Meeting Brochure and registration form      SMDM Homepage

Tuesday, October 23, 2007
P3-26

LESSONS FROM COMBINING SERVICE PLANNING MODELS WITH HEALTH ECONOMIC DECISION MODELS: A CASE STUDY ON COLPOSCOPY REFERRAL GUIDELINES IN THE ENGLISH CERVICAL SCREENING PROGRAMME

Alan Brennan, BSc, MSc1, Hazel Pilgrim1, Guillaume Le Douaron1, Richard Hadwin2, Patrick Walker2, Simon Eggington3, and Julietta Patnick4. (1) University of Sheffield, SHEFFIELD, United Kingdom, (2) Royal Free Hospital, London, United Kingdom, (3) IMS Health, London, United Kingdom, (4) NHS Cancer Screening Programme, Sheffield, United Kingdom

Purpose: To investigate feasibility and issues arising when a Service Planning Model is merged with Health Economic Decision Model. The case study evaluates the effects on English NHS colposcopy services of implementing new referral guidelines. Current guidelines suggest referral to colposcopy after two consecutive cervical smear tests showing mild dyskaryosis. New guidelines require referral to colposcopy after just a single mild dyskaryosis result. This will result in additional colposcopy workload, and extra costs but potentially reduced numbers of invasive cancers and improved survival.

Method: A mathematical model of each local colposcopy service in England using clinical pathways from smear result through to treatment has been constructed in Excel. The model incorporates questionnaire data on referral numbers and management practices, routine data and published research results. It predicts national average workload and impact in differing local services. A separate long term health economic model examines the costs and predicted change in quality-adjusted life years (QALYs) in Cervical Screening Programmes. We have merged these two models to assess the capacity impact and incremental cost-effectiveness of policy change in different local services and at the national level.

Results: Single mild dyskaryosis referral implies, on average, a 21.7% increase in colposcopy workload. Sensitivity analyses identify areas of current practice which may be modified to reduce the workload impact. The new screening strategy appears cost-effective in the national service economic analysis. However, variations in practice affect results substantially in terms of both capacity requirements and cost-effectiveness of the guideline change.

Conclusions: The general approach of integrating service planning modelling with health economic analysis was successful and can apply widely. Particular issues arose aligning different age structures and mismatched health state definitions. A user-friendly model is available via the NHS Cancer Screening Programme website and we aim to make the merged Health Economic version available also. This can be a useful tool for practitioners where local variation is wide.