A comparison of estimators for the harms of repeat cancer screening for use in health policy decision making
*Rebecca A. Hubbard, Group Health Research Institute 

Keywords: cancer screening, mammography, screening test performance

In 2009, the United States Preventive Services Task Force (USPSTF) released their updated screening mammography guidelines. These were based on a systematic review of evidence on the harms and benefits of mammography from randomized trials and observational data on mammography performance in community practice. These revised guidelines have provoked ongoing controversy and raise questions about the most appropriate measures for evaluating the harms and benefits of a screening test. In this talk we will review estimators of the harms of cancer screening used by the USPSTF as well as some of those proposed by critics of the USPSTF guidelines. We will then introduce an alternative class of estimators that more appropriately account for features of observational study designs common in evaluation of cancer screening tests. We demonstrate the implications of the choice of estimator for inference on the harms of repeat breast cancer screening using data on screening mammography.