Numerous studies have identified disparities in epilepsy care across racial and ethnic groups. However, the role of physician decision making has yet to be explored as a factor influencing disparities in epilepsy management.

Our study evaluated clinic visits from patients seen at the Penn Comprehensive Epilepsy Center between 2008-2023. We analyzed electronic health records data and extracted outcome variables using a previously validated NLP approach (Xie et al. 2022). Our primary independent variables were race, defined as Black, White, or other races, and ethnicity, defined as Hispanic/Latinx and non-Hispanic/Latinx. Our primary outcome measure was an increase in seizure medication dose in response to a breakthrough seizure. The data were analyzed using a generalized estimating equation multiple logistic regression, controlling for demographic and socioeconomic variables.

Our study evaluated 5515 patients and 35588 clinic visits. Of the patients included 2472 (45%) were male. 3468 (63%) were White, 1660 (30%) were Black, and 387 (7%) were other races. 136 (2%) were Hispanic/Latinx. In our unadjusted model, Black race (OR 1.00, 95% CI 0.95-1.06) and Hispanic/Latinx ethnicity (OR 1.08, 95% CI (0.95-1.24) were not significantly associated with dose increases in response to breakthrough seizures. Female sex (OR 0.94, 95% CI 0.89-0.99), lower income (OR 0.91, 95% CI 0.85-0.98), older age (OR 0.80, 95% CI 0.75-0.87), and public insurance (OR 0.91, 95% CI 0.87-0.96) were associated with a significantly decreased likelihood of increase in seizure medication dose in response to a breakthrough seizure. In our adjusted model, the associations between age, insurance, and income persisted.

Our study suggests treatment decisions by providers in response to breakthrough seizures differed according to several demographic variables and may contribute to disparities in outcomes.