope of the regression line through
6/10/2005
39
the red circles (overuse) is 0.50 (95% CI: 0.38-0.61). This indicates that the coverage
did not increase with the intensity of Pap smear use. Or on average, districts with a
higher intensity of Pap smear use, did not obtain a gain in coverage, only the excess
use was influenced.
Figure 15 shows the geographical distribution of the 3-coverage reached in 2000 on a
map. It is clear that South-Luxembourg, West-Flanders/Mouscron and East-Limburg
are under-screened; whereas a cluster of better screened districts is found in the north
of the Walloon-Region (Nivelles, Huy, Waremmes).
To control the effect of differences in age composition at district level the
Standardised Pap smear Ratio (SPR) was computed (see Table 19) using the national
coverage as reference. The ranking of districts by increasing SPR was almost
completely equal as that by crude screening coverage (Spearman coefficient 0.99).
We can conclude that age does not modify the geographical variation.
6/10/2005
40
Cervical cancer screening (Belgium, 2000)
3-year interval coverage
0.7 to 0.75 (1)
0.65 to 0.7
(2)
0.6 to 0.65 (7)
0.55 to 0.6 (20)
0.5 to 0.55 (10)
0.45 to 0.5
(2)
0.4 to 0.45 (0)
0.35 to 0.4
(1)
Figure 15. Geographical variation of the screening coverage considered over a 3-year interval (Belgium, by District, 1998-2000, 25-64 years old women).
6/10/2005
41
Table 19. Standardised Pap smear Ratio and 3-year cervical cancer screening coverage, by District
(Belgium, 1998-2000, women 25-64 years). The average age-specific coverage in Belgium over the
same period was used for standardisation.
District
Arlon
Mouscron
Disksmuide
Virton
Neufchâteau
Eeklo
Roeselare
Kortrijk
Maaseik
Tongeren
Bastogne
Tielt
Ieper
Dendermonde
Brugge
Aalst
Turnhout
Oostende
Oudenaarde
Brussels
Hasselt
Mons
Veurne
Antwerpen
Dinant
Charleroi
Ath
St-Niklaas
Leuven
Mechelen
Verviers
Tournai
Namur
Gent
Thuin
Marche-en-Famenne
Phillipeville
Soignies
Liège
Vilvoorde
Waremmes
Nivelles
Huy
Standardised Pap
smear Ratio
0.665
0.812
0.852
0.868
0.896
0.899
0.902
0.910
0.914
0.918
0.934
0.935
0.939
0.941
0.943
0.949
0.962
0.974
0.975
0.976
0.978
0.983
0.986
0.988
0.988
0.992
0.999
0.999
1.001
1.010
1.012
1.014
1.017
1.026
1.029
1.059
1.075
1.083
1.087
1.107
1.152
1.178
1.208
6/10/2005
3-year
coverage
39.3%
47.4%
49.7%
51.1%
52.6%
52.2%
52.4%
53.0%
53.8%
53.8%
55.0%
54.6%
54.7%
55.0%
54.7%
55.5%
56.6%
56.2%
56.9%
57.6%
57.4%
57.8%
56.5%
57.6%
57.7%
58.2%
58.5%
58.4%
58.8%
59.0%
59.3%
59.5%
59.8%
60.0%
60.3%
62.3%
62.9%
63.8%
63.4%
64.6%
67.7%
69.1%
71.1%
42
7.2. Profession of Pap smear takers
There is a different code for smears taken for specialists and general practitioners.
We assumed that gynaecologists are the only specialists taking Pap smears.
Over the five years, 5.25 million smears were taken, of which 0.82 million (15.5%)
were taken by GPs and 4.44 million (84.5%) by gynaecologists. We remark a
difference of 0.48 million between the number of Pap smears interpreted and taken. It
appears to happen that some Pap smears taken by gynaecologists are not billed (Dr. G.
Albertyn, personal communication). If this phenomenon is true the percentage of
smears taken by gynaecologists could be higher. Given the substantial differences
between the Regions we do not further detail the statistics on Pap smear collection at
the country level. Below, we present the proportion of smears taken by GPs ignoring
the gap of the 0.48 million Paps.
7.2.1.
Regional level
Important regional differences in the percentage of smears taken by GPs are observed.
Taking Pap smears is an almost exclusive activity of gynaecologists in the Walloon
Region (>95%), whereas in the Flemish Region more than one fifth of smears were
taken by GPs. In Brussels 8-10% of Paps were prepared by GPs. The trend is
decreasing, especially in the Flemish Region.
Flemish Region
Brussels
Fraction of smears taken by GPs
.3
.2
.1
0
1996
Walloon Region
1998
2000
.3
.2
.1
0
1996
1998
2000
Year
Figure 16. Trend in the proportion of Pap smears that is taken by GPs, by Region (Belgium, 19962000).
6/10/2005
43
Table 20. Proportion of Pap smears taken by GPs, by year and by Region.
Flemish
Walloon
Brussels
Region
Region
25.7%
10.2%
4.5%
24.7%
9.6%
4.2%
22.8%
9.0%
3.7%
21.5%
8.5%
3.6%
20.0%
8.3%
3.3%
Year
1996
1997
1998
1999
2000
The proportion of smears taken by general practitioners varies by age (see Figure 17).
The proportion of GP smears shows a dip in the age groups where the coverage and
intensity of screening is highest.
Flemish Region
Brussels
Fraction of smears taken by GPs
.3
.2
.1
0
10
Walloon Region
20
30
40
50
60
70
.3
.2
.1
0
10
20
30
40
50
60
70
Age
Figure 17. Relation between the proportion of Pap smears taken by GPs and age group, by Region.
7.2.2.
Provincial level
The variation of the contribution of GPs in Pap smear taking over time and province
is shown in Figure 18. The percentage was highest in Limburg, but here the decline
was also larger (from 34% to 25%) than in the other Flemish provinces. In Liège only
2 to 3% of smears are prepared by GPs.
6/10/2005
44
Antwerpen
West-Flanders
East-Flanders
Hainaut
Liège
Limburg
Luxembourg
Namur
Brussels
Flemish-Brabant
Walloon-Brabant
.4
.3
.2
.1
% smears taken by GP
0
.4
.3
.2
.1
0
1996
1998
2000
.4
.3
.2
.1
0
1996
1998
2000
1996
1998
2000
1996
1998
2000
Figure 18. Yearly trend in the proportion of cervical smears that are taken by GPs, by Province.
Year
7.2.3.
District level
Figure 19 and Table 21 show the variation in the proportion of smears taken by
general practitioners in 2000 over all the districts. The lowest proportion is noted in
Mons (only 1%). The highest proportion was found in Maaseik (43%).
The variation over age for 3 separate years in plotted in Figure 44 (see annexe 13.2).
6/10/2005
45
Table 21. The proportion of Pap smears taken by general practitioners, by District in 2000.
District
Mons
Huy
Waremmes
Verviers
Liège
Thuin
Ath
Charleroi
Soignies
Tournai
Namur
Phillipeville
Dinant
Virton
Arlon
Nivelles
Neufchâteau
Marche-en-Famenne
Bastogne
Mouscron
Brussels
Ieper
Oudenaarde
St-Niklaas
Gent
Oostende
Vilvoorde
Veurne
Brugge
Disksmuide
Tielt
Aalst
Kortrijk
Hasselt
Tongeren
Dendermonde
Roeselare
Antwerpen
Turnhout
Mechelen
Leuven
Eeklo
Maaseik
% of Pap smears
taken by GPs
1.0%
1.3%
2.1%
2.2%
2.3%
2.3%
2.4%
2.5%
2.7%
3.2%
3.2%
4.3%
4.9%
5.4%
5.8%
6.1%
7.1%
7.2%
8.2%
8.3%
8.3%
10.6%
11.6%
12.7%
12.8%
14.2%
14.6%
15.1%
16.0%
16.2%
16.5%
16.7%
17.0%
17.9%
19.3%
19.8%
20.3%
24.0%
24.3%
24.9%
29.1%
30.6%
42.7%
6/10/2005
46
Proportion of Pap smears taken by GPs
(Belgum, 2000, women aged 25-64 years)
0
to 0.06 (15)
0.06 to 0.12 (8)
0.12 to 0.18 (11)
0.18 to 0.24 (4)
0.24 to 0.3
(3)
0.3 to 0.36 (1)
0.36 to 0.43 (1)
Figure 19. Geographical variation in the proportion of Pap smears taken by general practitioners in
2000.
6/10/2005
47
7.3. Colposcopies & cervical biopsies
About 400 000 colposcopic examinations are performed each year in Belgium. The
time, age and geographical distribution of the number colposcopies, cervical biopsies
and interpretations of Pap smears and the ratios of this numbers are documented in
Table 22 - Table 25 and also in Table 38.
The average number of Pap smears interpreted per colposcopy is 2.9. This ratio
hardly varies with year and age, but changes considerably by geographical area.
In the Walloon Region this ratio is only 1.9, whereas 2.8 and 4.9 in Brussels and the
Flemish Region respectively. In Table 38, we have sorted districts by increasing Pap
smear/colposcopy ratio. Remarkable is that all Walloon districts are in top, separated
by Brussels from the Flemish districts. In Mouscron, the ratio is exactly one: for
every Pap smear also a colposcopy is performed.
Table 22. Number of colposcopies, cervical biopsies taken and Pap smears interpreted, ratio of the
number of biopsies over the number of colposcopies, the ratio of the number of Pap smears over the
number of colposcopies and the ratio of the number biopsies over the number of Paps, Belgium, by
year.
Year
1996
1997
1998
1999
2000
Total
Colposcopies
378,750
398,704
408,634
410,083
413,785
2,009,956
Biopsies Pap smears
22,063 1,042,821
20,143 1,095,337
22,048 1,171,347
21,108 1,201,152
20,776 1,223,544
106,138 5,734,201
Biopsy/colposcopy Pap/colposcopy Biopsy/Pap
ratio
ratio
ratio
5.8%
2.8
2.1%
5.1%
2.7
1.8%
5.4%
2.9
1.9%
5.1%
2.9
1.8%
5.0%
3.0
1.7%
5.3%
2.9
1.9%
Table 23. See Table 22, Belgium 1996-2000, by age group.
Age
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
>= 75
Total
Colposcopies Biopsies
Pap smears
2,667
130
4,260
48,972
2,414
123,802
157,159
9,362
442,454
245,852
14,908
692,356
272,003
16,502
790,245
265,565
15,835
780,372
242,440
13,728
694,605
225,029
11,008
625,377
193,096
7,938
537,123
122,395
5,013
357,656
98,854
3,799
294,414
69,624
2,576
203,103
40,161
1,614
114,270
26,139
1,311
74,164
2,009,956 106,138
5,734,201
ratio
ratio
ratio
4.9%
1.6
3.1%
4.9%
2.5
1.9%
6.0%
2.8
2.1%
6.1%
2.8
2.2%
6.1%
2.9
2.1%
6.0%
2.9
2.0%
5.7%
2.9
2.0%
4.9%
2.8
1.8%
4.1%
2.8
1.5%
4.1%
2.9
1.4%
3.8%
3.0
1.3%
3.7%
2.9
1.3%
4.0%
2.8
1.4%
5.0%
2.8
1.8%
5.3%
2.9
1.9%
6/10/2005
48
On average, 5.3% of colposcopies are accompanied by taking a cervical biopsy.
Again this percentage varies slightly by age and time. This percentage is respectively
3.3% and 8.7% in the Walloon and Flemish Region. It was highest in Antwerp
(13.3%) and lowest in Namur (2.3%).
On average the ratio of the number of smears over the number of cervical biopsies is
1.8%. The variation of biopsy/pap smear ratio by age is displayed in Figure 20. This
ratio varies only slightly by Region, indicating minor inter-regional differences in
prevalence of colposcopically visible cervical lesions. The range of variation at
district level is less important as well, if we exclude 2 districts with outlying
biopsy/pap smear ratio (Oudenaarde and Waremmes with a ratio of 6.4% or 7.1%).
The linear relation between the Pap/colopocopy ratio and biopsy/Pap ratio at district
level is weakly negative and non significant (slope: (-0.10%; CI: -0.21% to 0.02%).
We observe a discrepancy between the number of biopsies taken and interpreted.
106,138 cervical biopsies were taken, whereas only 72,757 (68.5%) were interpreted.
Probably other codes were used to impute histological interpretation of cervical
biopsies.
Table 24. See Table 22, Belgium, 1996-2000, by Region.
Region
Colposcopies
Biopsies Pap smears
Flemish Region
618,033
53,825 3,003,722
Brussels
207,587
12,143
586,105
Walloon Region
1,176,139
39,315 2,128,873
Total
2,001,759
105,283 5,718,700
ratio
ratio
ratio
8.7%
4.9
1.8%
5.8%
2.8
2.1%
3.3%
1.8
1.8%
5.3%
2.9
1.8%
Table 25. See Table 22, Belgium 1996-2000, by province.
Province
Colposcopies
Biopsies Pap smears
Antwerp
158,713
21,116
869,269
East-Flanders
67,346
6,722
531,930
West-Flanders
114,510
9,959
736,796
Hainaut
487,515
13,491
719,864
Liège
418,985
14,063
644,932
Luxembourg
82,686
5,664
397,248
Namur
62,377
1,450
110,906
Limburg
140,942
5,444
251,879
Brussels
207,587
12,143
586,105
Flemish-Brabant
136,522
10,584
613,848
Walloon-Brabant
124,576
4,647
255,923
6/10/2005
ratio
ratio
ratio
13.3%
5.5
2.4%
10.0%
7.9
1.3%
8.7%
6.4
1.4%
2.8%
1.5
1.9%
3.4%
1.5
2.2%
6.9%
4.8
1.4%
2.3%
1.8
1.3%
3.9%
1.8
2.2%
5.8%
2.8
2.1%
7.8%
4.5
1.7%
3.7%
2.1
1.8%
49
Biopsy/Pap smear ratio
3%
2%
1%
0%
15
20
25
30
35
40
45
50
55
60
65
70
75
Age group
Figure 20. Ratio of the number of cervical biopsies over the number of Pap smears taken, by age
(Belgium, 1996-2000).
biopapr
Fitted values
.08
Ratio Biopsies / Pap smears
War
Oud
.06
.04
Ath
Ant
Huy
Iep
Cha
Nam
Din
Bru
Neu
Liè Liv
NivPhi
Thu
Mar
Tou
Soi
Mon
Mou VirVer
.02
Arl
Bas
Has
Leu
Oos
Tur
Vil
Ton
Aal
Roe Den
Eek
Mec
Dis Tie
St-
Kor
Gen
Veu
Bru
Maa
0
0
2
4
6
8
10
Ratio Pap smears / colposcopies
12
Figure 21. Relation at district level between the Pap smears / colposcopy ratio and the biopsy / Pap
smear ratio.
6/10/2005
50
7.4. Hysterectomies
7.4.1.
National level
In total more than 84,500 total hysterectomies were performed in Belgium between
1996 and 2000. The number of hysterectomies by year varied from 15,454 in 2000
and 18,138 in 1998. The histogram in Figure 22 shows the distribution of the number
of hysterectomies by 5-year age group.
Table 26. The total number of total hysterectomies performed, each year, in Belgium between 1996
and 2000.
Year
1996
1997
1998
1999
2000
Total
Frequency
16622
16982
18138
17368
15454
84564
Percent
19.66
20.08
21.45
20.54
18.27
100
year==1996
year==1997
year==1998
year==1999
20 25 30 35 40 45 50 55 60 65 70 75
20 25 30 35 40 45 50 55 60 65 70 75
4000
3000
2000
Frequency
1000
0
4000
3000
2000
1000
0
Age group
Figure 22. Number of total hysterectomies by five year age group performed in Belgium in the years
1996, 1997, 1998 and 1999.
6/10/2005
51
Hysterectomies/1000 women-years
The average yearly incidence rate of total hysterectomy per thousand women and by
age, computed over the period 1996 to 2000, is plotted in Figure 23. The resulting
cumulative incidence (which is nearly equivalent to the prevalence of being totally
hysterectomiseda) by age is shown in Figure 24.
The overall incidence within the 25-64 years group was 5.0 per 1000 women-years.
The age-specific incidence rate increases up to a peak at 9.6/1000 women-years in the
50-54 year age group. The hysterectomy rate declines thereafter to reach a rather
constant level between 4 and 5 per 1000/year between in the age groups 55-70. The
increase in the last age group (women of 75 years or older) might be due to the
assumption that all hysterectomies in the age group 75+ were considered to occur all
within the age group 75-79.
Figure 25 shows the observed incidence rate of hysterectomy among women of 25-64
years old and also the linear trend obtained from a Poisson regression using the
calendar year as a continuous co-variate. The linear trend was mildly decreasing
(slope: 0.969; 95% CI: 0.964 to 0.974). For a multivariate Poisson-regression,
including district, age group and calendar-year as covariate, we refer to sub-chapter
7.4.4.
Figure 26 displays the time trend of the observed hysterectomy incidence rate for 3
separate age groups. The rate is mildly decreasing in the youngest age group (<45
years old women), and mildly increasing among women of 65 years or older. Among
45-64 years old, the rate increases mildly until 1998, but decreases thereafter.
10
7.5
5
2.5
0
40
25
55
70
Age group
Figure 23. Average incidence of total hystectomy by age (Belgium, 1996-2000).
a
Under the assumption of absent period effect.
6/10/2005
52
Cumulative incidence of hysterectomy up to a given age group
Belgium, 1996-2000
Cumulative incidence (%)
30
25
20
15
10
5
0
20
25
30
35
40
45 50 55
Age group
60
65
70
75
Figure 24. Average cumulative incidence of hysterectomy, up to a given age group (Belgium, 19962000).
Incidence of hysterectomy, observed rate & linear trend
Belgium, women 25-64 years
6
5
4
3
2
1
0
1996
1997
1998
Year
1999
2000
Figure 25. Observed incidence rates of hysterectomy (red circles) and linear trend between 1996 and
2000 among women between 25 and 64 years old.
6/10/2005
53
Incidence of hysterectomy
Trend by age group, Belgium
Age: 15-44 years
Age: >=65 years
Age: 45-64 years
8
6
4
2
0
1996
1997
1998
Year
1999
2000
Figure 26. Observed trend of hysterectomy rates by age group in Belgium between 1996 and 2000.
6/10/2005
54
7.4.2.
Regional level
The average age-specific incidence of hysterectomy is shown for the three Regions
separately in Figure 27. The shape of curve is similar for the three regions as
described for the whole of Belgium in the previous sub-chapter. However the age
specific incidences are substantially higher in the Flemish and substantially lower in
Brussels. Peak incidences occur systematically in the 45-49 year age group:
respectively 10.6, 6.4 and 8.7 per 100 women-years in the Flemish, Capital and
Walloon Region.
Flemish Region
Brussels
10
5
0
25
Walloon Region
40
55
70
10
5
0
25
40
55
70
Age group
Figure 27. Average incidence of total hystectomy by age (Belgium, by Region, 1996-2000).
The cumulative incidence of hysterectomy up to a given age is plotted in Figure 28.
Figure 29 shows the observed incidence rate of hysterectomy among women of 25-64
years old and also the linear trend obtained from a Poisson regression. This Poisson
regression included an interaction between Region and year, expressed as a
continuous variable. The contrast between the Flemish Region and the two other
Regions increases over time since the downward trend is larger (further from unity) in
Brussels and the Walloon Region. The coefficients for the linear slopes are: 0.987
(95% CI: 0.897 to 0.940) for the Flemish Region; 0.918 (95% CI: 0.897 to 0.940) for
Brussels and 0.954 (95% CI: 0.943 to 0.965) for the Walloon Region. For a
multivariate Poisson-regression, including district and separate age-group and
calendar-years as covariate, we refer to sub-chapter 7.4.4.
The time trend over the 5 calendar years for 3 separate age-groups is shown in Figure
30. Remarkable is the abrupt drop in the Brussels Region in the 45-64 year age group
between 1998 and 1999.
6/10/2005
55
Cumulative incidence of hysterectomy up to a given age group
Cumulative incidence (%)
Flemish Region
Brussels
30
25
20
15
10
5
0
20
Walloon Region
30
40
50
60
70
30
25
20
15
10
5
0
20
30
40
50
60
70
Age group
Figure 28. Average cumulative incidence of hysterectomy, up to a given age group (Belgium by
Region, 1996-2000).
Incidence of hysterectomy, observed rate & linear trend
Belgium by Region, women 25-64 years
Flemish Region
Brussels
6
5
4
3
2
1
0
1996 1997 1998 1999 2000
Walloon Region
6
5
4
3
2
1
0
1996 1997 1998 1999 2000
Year
Figure 29. Observed incidence rates of hysterectomy (red circles) and linear trend, by Region,
between 1996 and 2000 among women between 25 and 64 years old.
6/10/2005
56
Age: 15-44 years
Age: >=65 years
Age: 45-64 years
Flemish Region
Brussels
8
6
4
2
0
1996 1997 1998 1999 2000
Walloon Region
8
6
4
2
0
1996 1997 1998 1999 2000
Year
Figure 30. Observed trend of hysterectomy rates by age group in the three Belgian Regions between
1996 and 2000.
6/10/2005
57
7.4.3.
Provincial level
For the provinces, we just present the average yearly incidence, which is plotted as a
function of age in Figure 31. Remark the higher incidences (peak incidence at 45-49
year > 10/1000 women years) in the provinces of Antwerp, West- & East-Flanders
and Limburg.
Antwerpen
West-Flanders
East-Flanders
Hainaut
Liège
Limburg
Luxembourg
Namur
Brussels
15
10
5
0
15
10
5
0
15
10
5
0
25
40
55
70
Flemish-Brabant
25
40
55
25
70
40
55
70
Age
group
Walloon-Brabant
15
10
5
0
25
40
55
70
25
40
55
70
Figure 31. Average incidence of total hysterectomy by age (Belgium, by Province, 1996-2000).
6/10/2005
58
7.4.4.
District level
The age-specific incidences of total hysterectomy are plotted in Figure 45. The
relative risk of hysterectomy among women of 20 years and older, estimated from a
Poisson regression including District, age group and year as covariates, are shown in
Table 27. The district of Antwerp is used as reference. Districts are ranked from low
to high incidence. The 10 districts with lowest incidence are all from the Walloon
Region or Brussels, whereas the 17 districts with highest incidence belong all to the
Flemish Region. The range between lowest and highest rate is situated between
2.4/1000/y (Ath) and 5.5/1000/y (St.-Niklaas).
A clearer overview of the
geographical contrasts is obtained from the map in Figure 32.
Table 27. Relative risk (RR) of hysterectomy by District using the incidence in Antwerp as reference,
ranked by increasing risk. RRs are estimated from a Poisson regression, including age group and year
as additional covariates (Belgium, 1996-2000, ≥20 year old women).
District
Relative risk
Antwerpen
1
Ath
0.54
Tournai
0.57
Brussels
0.57
Namur
0.66
Soignies
0.69
Dinant
0.70
Mons
0.71
Arlon
0.73
Mouscron
0.75
Walloon-Brabant
0.76
Vilvoorde
0.76
Neufchâteau
0.76
Brugge
0.78
Veurne
0.80
Waremmes
0.81
Virton
0.82
Thuin
0.83
Phillipeville
0.85
Charleroi
0.85
Liège
0.86
Verviers
0.87
Age
20-24
25-29
30-34
35-39
40-44
45-49
Relative risk
Year
1996
1997
1998
Relative risk
1
4.83
14.74
35.03
65.26
76.86
1
1.01
1.06
(95% CI)
(0.91 to 1.00)
(0.98 to 1.06)
(0.55 to 0.59)
(0.73 to 0.79)
(0.84 to 0.90)
(0.72 to 0.79)
(0.75 to 0.82)
(0.91 to 1.10)
(1.16 to 1.31)
(0.95 to 1.04)
(1.00 to 1.11)
(0.86 to 0.97)
(1.13 to 1.30)
(0.72 to 0.88)
(0.93 to 1.02)
(0.95 to 1.05)
(0.93 to 1.08)
(0.99 to 1.06)
(0.92 to 1.04)
(1.20 to 1.31)
(0.49 to 0.59)
(95% CI)
(4.15 to 5.62)
(12.77 to 17.02)
(30.42 to 40.35)
(56.7 to 75.11)
(66.79 to 88.45)
(95% CI)
(0.99 to 1.03)
(1.04 to 1.08)
District
Relative risk
Leuven
0.87
Bastogne
0.87
Huy
0.88
Marche-en-Famenne
0.90
Roeselare
0.92
Mechelen
0.95
Tongeren
0.96
Aalst
0.97
Oudenaarde
0.98
Maaseik
0.98
Kortrijk
1.00
Dendermonde
1.00
Disksmuide
1.00
Eeklo
1.00
Turnhout
1.02
Gent
1.02
Oostende
1.05
Hasselt
1.07
Tielt
1.21
Ieper
1.23
St-Niklaas
1.25
(95% CI)
(0.84 to 0.90)
(0.68 to 0.75)
(0.68 to 0.82)
(0.82 to 1.00)
(0.78 to 0.88)
(0.91 to 1.00)
(0.81 to 0.94)
(0.83 to 0.89)
(0.83 to 0.91)
(0.74 to 0.88)
(1.03 to 1.11)
(0.94 to 1.03)
(0.91 to 1.10)
(0.65 to 0.81)
(0.77 to 0.98)
(0.82 to 1.00)
(1.00 to 1.11)
(0.73 to 0.91)
(1.13 to 1.30)
(0.62 to 0.69)
(1.20 to 1.31)
Age
50-54
55-59
60-64
65-69
70-74
75+
Relative risk
58.10
35.08
32.12
33.79
30.98
42.19
(95% CI)
(50.46 to 66.88)
(30.43 to 40.43)
(27.86 to 37.02)
(29.32 to 38.94)
(26.87 to 35.72)
(36.59 to 48.64)
Year
1999
2000
Relative risk
1.00
0.88
(95% CI)
(0.98 to 1.02)
(0.86 to 0.90)
6/10/2005
59
Total hysterectomy rate (per 1000 women-years)
Belgium, 1996-2000 (women between 25-64 years)
5 to 6
(3)
4.5 to 4.99 (3)
4 to 4.49 (11)
3.5 to 3.99 (13)
3 to 3.49 (9)
2.5 to 2.99 (2)
2 to 2.49 (2)
all others
(3)
IncHyster.WOR
IncHyster.WOR
IncHyster.WOR
IncHyster.WOR
Figure 32. Geographical variation in the rate of total hysterectomy by District (Belgium, 1996-2000; ≥20 year old women, per 1000 women-years).
6/10/2005
60
As for the other geographical levels, we present also the hysterectomy rate for the age
group 25-64 years (Table 28). The SHR (standardised hysterectomy ratio) is
computed using the national age-specific rate over the 1996-2000 period as reference.
The geographical contrasts look hardly influenced by age group.
Table 28. Hysterectomy incidence and standardised hysterectomy ratio (SHR) among women between
25 and 64 years old, ranked by increasing SHR (Belgium, 1996-2000).
District
Tournai
Ath
Brussels
Namur
Neufchâteau
Dinant
Soignies
Nivelles
Arlon
Vilvoorde
Mons
Mouscron
Brugge
Veurne
Bastogne
Thuin
Waremmes
Leuven
Virton
Verviers
Marche-en-Famenne
Roeselare
Charleroi
Liège
Phillipeville
Huy
Tongeren
Oudenaarde
Mechelen
Aalst
Maaseik
Disksmuide
Dendermonde
Kortrijk
Turnhout
Antwerpen
Gent
Eeklo
Hasselt
Oostende
Tielt
Ieper
St-Niklaas
Incidence rate (/1000
women-years)
2.87
2.96
2.99
3.81
3.88
3.93
4.06
4.30
4.01
4.25
4.34
4.40
4.50
4.66
4.52
4.82
4.82
4.89
4.86
4.98
4.98
4.98
5.14
5.21
5.24
5.28
5.45
5.48
5.48
5.50
5.62
5.57
5.71
5.71
5.74
5.89
5.82
6.11
6.21
6.40
6.79
7.31
7.37
6/10/2005
Standardised Hysterectomy
Ratio
0.56
0.58
0.62
0.75
0.78
0.78
0.80
0.82
0.82
0.82
0.84
0.85
0.89
0.92
0.93
0.94
0.94
0.97
0.98
1.00
1.00
1.01
1.01
1.02
1.03
1.03
1.08
1.10
1.10
1.10
1.13
1.13
1.14
1.14
1.15
1.17
1.18
1.22
1.23
1.26
1.38
1.46
1.47
61
7.5. Cytological screening in hysterectomised women
Women who have undergone total hysterectomy are not at risk for cervical cancer if
the indication for the intervention was not oncologic. In the next subchapter we will
study the intensity of Pap smear use in women who had total hysterectomy or cervixamputation. Since we do not know if smears taken in the same calendar year as the
hysterectomy, took place before or after the surgery, we will consider the calendar
years after the hysterectomy. Two types of post-hysterectomy periods are defined:
one-year (see sub-chapter 7.5.1) and three-year periods (see sub-chapter 7.5.2).
Linkage between Pap smear histories and hysterectomy using the anonymous IDnumber revealed that certain women had multiple hysterectomies. This artifact is
probably due to many-to-one ID mismatching. In the following sub-chapter all
second and following hysterectomies in a same woman were deleted. The
consequence is that the analyses concern only 81,751 hysterectomies instead of
84,564.
7.5.1.
Use of Pap smears one year after hysterectomy
The proportion of women with a hysterectomy in a given year in the period 19961999 and who had a least one Pap smear in the subsequent year, specified by age
group, is shown in Figure 33. No time trend effects can be discerned, since the curves
for the four years coincide. The proportion of screened women is highest among
young women: on average 49% among 20-29 years old women. It decreases with
age, to reach the lowest level in the age groups 40-54 (on average 13%). The
proportion increases a few percentages in the age group 55-64 (on average 18%) and
it decreases again until 13% in women being 70 or older.
The ratio of the number of smears taken over the number hysterectomised women is
plotted in Figure 34.
Table 29 shows by five-year groups the number of hysterectomies performed in 1999
and the number of Pap smears examined among the hysterectomised women in 2000.
It shows also the proportion of screened hysterectomised women and the #smears/
#women ratio which corresponds to the blue curves in Figure 33 and Figure 34. Both
measures are plotted in one graph (see Figure 35). The number of smears per
screened hysterectomised woman varies between 1.3 and 1.6 with lowest values
observed in 40-54 year age group.
Among the 15,883 women hysterectomised in 1999; 2,553 were screened
cytologically using on average 1.4 Pap smears per screened woman during the year
2000.
6/10/2005
62
Pap smears in women being hysterectomised 1 year before
Proportion with at least 1 smear
1996
1998
1997
1999
1
Fraction
.8
.6
.4
.2
0
20
30
40
50
Age group
60
70
Figure 33. Proportion of women with hysterectomy occurring in the year indicated in the legend, who
had at least one Pap smear in the subsequent year.
Table 29. Cytological screening in 2000 among women who underwent a total hysterectomy in 1999.
Number
of
Age
20
25
30
35
40
45
50
55
60
65
70
75
25-64
20+
Hysterectomies Smears
in 1999
in 2000
41
27
198
151
644
308
1,599
457
2,803
535
2,964
463
2,087
386
1,138
279
1,117
283
1,166
331
999
217
1,127
179
12,550
2,862
15,883
3,616
Women with
Proportion of
#smears/
# smears/
smear in
women with smear #women ratio screened woman
in 2000
20
48.8%
0.659
1.4
96
48.5%
0.763
1.6
198
30.7%
0.478
1.6
310
19.4%
0.286
1.5
388
13.8%
0.191
1.4
349
11.8%
0.156
1.3
291
13.9%
0.185
1.3
208
18.3%
0.245
1.3
197
17.6%
0.253
1.4
217
18.6%
0.284
1.5
155
15.5%
0.217
1.4
123
10.9%
0.159
1.5
2,037
16.2%
0.228
1.4
2,552
16.1%
0.228
1.4
6/10/2005
63
Pap smears in women being hysterectomised 1 year before
#smear/#women ratio; by year of hysterectomy and age
1996
1998
1997
1999
1
.8
Fraction
.6
.4
.2
0
20
30
40
50
Age group
60
70
Figure 34. Use of Pap smears in hysterectomised women: ratio of # smears/ # women, where # smears
is the number of Pap smears taken in the calendar-year following the year where hysterectomy took
place and # women is the number of women who underwent a total hysterectomy in the year defined in
the legend.
Pap smears taken in 2000 in women being hysterectomised in 1999
proportion with at >=1 smear
.8
.6
.4
.2
0
20
30
40
50
Age group
60
70
Figure 35. Pap smear use in hysterectomised women. Red full line curve: proportion of women with
hysterectomy in 1999 who had at least one Pap smear in 2000; Green interrupted line curve: ratio of the
number of Pap smears taken in 2000/ number of women hysterectomised in 1999.
6/10/2005
64
7.5.2.
Use of Pap smears in the 3-year period following
hysterectomy
The proportion of hysterectomised women who had a Pap smear taken in the three
subsequent years is plotted using a full line curve in Figure 36. The #smears/#women
ratio is plotted on the same graph using interrupted lines. These indicators are
documented for two hysterectomy years: 1996 and 1997. Absolute values and
computed indicators are enumerated in Table 30. The proportion of screened women
is highest in the youngest groups (63% in the 20-29 year age group) and it decreases
until 23% in the 40-45 year group. This proportion increases until 28% in the 55-64
year group and finally decreases again (18% in women of 70 years and older). The
number of smears per screened women was on average 2.3 and varied between 2.0
and 3.2 by age group.
#women/#smears ratio
Proportion w Pap, Hectomy 1996
#women/#smears ratio
Proportion w Pap, Hectomy 1997
2.5
2
1.5
1
.5
0
20
30
40
50
Age group
60
70
Figure 36. Pap smear use in hysterectomised women. Full line curves: proportion of women with at
least one smear in the 3-year period following the year of hysterectomy; interreputed line curves: #
smears taken in the three subsequent years/ # women hysterectomised in a given year; orange line
curves: concerning women hysterectomised in 1996; blue line curves: concerning women
hysterectomised in 1997.
6/10/2005
65
Table 30. Cytological screening in 1998-2000 among women who underwent a total hysterectomy in
1997.
Age
20
25
30
35
40
45
50
55
60
65
70
75
25-64
20+
# smears/screened
Women with Proportion of women
#smears/
woman
Hysterectomies
Smears
smear
with smear
#women ratio
in 1997
in 1998-2000 in 1998-2000
2.5
65
116
46
70.8%
1.785
3.2
213
416
130
61.0%
1.953
2.7
725
744
280
38.6%
1.026
2.2
1,719
1,025
467
27.2%
0.596
2.0
3,093
1,363
697
22.5%
0.441
2.0
3,436
1,538
768
22.4%
0.448
2.1
2,258
1,127
525
23.3%
0.499
2.3
1,122
733
318
28.3%
0.653
2.5
1,070
756
298
27.9%
0.707
2.5
1,148
669
268
23.3%
0.583
2.5
922
480
189
20.5%
0.521
2.3
857
319
139
16.2%
0.372
13,636
16,628
7,702
9,286
3,483
4,125
6/10/2005
25.5%
24.8%
2.2
2.3
0.565
0.558
66
8. Discussion
8.1. Quality of the dataset
The quality of the dataset was compromised by truncating details of information such
as date of the act, age of the women, locality of residence. This impeded computation
of exact screening intervals and excluded the possibility to assess the delay between
(or the simultaneity of) Pap smear collection and other investigations or interventions
The identities of concerned women were irreversibly encrypted. Nevertheless records
can refer to a unique person by their content when locality, age and date of the
medical act are known. Truncation of data was done to reduce the probability of
providing unique data, which theoretically include a risk of identifiability.
Nevertheless, one should not confound uniqueness with identifiability.
It must be remarked that organized population screening, includes by definition
registration of individual data with the possibility of linkage between population-,
screening cancer- and mortality-registers [Commission EU, 2003; Council EU, 2003].
Moreover the Belgian privacy protect law foresees derogation specifically for
population screening [Arbyn, 1999].
The amount of records with improbable results (women with multiple hysterectomies,
or with more then 20 smears) was limited (<1% of the complete dataset).
8.2. Comparison with results of the National Health Interview
Survey (HIS)
In 1997 and 2001, a nation-wide health interview survey (HIS) was conducted
including respectively 10 000 and 12 000 households, randomly chosen from the
national register using a stratified cluster sampling design [Demarest, 2002]. The
questionnaire contained seven questions on cervical cancer screening by Pap smear,
which were addressed to all selected women older than 14 years. The following
questions are of interest for this report: “Did you ever had a gynaecological
examination for cervical cancer (a cervical smear); b) If yes, how long ago did this
happen? (<1 year ago; 1-3 years ago, 3-5 years ago; more than 5 years ago, no
answer). From the answers to these two questions an indicator variable for 3-year
screening coverage was constructed (having had a Pap smear taken less then 3 years
ago). Another question of interest was: “Did you ever undergo a cervical ablation of
the cervix?”
We applied the same statistical weighting procedures as described by the authors of
the HIS report [Demarest, 2002].
Table 31 shows the coverage estimated from the health interview surveys in 1997 and
2001 confronted with the coverage computed from the individual health insurance
data set for respectively one year later and before at the level of Belgium and the three
Regions for women in target age range of 25-64 years. The coverage, estimated from
6/10/2005
67
HIS data, was substantially and statistically significantly higher than in our study. At
the national level this difference was 11.3% (CI: 9.3-13.9%). The difference between
2001 (HIS-data) and the 2000 coverage (IMA) was highest in the Flemish and Capital
Region, respectively 14.5% (CI: 11.3-17.5%) and 14.8% (CI: 11.0-18.3%). In the
Walloon Region, this difference was smaller (4.7%; CI: 0.5-8.8%).
Table 31. Comparison of the of 3-year cervical cancer screening coverage estimated from the health
interview surveys conducted in 1997 and 2001 and the coverage computed from individual health
insurance data in 1998 and 2000, among women between 25 and 64 years old, in Belgium and by
Region.
Health Interview Survey (HIS)
Year
1997
2001
Estimate
Estimate
(95% CI)
(95% CI)
69.6% (67.1% -71.9%)
69.9% (67.9% -72.5%)
Belgium
Flemish Region
Capital Region
Walloon Region
73.4% (69.9% -76.7%)
64.1% (59.8% -68.1%)
64.0% (59.8% -67.9%)
71.9% (68.7% -74.9%)
72.4% (68.6% -75.9%)
65.6% (61.4% -69.7%)
Individual patient
files from IMA
1998
2000
Coverage
55.90%
58.60%
55.00%
53.00%
58.40%
57.4%
57.6%
60.9%
Table 32 shows values of coverage at the provincial level estimated from the 2001
health survey or computed for 2000 from the individual health insurance data. The
HIS-coverage was always higher (excepted for Namur) and the difference was
statistically significant for Antwerp, Flemish-Brabant, West-Flanders, Limburg,
Brussels and Luxembourg.
Table 32. Comparison of the of 3-year cervical cancer screening coverage estimated from the health
interview survey (HIS, 2001) and the coverage computed from individual health insurance data for the
year 2000, among women between 25 and 64 years old, by province.
Province
HIS, 2001
(95% CI)
Antwerp
71.7% (65.9% -76.8%)
Flemish-Brabant
77.5% (68.8% -84.2%)
West-Flanders
67.7% (59.0% -75.3%)
East-Flanders
64.2% (56.8% -71.1%)
Limburg
80.1% (73.2% -85.6%)
Brussels
72.4% (68.6% -75.9%)
Walloon-Brabant 74.2% (62.1% -83.5%)
Hainaut
61.1% (55.6% -66.4%)
Liège
68.5% (61.2% -75.0%)
Luxembourg
61.5% (54.9% -67.7%)
Namur
59.7% (46.9% -71.3%)
Individual patient files
from IMA (2000)
57.60%
62.00%
54.10%
57.50%
55.50%
57.60%
69.10%
58.70%
63.40%
51.70%
59.80%
Table 33 and Figure 37 show the same type of contrast at the Belgian level but now
differentiated by age group.
6/10/2005
68
Table 33. Estimation of the 3-year cervical cancer screening coverage estimated from the national
health interview survey (HIS) in 2001 and the corresponding coverage computed from individual
health insurance data (provided by IMA) for the year 3000, Belgium, by age group.
Age group
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75+
HIS, 2001
Estimate
8.33%
48.51%
67.88%
78.23%
73.95%
71.15%
69.24%
73.43%
66.28%
53.96%
44.17%
25.97%
11.91%
(95% CI)
(4.49% -14.95%)
(39.84% -57.27%)
(61.12% -73.97%)
(72.29% -83.18%)
(67.24% -79.71%)
(64.63% -76.91%)
(62.90% -74.93%)
(66.48% -79.39%)
(58.99% -72.87%)
(46.07% -61.65%)
(35.76% -52.92%)
(19.05% -34.34%)
(7.95% -17.46%)
IMA, 2000
Coverage
15.6%
49.3%
63.7%
66.7%
65.6%
62.5%
59.1%
56.4%
46.5%
38.2%
26.9%
17.8%
16.4%
In the age groups 30-74, the coverage estimated from the HIS data was significantly
higher than that computed from the IMA data, and this difference tended to increase
with age. Nevertheless, in the youngest and oldest groups the differences were small
and the coverage became even higher in the IMA data than in the HI survey.
Substantial discrepancies in coverage derived from two independent data sources are
observed. We assume that the coverage computed from the individual health
insurance data has a higher likelihood to correspond more to the true coverage since it
does not suffer from reporting and selection biases inherent to health interview
surveys, and only influenced by “minor” administrative errors [R. Mertens; V. Fabri:
personal communications]. Several authors, who investigated the accuracy of selfreported utilisation of Pap smears, found systematic overestimation of the actual
screening status [Walter, 1988; Sawyer, 1989; Kottkel, 1995; Montano, 1995; Bowman,
1997; McGovern, 1998]. Selective participation of screened women in surveys is
another explanation [Arbyn, 1997; 2000]. Finally, another reason of overestimation
might be due to the fact that 11% of women in the age-range 25 64 year did not answer
the 2 questions on cervical cancer screening and probably the screening status in this
category was lower than among responding women.
Remarkable is the fact that the discrepancy between the HIS- and IMA coverage varies
with age and geographical area. According to the HIS the Flemish Region showed the
highest coverage being 9.4% in 1997 an 4.3% higher in 2001 than the Walloon Region.
The variation in erroneous answering to the questionnaire and selective participation in
surveys over different populations may require further investigation. A data linkage
based on a unique linkable identifier, might be an interesting research activity to
disentangle this question.
The fact that the years for which the coverage is documented in both types of studies
do not exactly correspond does not look a plausible explanation for the discrepancy,
given the limited trend effect observed in both studies.
6/10/2005
69
The proportion of women, being 25 to 64 year old that declared having undergone
hysterectomy, was estimated in the health interview survey to be 9.9% (CI: 8.8%11.0%).
The prevalence of total hysterectomy, derived from the cumulative incidence in the
individual health insurance data file was 8.9%. Both figures approximate each other
remarkably well. Answers to the hysterectomy question were probably more accurate
than those concerning cervical screening. Memory and projection biases are probably
negligible, since a hysterectomy is a drastic intervention in woman’s life.
8.3. Comparison with statistics from the National Health
Insurance Institute
The National Health Insurance Institute (INAMI/RIZIV)a produces yearly statistics
for each code from the list of medical acts. These reports do not contain details of
age or place of the patient and are only based on bookkeeping data.
8.3.1.
Collection and interpretation of cervical smears
Figure 38 shows the trend in the number of interpreted smears by year between 1983
and 2003. In that period (excluding 1983, which might be incomplete) the annual
number of Pap smear interpretations increased from about 0.6 million to nearly 1.3
million. Also the INAMI costs for interpretation and for the sum interpretation +
collection, expressed in EURO or EURO-equivalents, are plotted. The capital spent
in 2003 by the health authority to reimburse cervical examinations is around 24
million EURO (costs for the medical visit not included).
In Table 34, the numbers of smears, collected by GPs or by specialists and interpreted,
reported by the National Health Insurance Institute (INAMI) are compared with those
computed from the individual patient data set. Again we observe that the sum of the
number of smears taken is higher than the number of smears interpreted. Moreover,
important absolute and proportional differences are observed between the two
databases. The number of interpreted smears was substantially higher in the INAMI
database in 1996 and 1997. In the subsequent years the difference was less than 1%.
Differences between the two databases are not unexpected. Statistical information
from the National Health Insurance Institute is based on bookkeeping data, where the
invoice date is considered. Furthermore, according to the INAMI’s bookkeeping
system, a year spans the months January-October of the current year and NovemberDecember of the passed year. In 1996 more files were processed due to a delay in
bookkeeping in 1995,which explains the positive differences between INAMI and
IMA.
a
National Health Insurance Institute, in French: Institut Natioanl d’Assurance des Maladies et des
Invalidités (INAMI); in Dutch: Rijks-Instituut voor Ziekte en Invaliditeits verzekering (RIZIV). In this
report we will use the French abbreviation “INAMI”.
6/10/2005
70
Table 34. Number of smears collected by GPs, specialists and number of smears interpreted by year in
Belgium, as reported in the yearly statistics of the National Health Insurance Institute (INAMI) and as
computed from the individual patient data provided by IMA. Absolute and relative differences
between both databases.
Proportional difference
INAMI
Collection Pap by
Year
GP
IMA
Inter-
Difference (INAMI-IMA)
Collection Pap by
specialist pretation
GP
Inter-
Collection Pap by
GP
Inter-
1996
173,265
813,131 1,199,523
165,705
775,255 1,042,821
7,560
37,876 156,702
1997
167,323
825,610 1,167,172
169,945
839,161 1,095,337
2,622
-13,551
1998
161,218
867,308 1,176,029
167,574
909,139 1,171,347
6,356
1999
152,362
891,876 1,207,414
160,524
942,069 1,201,152
8,162
2000
142,828
911,903 1,219,127
152,034
969,792 1,223,544
9,206
(INAMI-IMA)/INAMI]
Collection Pap by
GP
Inter-
specialist
pretation
4.4%
4.7%
13.1%
71,835
1.6%
-1.6%
6.2%
-41,831
4,682
3.9%
-4.8%
0.4%
-50,193
6,262
5.4%
-5.6%
0.5%
-57,889
-4,417
6.4%
-6.3%
-0.4%
Economical considerations
In 2003, the National Health Insurance Institute reimbursed 24,1 million EURO for
collection and interpretation of Pap smears from the uterine cervix (0.4 million for
collection by GPs, 3.2 million for collection by specialists and 20.5 million for
cytological interpretation). In total, 1.3 million smears were interpreted that year.
Considering the 3-year coverage measured over the period 1998-2000 in Table 9, we
can derive that 1.6 million in the target population of 25-64 years old women received
at least one Pap smear less than 3 years before. Three million (3.0) smears were
interpreted in that period taken from women in the target population. 1.1 million
women of the target population were not screened at all.
Assuming that 10 % of the overuse is spent for follow-upa, we can estimate that about
400,000 Paps are taken yearly that do not contribute to screening coverage or followup. This corresponds with an estimated capital of 8.2 million € per year reimbursed
by the National Health Insurance Institute for taking and reading Pap smears with
limited clinical utilityb.
Moreover, in the period 1998-2000, 0.6 million smears were taken from women
beyond the target age range, representing a yearly expenditure of 3.9 million EURO
for the INAMI.
a
In fact the proportion of smears for reason of clinical follow-up is not known for Belgium.
Nevertheless, according to an analysis of data from the provincial cervical cytology registers of
Flemish-Brabant and Antwerp for the period 1996-97, 91% of smears were taken for screening, 4% for
clinical indications and 6% for follow-up. This estimate should be considered with caution since the
variable “reason for the Pap smear” was only available for 67% of the registered cytological records
[Arbyn, 1999c].
b
The average unit cost at charge of the National Health Insurance Institute for one smear was compute
by dividing the total INAMI expenditure of 2003 for collection and interpretation of smears by the total
number of corresponding acts: € 3.24 for collection and € 16.21 for interpretation.
6/10/2005
71
90%
3-year screening coverage
Health Interview Survey, 2001
80%
HIS (95% CI)
70%
Individucal Health Insurance
Data, 2000
60%
50%
40%
30%
20%
10%
0%
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75+
Age group
Figure 37. Estimation of the 3-year cervical cancer screening coverage estimated from the national health interview survey (HIS) in 2001 (with 95% intervals) and the
corresponding coverage computed from individual health insurance date for the year 2000, Belgium, by age group.
6/10/2005
72
25
1,200,000
20
Costs (*106 EURO)
Number of Pap smears
1,000,000
800,000
15
600,000
10
400,000
Number of Pap smear interpretations
200,000
5
Costs for interpretation of smears
Costs for taking & intrepretation of smears
0
cvxriziv.xls/Pap_cht
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
0
2003
Year
Figure 38. Intensity and health insurance cost of cervical cancer screening in Belgium between 1983 en 2003: number of Pap smears interpreted (red curve) and costs for
interpretation (blue squares) and for collection & interpretation of cervical smears (RIZIV/WIV).
6/10/2005
73
450,000
4,500,000
Number of colposcopies
Costs
3,500,000
350,000
300,000
2,500,000
Expences (EUROS)
Number of colposcopies
400,000
250,000
200,000
1984
1,500,000
1986
1988
1990
1992
1994
1996
1998
2000
2002
Year
Figure 39. Number of colposcopic examinations performed and expenses for colposcopy spent between 1984 and 2003 in Belgium, as registered in the reimbursement
statistics of the National Health Insurance Institute.
6/10/2005
74
8.3.2.
Colposcopies
In Table 35 we compare the number of colposcopies derived from the two databases.
The difference (INAMI-IMA) was positive in 1996. Thereafter the difference became
progressively more negative (from –1.5% to –5.9%).
Table 35. Number of colposcopies by year reported by the National Health Insurance Institute
(INAMI) and computed from the individual patient data provided by IMA.
Yeaer
1996
1997
1998
1999
2000
Proportional
IMA
INAMI-IMA difference
378,750
23,947
5.9%
398,704
-5,702
-1.5%
408,634
-12,946
-3.3%
410,083
-16,338
-4.1%
413,785
-19,598
-5.0%
INAMI
402,697
393,002
395,688
393,745
394,187
The consumption of colposcopic examinations in Belgium is huge. On average, one
colposcopy is performed for every 3 Pap smears. Nevertheless the capital reimbursed
by the INAMI (4.2 million € in 2003) is not so elevated given the low unit price of
colposcopy (10.6 €, May 2004).
8.3.3.
Hysterectomies
The total number of total hysterectomies performed in Belgium each year between
1996 and 2000, reported by the INAMI and derived from the IMA file is shown in
Table 36. The number was always larger in the INAMI database, particularly in the
first two years, but the difference decreases over time.
Table 36. Number of total hysterectomies and cervix amputations by year reported by the National
Health Insurance Institute (INAMI) and computed from the individual patient data provided by IMA.
Year
1996
1997
1998
1999
2000
INAMI
19,862
18,685
18,954
18,237
15,810
IMA
16,622
16,982
18,138
17,368
15,454
INAMI- Proportional
IMA
difference
3,240
16.3%
1,703
9.1%
816
4.3%
869
4.8%
356
2.3%
In Figure 40 we give the trend in the number of all types of total hysterectomy and
cervix amputation performed each year in Belgium according to available statistics
published by the National Health Insurance Institute.
6/10/2005
75
20,000
18,000
16,000
number of interventions
14,000
12,000
10,000
8,000
6,000
4,000
2,000
0
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
year
Total abd. hysterectomy
Total vag. hysterectomy
Wertheim
Total hysterectomy & pelvic lymphadenectomy
Amputations cervix
Amputation resting cervix
Total
Figure 40. Number of total hysterectomies and cervix amputations performed each year in Belgium
between 1983 and 2003, according to the statistics published by the National Health Insurance Institute.
6/10/2005
76
8.4. Prevalence of cytological cervical lesions
In the framework of the Flemish cervical cancer screening programme, a cytological
register was set up. In Table 37, we show the overall distribution of cytological
findings and the range of minimum and maximum values from the first results from
provinces [Arbyn, 1999]. For more details, we refer to a chapter on cytological
registration in Gezondheidsindictoren 1997, published by the Ministerie van de
Vlaamse Gemeenschap [Arbyn, 1999]. These values are indicative to estimate the
need for follow-up cytology, colposcopy and histology.
Table 37. Distribution of specimen adequacy and cytological abnormalities, as registered from
laboratories in Flemish-Brabant and Antwerp (1996-1998.
Overall % (range: lowest-highest value)
Distribution of specimen adequacy
% Sub-optimal
Inadequate
% without endocervical cells
Prevalence of squamous lesions
ASCCUS
Low-grade SIL
High-grade SIL
Suspicion of cancer
Presence of glandular lesions
ASCUS, atypia, suspicion adenoca
Flemish-Brabant
Antwerp
175,260
47,721
38.5 (3.5-46.1)
0.6 (0.1-1.7)
27.2 (0.1-42.8)
25.7 (4.1-41.8)
0.7 (0.0-1.5)
21.3 (2.6-36.6)
1.6 (0.9-3.6)
0.7 (0.4-1.2)
0.5 (0.3-0.8)
0.17
2.1 (0.1-4.3)
1.1 (0.2-2.4)
0.5 (0.1-1.0)
0.06
0.6 (0.0-14.6)
1.8 (0.0-7.6)
6/10/2005
77
8.5. Conclusions
Screening coverage
The cervical cancer screening coverage, defined as the proportion of the target
population of women between 25 and 64 years old, that had a Pap smear taken in the
last 3 years, measured in 2000, was 59%. The range of variation in screening
coverage at the level of the Regions was small: 57% in the Flemish Region, 58% in
Brussels and 61% in the Walloon Region. The increase in coverage measured over
the period 1996-2000 is limited and respectively 2%, 5% and 3% in the Flemish,
Brussels and Walloon Region. This limited increase in the Flemish Region is
remarkable, given the fact that only in this region a screening campaign was
organized.
Differences at lower geographical levels are important: at provincial level in 2000, it
is situated between 52% (Luxembourg) and 69% (Walloon-Brabant); whereas at
district level the highest coverage is found in Huy (71%) and the lowest in Arlon
(only 40%).
At the national level, in the youngest 5-year age group of the target population,
women of 25-29 years old, the coverage is 64%. It increases until 67% among
women in the age range 30-39 years. From the age of 40 the coverage decreases
gradually until 56% in the 50-54 year age group. Thereafter the coverage declines
more rapidly. In the Brussels and Walloon Region this decline by age group is less
pronounced.
The coverage computed from the individual health insurance data file is substantially
lower than the estimates derived from the national health interview surveys. For
instance at national level, for the years 2000-2001, this difference was 11%. This
discrepancy is probably due to reporting biases, which are according to the
international literature inherent to interview surveys.
Screening interval and target age range
The modal screening interval is one year. A time span of 3 years is observed in only
5% of women with 2 or more smears in the studied time period. Ten percent of all
interpreted Pap smears are taken from women younger than 25 years and 7% in
women older than 64 years.
Pap smear use
The amount of used smears is theoretically sufficient to cover more than 100% of the
target population over a time span of 3 years. In 2000, the ratio of the number of Pap
smear interpretations over the concerned mid-year target population concerned was
1.1. This ratio should be higher if we were able to adjust the size of the target
population by excluding women whose cervix is removed. Nevertheless, only 59% is
covered with one or more smears.
In absolute figures: 3 million Pap smears were interpreted in the period 1998-2000
which were taken from only 1.6 million women in the age range 25 to 64 years. 1.1
million women did not get a Pap smear. The excess use of cervical cytological
examinations was 88%, which means that each screened woman received 1.8 smears
6/10/2005
78
over a 3-year period. The excess smear use was high in all parts of Belgium: it was
less high in the Flemish Region (86%), and was highest in the Capital Region (99%),
and intermediate in the Walloon Region (90%). Of course, a part of this “overuse” is
used for follow-up of women with previous cervical abnormality. Assuming that 10
% of the overuse is spent for follow-up, we can estimate that yearly about 400,000
Paps are taken that do not contribute to screening coverage or follow-up. This
corresponds with an estimated amount of € 8.2 million per year reimbursed by the
National Health Insurance Institute for taking and reading Pap smears with limited
utilitya. This amount is further increased with € 3.9 million for screening beyond the
target age range.
Colposcopy use
An impressive amount of colposcopies are performed in Belgium. At the national
level, on average, one colposcopic examination is done for every 3 Pap smears. In the
Walloon Region the ratio of the number of Pap smears over the number of
colposcopies is even less than two, whereas for the Flemish Region this ratio is five.
In certain Walloon districts the number of examined Pap smears equals the number of
performed colposcopies. The biopsy/colposcopy ratio is low (on average 5%) which
is due to the very high frequency of colposcopy in normal women.
It is clear that colposcopy is not used as indicated in national or international
guidelines. Colposcopic exploration is indicated in case of a first observation of HSIL
or glandular abnormality or after a second observation of ASC-US or LSIL or after a
positive HPV triage result in ASCUS-cases and in follow-up after treated lesions
[Arbyn, 1997; Arbyn 2004a: Arbyn; 2004b; BSCC, 2004; Wright, 2002].
General conclusion
The cervical cancer screening coverage in Belgium, defined as the proportion of
women in the 25-64 year age range that received a Pap smear in the last 3 years, is
only 59%. Nevertheless, the amount of smears used is theoretically sufficient to cover
the whole target population.
Estimation of the cervical cancer screening coverage, derived from interview surveys
should be interpreted with caution, given the inherent risk of overestimation.
Colposcopy is too often used simultaneously with the Pap smear, whereas its main
clinical indication is the exploration of women with previous cytological abnormality.
The overuse of colposcopy is systematic in the Walloon region.
Structural reduction of overuse and re-investment in coverage and quality
improvement can potentially result in more life-years saved, without increase in
public funding.
Measures foreseen in the European Council Recommendation on Cancer screening
should be binding and universal. Regionally isolated actions such as those undertaken
a
Costs for consultation by a gynaecologist or GP were not taken into account.
6/10/2005
79
in the Flemish screening programme, did not yield the desired result, because they
were based on voluntary participation without financial compensation for involved
health professionals.
Health authorities of the Federal and Community Governments and representatives of
the scientific societies should meet as soon as possible in order to define a rational,
evidence-based and cost-effective screening policy.
8.6. Propositions for further research
The discrepancy between the cervical cancer screening coverage, estimated from the
national health interview surveys and the coverage computed from individual patient
data provided by IMA, merits further research. An individual data linkage between
the health survey records and health insurance data files might be an interesting
initiative.
The reason for smear taking is insufficiently known. We have assumed that 10% of
the overused smears could be clinically indicated because of a previous minor cervical
abnormality or for surveillance after treatment of a high-grade lesion. This 10% was
based on previous analyses of two provincial screening registers. Longitudinal
analysis of the pooled data from all cytological registers is planned and will yield
more reliable information. Nevertheless, cytological registration remains until today
incomplete and the existing registers do not contain biopsy or treatment data.
The discrepancy between number of cervical smears and biopsies taken and
interpreted merits further investigation as well, preferentially using data files with
exact calendar data.
The evaluation of effectiveness of cervical cancer screening requires more performant
information systems which in the future must become mandatory, so that screening
and follow-up data can be linked to cancer registers and mortality registers, as
recommended in by the European Council.
9. Acknowledgements
We thank the administrators of the Europe Against Cancer Programme of the
European Commission (Directorate of SANCO, Luxemburg) for their support for the
European Network for Cervical Cancer Screening, in general, and for cervical cancer
screening in Belgium, in particular. At the same time, we acknowledge, Prof. Dr.
Herman Van Loon, Director of the General Directorate of Primary Health Care of the
Federal Public Service of Health, Food Chain Security and Environment for according
the matching Belgian funds for the European subsides, allowing us to conduct the
present study.
We are grateful to Dr. Valérie Fabri of the Mutualité socialiste for the coordination of
the pooling of data files provided by the health insurance agencies to the Agence
6/10/2005
80
Inter-mutualiste/Intermutualistisch Agentschap. Further, we acknowledge Dr. Jean
Taffereau, Dr. Johan Vanderheyden and Mr. Stefan Demarey for providing data from
the national Health Interview Surveys, and Mr. M. Breda and Mr. W. Vandeputte for
providing INAMI/RIZIV data.
We appreciated critical remarks from Dr. Valérie Fabri and Dr. Cinthia Lemos.
10. Glossarium
Screening coverage
Proportion of women having had a Pap smear within the period indicated by the
screening interval. Formula: # women with Pap < x time ago / midyear population
over period x. Global population coverages are computed for the 25-64 year age
group. For age-specific coverage, the age group reached at the end of the interval is
considered.
# smears / # women ratio
Ratio of the number of smears to the size of the mid-year population of women in a
given period.
Excess smear use and overuse
The percentage of all Pap smears that does not contribute to the coverage considered
over a certain screening interval, in this case 3 years, in a target age group.
Formula: [total number of smears taken in the target age group during a screening
period)*100 / number of smears needed to reach the observed screening coverage] 100% [Van Ballegooijen, 2000]. A related indicator is “overuse”: ratio of the number
of smears taken in the target age group during a screening period/number of smears
needed to reach the observed screening coverage. (Overuse-1)*100 gives the “excess
use” as a percentage.
Target population
The population of women at risk for cervical cancer, for whom screening is
recommended. In the European Guideline for Quality Assurance in Cervical Cancer
Screening cytological screening is recommended to target the age groups 25 to 64
years [Coleman, 1993].
Real target population
Population of women between 25-64 years, still having a cervix. The real target
population is computed as following: Popi* (1-Cumulative Incidence of total
hysterectomy up to agei.), where i is the index for 5-year age group.
Recommended screening interval
Period between 2 successive smears according to the screening policy, which is for
Belgium 3 years. The screening interval recommended in the European Guideline for
Quality Assurance in Cervical Cancer Screening is three to five years.
In this report we also compute the screening coverage and consumption of Pap smears
over a 1-year period, which, in opportunistic screening, is the usual interval.
Throughout the report it is always specified which screening interval is considered.
6/10/2005
81
11. Abbreviations
AIM: Agence Inter-Mutualitste
ASC-US: atypical squamous cells of unspecified significance
AGUS: atypical glandular cells of unspecified significance
CI: 95% confidence interval
EU: European Union
GP: general practitioner
HIS: Health Interview Survey
HSIL: high-grade squamos intra-epithelial lesions
IHID: individual health insurance data.
INAMI: Institut national pour l’Asurrance des Maladies et des Invalidités
IMA: Intermutalistisch Agentschap
ISP: Scientific Insitute of Public Health.
LSIL: low-grade squamos intra-epithelial lesions
NHII: National Health Insurance Institute (translation of RIZIV/INAMI)
RIZIV: Rijksinstituut voor Ziekte en Invaliditeitsverzekering
SHR: standaridised hysterectomy ratio
SPR: standardised Pap smear ratio
12. References
Arbyn M, Albertyn G, De Groof V. Consensus omtrent follow-up adviezen bij
cytologische screening naar baarmoederhalskanker. Tijdschr voor Geneeskunde 1996;
52: 709-15.
Arbyn M, Quataert P, Van Hal G, Van Oyen H. Cervical cancer screening in the
Flemish Region (Belgium): measurement of the attendance rate by telephone
interview. Eur J Cancer Prev 1997; 6: 389-98.
Arbyn M, Van Oyen H, Vergote I, Mertens R, Cluyse L, Diels J. Incidence and
prevalence of hysterectomy in the Flemish Region (Belgium). Full paper published
in: 11th International Meeting of Gynaecological Oncology; organized by the
European Society of Gynaecological Oncology, Budapest, May 8-12, 1999a.
Monduzzi Editore, International Proceedings Division, pp 189-193.
Arbyn M, Bourgain C, Cuvelier C, Drijkoningen M, Van Marck E, Willocx F and
WUCC. The Flemish Cervical Cancer Screening Register: creation and first results.
Acta Cytol 1999b; 43: 708-709.
Arbyn M, Van Oyen H, Cuvelier C, Nelen V, Van de Mieroop E, Willems H.
Registratie van cytologische screening naar baarmoederhalskanker. In: Gezondheidsindicatoren 97. Brussel: Ministerie van de Vlaamse Gemeenschap, 1999c, 104-7.
6/10/2005
82
Arbyn M, Van Oyen H, Mertens R, Vergote I. Incidence and prevalence of
hysterectomy in the Flemish Region (Belgium). Eur J Gynaecol Oncol 1999; 20: 116.
Arbyn M, Wallyn S, Van Oyen H, Nys H, Dhont J, Seutin B. The new privacy law in
Belgium: a legal basis for organised cancer screening. Eur J Health Law 1999d; 6:
401-407.
Arbyn M, Van Oyen H. Cervical cancer screening in Belgium. Eur J Cancer 2000; 36:
2191-2197.
Arbyn M, Geys H. Trend of cervical cancer mortality in Belgium (1954-94): tentative
solution for the certification problem of not specified uterine cancer. Int J Cancer
2002; 102: 649-654.
Arbyn M, Dillner J, Van Ranst M, Buntinx F, Martin-Hirsch P, Paraskevaidis E. Re:
Have we resolved how to triage equivocal cervical cytology? J Natl Cancer Inst
2004a; 96: 1401-1402.
Arbyn M, Buntinx F, Van Ranst M, Paraskevaidis E, Martin-Hirsch P, Dillner J.
Virologic verus cytologic triage of women with equivocal Pap smears: a metaanalysis of the accuracy to detect high-grade intraepithelial neoplasia. J Natl Cancer
Inst 2004b; 96: 280-293.
Bowman JA, Girgis A, Sasson-Fisher RW, SavolainenNJ, Hons BA. The accuracy of
self-reported Pap smear utilisation. Soc Sci Med 1997; 44: 969-976.
BSCC (Belgian Society of Clinical Cytology). Follow-up guidelines for cervical
cytology – Abnormal cervical smears. Presented at the Spring meeting of the
BVKC/SBCC, Aalst, 19 June, 2004.
Coleman D, Day N, Douglas G, Farmery E, Lynge E, Philip J et al. European
Guidelines for Quality Assurance in Cervical Cancer Screening. Europe Against
Cancer Programme. Eur J Cancer 1993; 29A Suppl 4: S1-S38.
Commission of the European Communities. Proposal for a Council Recommendation
on Cancer Screening. 2003/0093 (CNS), Brussels, 5 May 2004, pp 1-22.
Council of the European Union. Council Recommendation of 2 December on Cancer
Screening. Off J Eur Union 2003; 878: 34-38.
Demarest S, Van der Heyden J, Gisle L, Buziarist J, Miermans PJ, Sartor F et al.
Health Survey by Interview, Belgium 2001. pp 1-2480. 2002. Brussels, Scientific
Institute of Public Health. IPH/EPI REPORTS 2002-25.
IARC expert team. Cervix Cancer Screening Meeting, Lyon, 20-27 April 2004: preparation
of the IARC Handbook of Cancer Prevention – Volume 10, to be published in 2005.
Kleinbaum DG, Kupper LL, Morgenstern H. 1982. Chapter 6: Measures of disease
frequency. In: Epidemiologic Research, Principles and Quantitative Methods. Ed.
Van Nostrand Reinhold, New York, pp 96-116.
6/10/2005
83
Kottke TE, Trapp MA, Fores MM, Kelly AW, Jung SH, Novotny PJ, Panser LA.
Cancer screening behaviors and attitudes of women in southeastern Minnesota. JAMA
1995; 273: 1099-105.
McGovern PG, Lurrie N, Margolis KL, Slater JS. Accuracy of self-report of
mammorgraphy and Pap smear in low-income urban population. Am J Prev Med
1998; 14: 201-218.
Mertens, R. Hysterectomy in Belgium. Thematic Files of the CSF N° 1, May 1999,
pp 1-28.
Montano DE, Phillips WR. Cancer screening by primary care physicians: a
comparison of rates obtained from physician self-report, patient survey, and chart
audit. Am J Public Health 1995, 85: 795-800.
Peto J, Gilham C, Fletcher O, Matthews FE. The epidemic that screening has
prevented in the UK. Lancet 2004; 364: 249-256.
Sawyer JA, Earp JA, Fletcher RH, Daye FF, Wynn TM. Accuracy of women's self
report of their last pap smear. Am J Public Health 1989; 79: 1036-7.
van Ballegooijen M, van den Akker van Marle ME, Patnick J, Lynge E, Arbyn M,
Anttila A et al. Overview of important cervical cancer screening process values in
EU-countries, and tentative predictions of the corresponding effectiveness and costeffectiveness. Eur J Cancer 2000; 36: 2177-88.
Vlaams Kankerregistratienetwerk. Kankerincidentie in Vlaanderen 2000. Available
at: http://www.tegenkanker.be. Last consulted on 28 October, 2004.
Walter SD, Clarke EA, Hatcher J, Stitt LW (1988). A comparison of physician and
patient reports of Pap smear histories. J Clin Epidemiol 1988; 41: 401-410.
Wright TC, Cox JT, Massad LS, Wilkinson EJ. 2001 Consensus guidelines for the
management of women with cervical cytological abnormalities. JAMA 2002; 287:
2120-9.
6/10/2005
84
13. Annexes
13.1. Pap smear use at District level
Figure 41. Evolution of the one-year screening coverage and of the #smears/#women
ratio between 1996 and 200, by District.
Figure 42. Variation of the screening coverage and the # smears / # women ratio
considered over a 1-year interval according to 5-year age group, by District (Belgium,
2000).
Figure 43. Variation of the screening coverage and the # smears / # women ratio
considered over a 3-year interval according to 5-year age group (Belgium, by District,
1998-2000).
13.2. Profession of smear takers, by district
Figure 44. Relation between the % of Pap smears taken by GPs and age group for the
years 1996, 1998 and 2000, by District.
13.3. Colposcopies and cervical biopsies, by district
13.4. Hystectomies at District level
Figure 45. Average incidence of total hysterectomy by age (Belgium, by District,
1996-2000).
6/10/2005
85
Antwerpen
Mechelen
Turnhout
Brussels
Vilvoorde
Leuven
Nivelles
Brugge
Disksmuide
.4
.3
.2
.1
0
.4
.3
.2
.1
0
.4
.3
.2
.1
0
1996
1998
2000
1998
1996
2000
1996
1998
2000
Year
Ieper
Kortrijk
Oostende
Roeselare
Tielt
Veurne
Aalst
Dendermonde
Eeklo
.4
.3
.2
.1
0
.4
.3
.2
.1
0
.4
.3
.2
.1
0
1996
1998
2000
1996
1998
2000
1996
1998
2000
Year
1996 and 200, by District.
6/10/2005
86
Gent
Oudenaarde
St-Niklaas
Ath
Charleroi
Mons
Mouscron
Soignies
Thuin
.4
.3
.2
.1
0
.4
.3
.2
.1
0
.4
.3
.2
.1
0
1996
1998
2000
1996
1998
2000
1996
1998
2000
1998
2000
Year
Tournai
Huy
Liège
Verviers
Waremmes
Hasselt
Maaseik
Tongeren
Arlon
.4
.3
.2
.1
0
.4
.3
.2
.1
0
.4
.3
.2
.1
0
1996
1998
2000
1996
1998
2000
1996
Year
6/10/2005
87
Bastogne
Marche-en-Famenne
Neufchâteau
Virton
Dinant
Namur
.4
.3
.2
.1
0
.4
.3
.2
.1
0
1996
Phillipeville
1998
2000
1996
1998
.4
.3
.2
.1
0
1996
1998
2000
Year
6/10/2005
88
2000
Antwerpen
Mechelen
Turnhout
Brussels
Vilvoorde
Leuven
Nivelles
Brugge
Disksmuide
10 20 30 40 50 60 70
10 20 30 40 50 60 70
10 20 30 40 50 60 70
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
Ieper
Kortrijk
Oostende
Roeselare
Tielt
Veurne
Aalst
Dendermonde
Eeklo
10 20 30 40 50 60 70
10 20 30 40 50 60 70
10 20 30 40 50 60 70
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
Age
Figure 42. Variation of the screening coverage and the # smears / # women ratio considered over a 1year interval according to 5-year age group, by District (Belgium, 2000).
6/10/2005
89
Gent
Oudenaarde
St-Niklaas
Ath
Charleroi
Mons
Mouscron
Soignies
Thuin
10 20 30 40 50 60 70
10 20 30 40 50 60 70
10 20 30 40 50 60 70
Tournai
Huy
Liège
Verviers
Waremmes
Hasselt
Maaseik
Tongeren
Arlon
10 20 30 40 50 60 70
10 20 30 40 50 60 70
10 20 30 40 50 60 70
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
Age
6/10/2005
90
Bastogne
Marche-en-Famenne
Neufchâteau
Virton
Dinant
Namur
10 20 30 40 50 60 70
10 20 30 40 50 60 70
.5
.4
.3
.2
.1
0
.5
.4
.3
.2
.1
0
Phillipeville
.5
.4
.3
.2
.1
0
10 20 30 40 50 60 70
Age
6/10/2005
91
Antwerpen
Mechelen
Turnhout
Brussels
Vilvoorde
Leuven
Nivelles
Brugge
Disksmuide
10 20 30 40 50 60 70
10 20 30 40 50 60 70
10 20 30 40 50 60 70
1.5
1
.5
0
1.5
1
.5
0
1.5
1
.5
0
Age
Ieper
Kortrijk
Oostende
Roeselare
Tielt
Veurne
Aalst
Dendermonde
Eeklo
10 20 30 40 50 60 70
10 20 30 40 50 60 70
10 20 30 40 50 60 70
1.5
1
.5
0
1.5
1
.5
0
1.5
1
.5
0
Age
Figure 43. Variation of the screening coverage and the # smears / # women ratio considered over a 3year interval according to 5-year age group (Belgium, by District, 1998-2000).
6/10/2005
92
Gent
Oudenaarde
St-Niklaas
Ath
Charleroi
Mons
Mouscron
Soignies
Thuin
10 20 30 40 50 60 70
10 20 30 40 50 60 70
10 20 30 40 50 60 70
1.5
1
.5
0
1.5
1
.5
0
1.5
1
.5
0
Age
Tournai
Huy
Liège
Verviers
Waremmes
Hasselt
Maaseik
Tongeren
Arlon
10 20 30 40 50 60 70
10 20 30 40 50 60 70
10 20 30 40 50 60 70
1.5
1
.5
0
1.5
1
.5
0
1.5
1
.5
0
Age
6/10/2005
93
Bastogne
Marche-en-Famenne
Neufchâteau
Virton
Dinant
Namur
10 20 30 40 50 60 70
10 20 30 40 50 60 70
1.5
1
.5
0
1.5
1
.5
0
Phillipeville
1.5
1
.5
0
10 20 30 40 50 60 70
Age
6/10/2005
94
1996
2000
1998
Antwerpen
Mechelen
Turnhout
Brussels
Vilvoorde
Leuven
Nivelles
Brugge
Disksmuide
.6
.4
.2
0
.6
.4
.2
0
.6
.4
.2
0
0
20
40
60
80
0
20
40
60
80
0
20
40
60
80
40
60
80
Age
Ieper
Kortrijk
Oostende
Roeselare
Tielt
Veurne
Aalst
Dendermonde
Eeklo
.6
.4
.2
0
.6
.4
.2
0
.6
.4
.2
0
0
20
40
60
80
0
20
40
60
80
0
20
Age
Figure 44. Relation between the % of Pap smears taken by GPs and age group for the years 1996,
6/10/2005
95
1996
2000
1998
Gent
Oudenaarde
St-Niklaas
Ath
Charleroi
Mons
Mouscron
Soignies
Thuin
.6
.4
.2
0
.6
.4
.2
0
.6
.4
.2
0
0
20
40
60
80
0
20
40
60
80
0
20
40
60
80
40
60
80
Age
Huy
Liège
Verviers
Waremmes
Hasselt
Maaseik
Tongeren
Arlon
Bastogne
.6
.4
.2
0
.6
.4
.2
0
.6
.4
.2
0
0
20
40
60
80
0
20
40
60
80
0
20
Age
Figure 44 (continued). Relation between the % of Pap smears taken by GPs and age group for the
6/10/2005
96
1996
2000
1998
Marche-en-Famenne
Neufchâteau
Virton
Dinant
Namur
Phillipeville
.6
.4
.2
0
.6
.4
.2
0
0
20
40
60
80
0
20
40
60
80
0
20
40
60
80
Age
Figure 44 (continued). Relation between the % of Pap smears taken by GPs and age group for the
6/10/2005
97
Table 38. Number of colposcopies, cervical biopsies taken and Pap smears interpreted, ratio of the
number of biopsies over the number of colposcopies and the ratio of the number of Pap smears over the
number of colposcopies, 1996-2000, by district and ranked increasing by Pap smear/colposcopy ratio.
District
Mouscron
Huy
Soignies
Tournai
Mons
Waremmes
Ath
Bastogne
Dinant
Liège
Virton
Unknown
Arlon
Thuin
Namur
Marche-en-F.
Verviers
Living abroad
Charleroi
Nivelles
Phillipeville
Neufchâteau
Brussels
Tongeren
Vilvoorde
Roeselare
Oudenaarde
Hasselt
Dendermonde
Antwerpen
Aalst
Oostende
Turnhout
Leuven
Disksmuide
Tielt
Eeklo
St-Niklaas
Mechelen
Gent
Kortrijk
Veurne
Maaseik
Ieper
Brugge
Colposcopies Biopsies Pap smears
24,586
330
25,667
58,492 1,958
69,862
89,181 1,629
109,031
62,831 1,210
78,367
116,261 2,088
145,926
33,701 3,173
44,767
32,299 1,401
42,915
13,912
138
19,421
36,802 1,086
51,704
249,396 7,094
379,438
13,653
281
21,732
2,433
651
3,921
10,899
138
18,693
46,737 1,428
83,721
89,643 3,697
164,277
14,997
454
28,142
77,396 1,838
150,865
5,764
204
11,580
115,620 5,405
234,237
124,576 4,647
255,923
14,497
661
35,898
8,916
439
22,918
207,587 12,143
586,105
27,212
962
94,103
97,786 4,796
367,528
14,703
458
58,587
14,095 3,629
56,438
46,334 4,094
200,457
20,264
822
96,755
104,308 15,509
504,660
27,206 1,311
137,373
13,532 1,439
72,399
36,173 3,574
204,377
38,736 5,788
246,320
2,847
125
18,269
5,693
243
40,167
4,970
639
35,403
15,864 1,156
126,931
18,232 2,033
160,232
32,111 2,402
283,896
14,037 1,945
132,314
2,552
281
27,901
9,140
608
102,688
4,414 1,359
49,812
9,568
872
132,481
ratio
ratio
ratio
1.3%
1.0
1.3%
3.3%
1.2
2.8%
1.8%
1.2
1.5%
1.9%
1.2
1.5%
1.8%
1.3
1.4%
9.4%
1.3
7.1%
4.3%
1.3
3.3%
1.0%
1.4
0.7%
3.0%
1.4
2.1%
2.8%
1.5
1.9%
2.1%
1.6
1.3%
26.8%
1.6
16.6%
1.3%
1.7
0.7%
3.1%
1.8
1.7%
4.1%
1.8
2.3%
3.0%
1.9
1.6%
2.4%
1.9
1.2%
3.5%
2.0
1.8%
4.7%
2.0
2.3%
3.7%
2.1
1.8%
4.6%
2.5
1.8%
4.9%
2.6
1.9%
5.8%
2.8
2.1%
3.5%
3.5
1.0%
4.9%
3.8
1.3%
3.1%
4.0
0.8%
25.7%
4.0
6.4%
8.8%
4.3
2.0%
4.1%
4.8
0.8%
14.9%
4.8
3.1%
4.8%
5.0
1.0%
10.6%
5.4
2.0%
9.9%
5.6
1.7%
14.9%
6.4
2.3%
4.4%
6.4
0.7%
4.3%
7.1
0.6%
12.9%
7.1
1.8%
7.3%
8.0
0.9%
11.2%
8.8
1.3%
7.5%
8.8
0.8%
13.9%
9.4
1.5%
11.0%
10.9
1.0%
6.7%
11.2
0.6%
30.8%
11.3
2.7%
9.1%
13.8
0.7%
6/10/2005
98
Antwerpen
Mechelen
Turnhout
Brussels
Vilvoorde
Leuven
Nivelles
Brugge
Disksmuide
15
10
5
0
15
10
5
0
15
10
5
0
25
40
55
70
25
40
55
70
25
40
55
70
55
70
Age group
Ieper
Kortrijk
Oostende
Roeselare
Tielt
Veurne
Aalst
Dendermonde
Eeklo
15
10
5
0
15
10
5
0
15
10
5
0
25
40
55
70
25
40
55
70
25
40
Age group
Figure 45. Average incidence of total hysterectomy by age (Belgium, by District, 1996-2000).
6/10/2005
99
Gent
Oudenaarde
St-Niklaas
Ath
Charleroi
Mons
Mouscron
Soignies
Thuin
15
10
5
0
15
10
5
0
15
10
5
0
25
40
55
70
25
40
55
70
25
40
55
70
55
70
Age group
Tournai
Huy
Liège
Verviers
Waremmes
Hasselt
Maaseik
Tongeren
Arlon
15
10
5
0
15
10
5
0
15
10
5
0
25
40
55
70
25
40
55
70
25
40
Age group
6/10/2005
100
Bastogne
Marche-en-Famenne
Neufchâteau
Virton
Dinant
Namur
15
10
5
0
15
10
5
0
25
Phillipeville
40
55
70
25
40
55
70
15
10
5
0
25
40
55
70
Age group
6/10/2005
101
13.5. Data files and statistical syntaxis files
-
directory: /intermut/
pap smear nterpretation: several do files
profession of smear takers: takingpap.do
colposcopy and cervical biopsies: colpo.do; colpo.xls
hysterectomy: hyster.do; hyster.xls
screening in hysterectomised women: cyt_hyst.do; cyt_hyst.xls
district maps:
o 3-year Pap smear coverage: Arr3y.wor
o Proportion smears taken by GPs: takingpap.wor
o Hysterectomy rate: Inchyster.wor
-
Comparison with health interview survey:
o Directory: \his\
o Command files: his.do; hyster.do
-
Data from RIVIZ: \xls\cvxriziv.xls; cvxrizivtotal.xls
6/10/2005
102

analysis of individual health insurance data pertaining to pap

Transcription

Similar documents

JFK`tje 01

Untitled - Studio Berkhout

the blue folder De 9 Straatjes

Untitled - First Covenant Church

2015 Beach Volleyball

HAIDER ACKERMANN ELSA BY ELS PROOST

LCPL-Pathan jaarverslag van het jaar 2009-2010

Additional national page for Country Report Issue 9