Abstracts

Rationale: Anti-epileptic drugs (AEDs) differ in several different attributes including efficacy, adverse event (AE) profile, and dosing. Given the chronic nature of the disease, these attributes are meaningful to patients for whom AEDs are a life-long consideration. Empirically however, little is known about how patient characteristics and AED attributes interact to form AED preferences. Methods: Adult survey respondents currently treated with an AED were recruited from nationally-representative US Harris Panel to complete an online survey that included a discrete choice experiment (DCE). Respondents first provided relevant demographic and seizure-related clinical information, and then completed two validated instruments to assess generic and disease-related quality of life: the EQ-5D-5L, and the Impact of Epilepsy Scale (IES), respectively. Subsequently, a DCE was administered whereby respondents viewed 15 sets of factorial designed hypothetical epilepsy treatments (2 treatments/set) and selected the preferred treatment from each set. Hypothetical treatments were characterized by 6 attributes pertaining to efficacy, adverse effects, and dosing. Preferences were derived from respondents' selections by a hierarchical Bayes estimation, and the importance of each AED attribute was computed as a percentage. The impacts of patient characteristics on the relative preference for each AED attribute were computed by linear regression. Results: Among 518 respondents, 71% were female; mean age was 21.8. Variables listed in Table 1 served as covariates for the linear regression in order to determine the magnitude and directionality of any associations between these variables and the importance of each AED attribute (Table 2). Respondents with worse current seizure control placed significantly more importance in AED effectiveness versus other attributes. Conversely, respondents with better current seizure control placed significantly more importance in avoiding psychiatric AEs versus other attributes (p < 0.05). The IES and EQ-5D had a positive association with the importance of treatment effectiveness, while the IES had a negative association with daily administration frequency, associated fatigue / drowsiness, and dietary restrictions, and the EQ-5D had a negative association with psychiatric side effects. Age and the presence of tonic- clonic seizures, were both negatively associated with dietary restrictions. No significant relationships were observed for gender, disease duration, insurance type, number of seizure days in the last 30 days, or physician type. Conclusions: Specific demographic, clinical, and QOL characteristics were significant drivers of AED treatment preferences. Further, respondents with higher QOL scores and those reporting better current epilepsy control were less willing to accept greater risks of adverse events, while younger patients, and those with higher QOL scores, were less accepting of increased administration frequency and increased dietary restrictions. Funding: This study was sponsored by Sunovion, where FV is an employee. JC is an employee of EPI-Q, which received funding from Sunovion associated with the development and execution of this study. AE received no compensation (monetary or otherwise) for his role as an investigator in this study.