The goal of this project was to (1) assess the validity of medical claims information for tracking breast cancer diagnoses, treatments, and outcomes; (2) use Medicare data, linked SEER cancer registry data, and claims data from large firms to analyze trends in diagnosis rates and staging, treatment, expenditures, and outcomes for Americans with breast cancer; and (3) analyze the cost effectiveness of alternative patterns of breast cancer diagnosis and treatment. The first part of Section B reviews the work done on creating cohorts of breast cancer patients and assessing the validity of the selection procedure. The second part discusses our findings on trends in breast cancer costs, treatments, and outcomes with a particular emphasis on changes in the productivity of breast cancer care. The third part reviews our work on measuring the cost effectiveness of alternative patterns of diagnosis and treatment, and introduces some innovative statistical methods we developed to help address the problem of bias. The conclusion section reviews the main findings and describes the areas for future research identified by this project.