Acute care utilization in patients with epilepsy on antiepileptic monotherapy

To identify clinical and demographic predictors of urgent health care services utilization in monotherapy-treated patients with epilepsy. Patients aged 18-64 years with a primary or secondary diagnosis of epilepsy and >1 prescription claim for an antiepileptic drug (AED) pre-index were included. Innovus Invision™ Data Mart insurance claims from January 1, 2007 to September 30, 2010 were retrospectively analyzed. The primary outcome was incidence of seizures defined as an occurrence of an emergency room visit, ambulance service use or hospitalization with a primary or secondary diagnosis of epilepsy during the 1-year follow-up. Predictor variables included AED adherence, general comorbidity, any mental health comorbidity, evidence of a prior seizure, type of epilepsy diagnosis, presence of AED-interacting medications and any bioequivalent AED switch. The covariates included age, gender and geographic region of residence. The overall incidence of post-index seizures in the 1-year follow-up for four monotherapy cohorts combined was 5.3 % (n=166/3140). The combined cohort analysis demonstrated that pre-index seizures (odds ratio [OR] = 4.28; 95% CI, 2.81-6.53), any mental health comorbidity ([OR] = 3.41; 95% CI, 2.10-5.54), Charlson Comorbidity Index ≥1 ([OR] = 2.88; 95% CI, 1.96-4.24) and monotherapy with levetiracetam ([OR] = 1.54; 95% CI, 1.03-2.31) were significant predictors of seizure recurrence. Among covariates, only geographic region was a significant predictor, with patients residing in the Northeast U.S. having higher odds of post-index seizure ([OR] = 1.92; 95% CI, 1.19-3.10), while controlling for clinical, medication and demographic characteristics. A bioequivalent AED switch, type of epilepsy diagnosis, AED adherence and presence of interacting medications were not significant predictors of seizure recurrence (p>0.05). Results indicate that epilepsy patients with comorbid conditions (both mental and somatic diseases), as well as patients who may have initially been unstable (with previous seizure occurrences) were more likely to experience seizures.