Objective
Evaluate the accuracy of an algorithm at identifying ethnic minorities from administrative claims for enrollment into a clinical trial.
Data Sources/Study Setting
Claims data from a health benefits company.
Study Design
We compared results of a three-step algorithm to self-reported race/ethnicity.
Data Collection/Extraction Methods
Using the algorithm, we identified subjects with high probability of being minority and ascertained self-reported race/ethnicity.
Principal Findings
We identified 164 subjects as likely minority based on our algorithm. Of these, 94 completed the survey and 87 identified themselves as black or Hispanic. The positive predictive value of the algorithm was 93 percent (CI: 85–97).
Conclusions
Claims data can be used to efficiently identify minorities for participation in clinical trials.