What do researchers say about proposed Medicare claims public use files?

Medicare claims data sets have been used widely in outcomes research over the last decade. In 2010, the Centers for Medicare & Medicaid Services established a program to create de-identified basic standalone public use files (PUFs) from Medicare claims data, each one containing claims information from a 5% sample of beneficiaries.

We conducted a series of structured key informant interviews with research stakeholders to compile recommendations that would guide the creation of these PUFs. In this paper, we describe the interview methodology and present our findings. Fifteen researchers, representing a range of clinical health services and health policy expertise, were interviewed.

All respondents supported the use of Medicare claims in comparative effectiveness research and responded favorably to the creation of PUFs for this purpose. The interviews resulted in administrative-, technical- and content-related recommendations, some of which led to important changes in the PUFs.

A primary trade-off in the development of the proposed PUFs involved assuring maximum research utility of the files while assuring security of beneficiaries' protected health information. Protection of protected health information was considered a requirement. Given this constraint, the proposed PUFs may be most useful for two primary activities in comparative effectiveness research: first, working through the beginning stages of a research project; and second, examining high-level questions.