Abstracts

Rationale: To perform a nationwide epidemiologic study for treated epilepsy, we trained the professional medical record administrators (MRAs) to review the record associated with epilepsy with a structured case record form (CRF). In this study, we intended to validate the medical record survey with trained MRAs.Methods: Subjects for survey were the 80 patients who were prescribed anticonvulsants for one year in 2009 at two tertiary hospitals. The contents of CRF were demography, anticonvulsants, diagnosis, disease activity, type of seizure, cause and classification of epilepsy, and results of the electroencephalography and brain imaging. The CRF consists of two parts, MRAs documentation and epileptologists interpretation of the documentation collected by MRAs. For validation, two MRAs and one in-house epileptologist documented the medical records of same patients independently. Interpretation was done by an independent epileptologist for MRAs documentation. We assessed the accuracy, sensitivity and specificity of the results by MRAs, using those by the in-house epileptologist as standard. For the test-retest reliability, MRAs repeatedly surveyed the same patients. Results: The accuracy of the survey conducted by trained MRAs was 95% to 100% for the diagnosis and disease activity and 85% to 100% for the cause and classification. The sensitivity was 90.9% to 100% for diagnosis and 100% for disease activity. The specificity was 100% in diagnosis and activity. The test-retest reliability was higher than 90% for diagnosis, activity and classification. The reliabilty for cause of epilepsy was 81% to 85%. Conclusions: Our structured CRF was appropriate for epilepsy survey using medical records, and trained MRAs performed as successfully as an epileptologist in terms of collecting the information associated with epilepsy. The test-retest reliability of MRAs survey was also good. This method is useful for large-scaled epidemiologic studies of epilepsy.