Abstracts

Rationale: The purpose of this study is to evaluate the clinical and economic burden of uncontrolled epilepsy when compared to controlled epilepsy in a Medicaid population. Methods: Medical and pharmacy claims from Florida (1997Q3-2008Q2), Iowa (1998Q1-2006Q2), Kansas (2001Q1-2009Q2), Missouri (1997Q1-2008Q2), and New Jersey (1997Q1-2008Q4) Medicaid databases were analyzed. Patient selection criteria were: ? 18 years old; ? 1 medical visit with ? 1 diagnosis of epilepsy (ICD-9 345.xx) or ? 2 diagnoses of non-febrile convulsions (ICD-9 780.3 or 780.39) occurring ? 30 days apart; ? 1 pharmacy dispensing of any AED; and continuous enrollment throughout observation (minimum 365 days). A retrospective longitudinal matched-cohort design was used to classify patients into mutually-exclusive cohorts of uncontrolled , intermediate , and well-controlled epilepsy. Uncontrolled epilepsy was defined as ? 2 consecutive changes in AED therapy occurring ? 30 days apart, followed by ? 1 epilepsy-related inpatient or ER visit within the next 365 days. Well-controlled epilepsy was defined as no AED change, and no epilepsy-related inpatient or ER visit. Patients not uncontrolled or well-controlled were classified in the intermediate group. Patients of the well-controlled and intermediate groups were matched 1:1, respectively, with those with uncontrolled epilepsy using propensity score matching. Matched cohorts were compared for resource use, and occurrence of negative clinical events. Statistical differences between cohorts were assessed using multivariate regression models, adjusted for demographics, treatment characteristics, and comorbidities. Results: From 110,425 eligible patients with epilepsy, 3,562 patients (mean age=41.4, 38.9% male) with uncontrolled epilepsy were identified. In total, 3,318 well-controlled and 3,560 intermediate patients were matched 1:1 with uncontrolled epilepsy patients. Compared to well-controlled epilepsy, patients with uncontrolled epilepsy had significantly higher all-cause hospitalizations rates (1.4 vs. 0.2 visits/patient-year, adjusted rate ratio [ARR]=6.99, p<.001), longer hospital stays (9.7 vs. 1.5 days/patient-year, ARR=6.76, p<.001), more frequent emergency-room visits (4.1 vs. 1.1 visits/patient-year, ARR=4.05, p<.001), and more neurologist visits (0.7 vs. 0.2 visits/patient-year, ARR=3.08, p<.001). Negative clinical events occurred more frequently in the uncontrolled epilepsy group (ARRs: fractures: 1.97, motor vehicle accident-related injuries: 2.89, head injuries: 1.89, and status epilepticus events: 38.43; p-values<.001) relative to well-controlled patients. Similar findings were observed between uncontrolled versus intermediate groups. Adjusted differences in pharmacy and medical costs will also be presented.