Abstracts

Rationale: Epilepsy affects people of all ages, gender and socioeconomic status and accounts for one percent of the global burden of disease. The prevalence of epilepsy is expected to rise with aging populations. It is thus necessary to develop surveillance programs to monitor and project the future social and clinical demands of epilepsy. The purpose of this study was to assess the validity of clinical administrative ICD-9 data to define epilepsy cases and to compare coding variations between hospital settings (adult vs. children, teaching vs. community).Methods: All emergency and inpatient visits in calendar year 2000 coded in the primary diagnostic position with the ICD-9 epilepsy, transient ischemic attack, syncope or classical migraine codes were identified for patients of all ages. A randomly selected sample of charts coded for each of the above conditions (n=424) were then blindly reviewed by an epileptologist and a neurology resident and independently coded with an ICD-9 code. Results: Epilepsy coding was equally good regardless of hospital setting (community or teaching site, adult or pediatric site) with high sensitivity (Sn) and specificity (Sp). Sn and Sp were highest for inpatient admissions with Sn 98% and Sp 99%. Agreement in coding between the epileptologist and the professional coders or between the epileptologist and the neurology resident was excellent with a Kappa value of 0.97 and 0.98 respectively.Conclusions: Epilepsy cases can be accurately identified using emergency and inpatient administrative data. Validation of outpatient databases will be required to develop a comprehensive population-based epilepsy surveillance program.