Building a model to determine the cost-effectiveness of breast cancer screening in France.

This paper describes the methods and initial validation of a cost-effectiveness model developed to simulate the breast cancer screening situation in France. The first screening pilot programmes were set up in France in 1989 to test the feasibility of a decentralized screening model based in a large number of existing non-dedicated radiology centres. The present cost-effectiveness model was built as a tool to help guide current policy discussions on the future of screening in France. This Markov model compares the costs and effects expected when a screening programme is offered to a given cohort of women to those expected in the absence of screening. The model was initially validated using current results from the Bas-Rhin screening programme and local cancer registry epidemiological data. Over a 20-year period, 315 274 women would attend for screening, of whom 12 491 would be recalled for further assessment. 4423 cancers would be detected, resulting in 637 deaths. Screening allows the detection of 106 additional cancer cases, thereby preventing 92 deaths, and saves 1522 life-years compared with a situation without screening. Breast cancer mortality is reduced by 12.6%, yielding a cost-effectiveness ratio of 137 000 FF per life-year saved. The results of initial analyses suggest that the model is capable of suitably assessing the impact of breast cancer screening in terms of costs and effects. Further scenario analyses are needed to understand the impact of screening policy changes on the costs and effectiveness of future screening programmes.