Administrative health-care claims databases offer a number of attractive features from which to form the evidence base for conducting pharmacoeconomic studies. These features include the ready availability of data, real-world health-care practice patterns, and potentially large sample sizes and long follow-up. However the ease of access to retrospective data also increase the potential for analytic mining and spurious findings, especially in the absence of prespecified analytic plans. Hence, I believe that the publication of the report of the ISPOR Task Force on Retrospective Databases in this issue represents an important and timely step toward ensuring the quality of published pharmacoeconomic evaluations that use claims data as the basis for their evidence.