Women's Benefits and Harms Trade-Offs in Breast Cancer Screening: Results from a Discrete-Choice Experiment.

BACKGROUND: Over the past decade, the benefits and harms balance of breast cancer (BC) screening has been widely debated. OBJECTIVES: To elicit women's trade-offs between the benefits and harms of BC screening and to analyze the main determinants of these trade-offs. METHODS: A discrete-choice experiment with seven attributes depicting BC screening programs including varying levels of BC mortality, overdiagnosis, and false-positive result was used. Eight hundred twelve women aged 40 to 74 years with no personal history of BC recruited by a survey institute and representative of the French general population (age, socioeconomic level, and geographical location) completed the discrete-choice experiment. Preference heterogeneity was investigated using generalized multinomial logit models from which individual trade-offs were derived, and their main determinants were assessed using generalized linear models. Screening acceptance rates under various benefits and harms ratios were simulated on the basis of the distribution of individual preferences. RESULTS: The women would be willing to accept on average 14.1 overdiagnosis cases (median = 9.6) and 47.8 false-positive results (median = 27.2) to avoid one BC-related death. After accounting for preference heterogeneity, less than 50% of women would be willing to accept 10 overdiagnosis cases for one BC-related death avoided. Screening acceptance rates were higher among women with higher socioeconomic level and lower among women with poor health. CONCLUSIONS: Women are sensitive to both the benefits and the harms of BC screening and their preferences are highly heterogeneous. Our study provides useful results for public health authorities and clinicians willing to improve their recommendations of BC screening on the basis of women's preferences. Copyright (c) 2018 International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. All rights reserved.