Epilepsy-related research studies that rely on medical claims data have significant limitations that can be improved by choosing better selection algorithms. When Moura et al performed a medical chart review of 1377 patients, they found that the best algorithms to identify people with epilepsy included diagnostic and prescription drug data, as well as current antiepileptic drug usage, the location of the site that patients received medical care, and the specialty of the physician providing that care.
Moura LM, Price M, Cole AJ, et al. Accuracy of claims-based algorithms for epilepsy research: Revealing the unseen performance of claims-based studies. Epilepsia. 2017; Feb 15. doi: 10.1111/epi.13691. [Epub ahead of print]