Abstracts

Rationale: Limited data are available regarding the lack of treatment of patients with epilepsy in the US. We therefore estimated the proportion of newly diagnosed epilepsy patients who remain untreated after diagnosis, and compared the characteristics of treated and untreated patients in a US population followed for at least 6 months. Methods: This retrospective cohort study analyzed data from the US-based PharMetrics Plus claims database that was linked to the IMS LRx prescription database, which captures information on dispensed prescriptions sourced from retail, mail, long-term care, and specialty pharmacies, including cash paid prescriptions, and therefore has a wider coverage than the PharMetrics database and an electronic medical records (EMR) database. Incident epilepsy cases with an index date between Jan 2009 and Dec 2014 were identified in the PharMetrics database. A 2-year baseline pre-index date was required for those aged >2 years (1 year for those aged 1 to < 2 years; no enrollment period for those < 1 year). Epilepsy cases were defined using ICD?"9-CM codes (Table 1). Patients were required to have continuous medical and pharmacy enrollment and no claims records with a diagnosis of epilepsy in the PharMetrics database, or pharmacy claim for antiepileptic drugs (AEDs) in the PharMetrics and/or LRx Prescription (if linked) databases during the baseline period. Patients were followed for ?-180 days after the index date. In the follow-up period, information was obtained on AED prescriptions in the PharMetrics and LRx prescription databases. The proportion of untreated epilepsy patients was estimated, and the baseline characteristics of patients with epilepsy who were untreated were compared to treated patients. Results: A total of 37,338 incident cases of epilepsy were identified, of whom 21,680 cases were linked to the LRx database. Overall, 10,510 cases (28.2%) in the PharMetrics database did not have an AED prescription during the follow-up period; however, 651 patients had ?-1 AED prescription in the LRx database and 1 patient in the EMR database. Therefore, more than a quarter of the cases (9,859 cases [26.4%]) remained untreated during follow-up from 6 months to 3 years after diagnosis. Based on data from patients that were linked to the LRx or EMR databases, there were no differences in the baseline characteristics between the treated (n=652) and untreated (n=4,909) patients except for the prevalence of stroke which was more frequent in the treated patients (Table 2). Conclusions: A significant proportion of newly diagnosed patients remain untreated for a period of up to 3 years after an epilepsy diagnosis. Reasons may include prescription drugs not being captured in databases, misclassification of epilepsy diagnosis, a decision not to treat by physician and patients, and economic constraints. Further investigations are needed to validate diagnosis of epilepsy in electronic medical records databases. Funding: UCB Pharma-sponsored.