The 3 primary administrative data sets developed by the Centers for Medicare and Medicaid services (CMS) to support the Medicare Part D program implementation represent a valuable source of data for health services researchers. This paper describes the structure of the Medicare Part D program and the related databases, and discusses their utilization for research purposes. The Medicare Part D administrative data include information on plan benefits (integrated into the Health Plan Management System), beneficiary enrollment files, and prescription drug event (PDE) claims-type data. The enrollment data may be of use in investigating the benefits and plan types that appeal to beneficiaries, but their application is limited by the fact that, although individual beneficiaries' enrollment choices are recorded, only summary enrollment data are currently publicly available. PDE data are likely to be of most interest to researchers as they are detailed (including beneficiary identifiers, contract identifiers pharmacy provider information on drugs provided, drug cost, and insurance status), beneficiary-specific (allowing them to be linked to beneficiary characteristics), and an unusual output for a program reimbursed under a capitation-based system. Because PDE data are highly sensitive, only summary data on the number of Part D prescriptions filled are publicly available. Although the data collected in relation to the Medicare Part D program could be applied to many questions of interest to health services researchers, their utility is limited by the sensitive natures of many of these data, making it difficult currently to obtain access for research purposes.