QUANTITATIVE ESTIMATES OF THE IMPACT OF SENSITIVITY AND SPECIFICITY IN MAMMOGRAPHIC SCREENING IN GERMANY

Study objective - To estimate quantitatively the impact of the quality of mammographic screening (in terms of sensitivity and specificity) o n the effects and costs of nationwide breast cancer screening. Design - Three plausible ''quality'' scenarios for a biennial breast cancer s creening programme for women aged 50-69 in Germany were analysed in te rms of costs and effects using the Microsimulation Screening Analysis model on breast cancer screening and the natural history of breast can cer. Firstly, sensitivity and specificity in the expected situation (o r ''baseline'' scenario) were estimated from a model based analysis of empirical data from 35 000 screening examinations in two German pilot projects. In the second ''high quality'' scenario, these properties w ere based on the more favourable diagnostic results from breast cancer screening projects and the nationwide programme in The Netherlands. T hirdly, a worst case, ''low quality'' hypothetical scenario with a 25% lower sensitivity than that experienced in The Netherlands was analys ed. Setting - The epidemiological and social situation in Germany in r elation to mass screening for breast cancer. Results - In the ''baseli ne'' scenario, an 11% reduction in breast cancer mortality was expecte d in the total German female population, ie 2100 breast cancer deaths would be prevented per year. It was estimated that the ''high quality' ' scenario, based on Dutch experience, would lead to the prevention of an additional 200 deaths per year and would also cut the number of fa lse positive biopsy results by half. The cost per life year gained var ied from Deutsche mark (DM) 15 000 in the ''high quality'' scenario to DM. 21 000 in the ''low quality'' setting. Conclusions - Up to 20% of the total costs of a screening programme can be spent on quality impr ovement in order to achieve a substantially higher reduction in mortal ity and reduce undesirable side effects while retaining the same cost effectiveness ratio as that estimated from the German data.