Research use of insurance claims data presents unique challenges and requires a series of value judgments that are intended to improve the data quality. In this study, medical insurance claims from 2 large companies were combined to assess utilization of complementary and alternative medicine. Challenges included assessing and improving the quality of data, combining data from 2 different companies with dissimilar coding systems, and determining the most appropriate ways to describe utilization. This article addresses 4 methodologic challenges in creating the analytic files: (1) conversion of claims into unique visits, (2) identification of incomplete claims data, (3) categorization of providers and locations of service, and (4) selecting the most useful measures of utilization and expenditures.