Abstracts

Rationale: One of the Veterans Health Administration Epilepsy Centers of Excellence (VHA ECoE) missions is to obtain accurate estimates of disease frequency measures in the VHA for surveillance and distribution of resources. Using a validated algorithm, the ECoE started reporting epilepsy prevalence in Fiscal Year (FY) 2012. From FY12 to FY15 the annual prevalence ranged between 13.8 to 14.2 per 1,000 VHA patients. The algorithm used had a positive predictive value (PPV) of 82.0% (95% confidence interval (CI): 78.6% to 85.4%). After the implementation of ICD-10 in the VHA, it was important to verify that the algorithm continued to accurately identify Veterans with epilepsy when using the new diagnosis codes. Methods: The VHA Support Service Center (VSSC) and Pharmacy Benefits Management (PBM) data were used. Patients who were given an ICD-09 diagnosis code 345.* (epilepsy) or 780.39 (other convulsions) between FY14-FY15 or ICD-10 diagnosis code G40.* (epilepsy), R40.4 (Transient Alteration of Awareness), R56.1 (Post Traumatic Seizures) or R56.9 (Unspecified Convulsions) on clinical encounters (excluding data from diagnostic clinics) during FY16 were initially considered. Among those, patients prescribed an anti-seizure drug (ASD) during FY16 were identified as having epilepsy. Also, in the updated algorithm gabapentin was excluded as an ASD unless associated with a G40.* or 345.* diagnosis due to controversial views on gabapentin prescribed for epilepsy. Results: A total of 79, 562 Veterans were identified as epilepsy patients with an estimated prevalence of 12.6 per 1,000 VHA patients during FY16. Most were males (91.8%). The estimated prevalence per 1,000 patients among males was 12.9 and among females was 10.1. Among the Operation Enduring Freedom, Iraqi Freedom, New Dawn (OEF/OIF/OND) Veterans the estimate of prevalence was 8.0 per 1,000. Chart evaluation of 625 randomly selected patients by epilepsy experts showed an estimated PPV of 85.1% (95% CI: 82.1% to 87.8%) of the algorithm used for the identification of epilepsy cohort. Conclusions: The algorithm used had a high accuracy for the identification of epilepsy patients. Continuous surveillance of epilepsy patients in the VHA is imperative for determining the trends of epilepsy prevalence among VHA patients in general, to include OEF/OIF/OND Veterans. In this regard, ECoE can play a pivotal role for the optimal specialty care of epilepsy patients in the VHA. Funding: No funding