Abstracts

Rationale: Understanding patterns of inpatient utilization is important for optimizing longitudinal care for patients with epilepsy. In particular, patients readmitted to the hospital after a seizure-related admission may represent a particularly high-risk group in terms of utilization and morbidity. Reducing readmissions is crucial because they are costly and burdensome, and because they may reflect gaps in transitional care out of the hospital. However, little is known regarding reasons and predictors of readmissions among patients with epilepsy. We sought to determine the prevalence of and risk factors for hospital readmissions and to identify the most common diagnoses in patients hospitalized primarily for seizures. Methods: We performed a retrospective cohort study of all adult patients hospitalized with a primary discharge diagnosis of seizure or convulsions (identified by ICD 9 CM codes) contained within the State Inpatient Database in years 2009 to 2012 across 11 states. Information regarding hospital and community characteristics were obtained from the American Hospital Association and Robert Wood Johnson Foundation, respectively. We performed serial logistic regressions to explore the effect of a broad array of patient, hospital, and community level risk factors on all-cause 30-day readmissions. Results: Of 98,712 included patients, 13,929 (14.1%) had a 30-day readmission. The most common primary reasons for readmission included epilepsy/convulsions (29.9% of readmissions), mood disorders (5.3%), schizophrenia (4.4%), and septicemia (3.9%). The strongest patient-level predictors of readmission were CNS tumor (2.1, 1.9-2.4), urgent index admission (2.0, 1.8-2.2), psychosis (OR, 95% CI: 1.7, 1.6-1.8), and long length of stay (11 + days versus 0-1 days: 1.7, 1.5-1.9). Aggregate county-level variables were generally not significantly associated with readmission aside from uninsurance (1.2, 1.1-1.3 for every 10% increase in uninsured proportion). Full results are listed in Table 1. The model c-statistic for the adjusted logistic regression for all-cause readmissions was 0.63. Index admission for status epilepticus (3.5, 2.6-4.8) and low hospital epilepsy volume (0.4, 0.3-0.8 comparing yearly hospital epilepsy volume 200 + patients per year versus < 40 patients per year) were predictors of readmission for status epilepticus, but had little effect on all-cause readmission (respectively: 1.0, 0.9-1.0 and 1.0, 0.9-1.1). Conclusions: Readmissions are common after an acute hospitalization related to epilepsy. Recurrent seizures and mental health conditions were the most common reasons for readmission. It was noteworthy that care at hospitals with higher volumes of epilepsy admissions tended to have lower readmissions for status epilepticus which may have important implications for how care is delivered for epilepsy emergencies. While many patient-level factors meaningfully predict readmission, readmissions are challenging to predict even with extensive administrative data. Clarifying drivers for post-seizure readmissions would benefit our ability to counsel patients and better design systems of care. Funding: Dr Burke is funded by National Institute of Neurological Disorders and Stroke K08 NS082597 and National Institutes of Health National Institute on Minority Health and Health Disparities R01 MD008879