Abstracts

Rationale: We used hospital discharge data to examine utilization of emergency department and inpatient hospitalization by individuals presenting to the ED with seizures. Methods: Northwell Health is a not-for-profit health system with emergency services through 5 tertiary, 1 children’s, and 11 community hospitals, along with 1 freestanding emergency department.  We queried the system to identify all epilepsy-related ED visits in 2017. Visits were identified based on the presence of an epilepsy or seizure-related ICD-10 discharge code. Additional data fields extracted were an anonymized patient ID, visit type (ED only, or ED to inpatient), facility, chief complaint, ED admission diagnosis, and ED discharge diagnosis. Discharge diagnoses were reviewed and categorized into the following categories: seizure, epilepsy NOS, withdrawal seizure, status epilepticus, psychogenic non-epileptic seizures (PNES), syncope, epilepsy not intractable, epilepsy intractable, miscellaneous. Results: 3945 visits by 2998 patients occurred in 2017.  1 patient (0.03%) accounted for 65 visits (1.6%), 9 patients (1.4%) accounted or 398 visits (9.8%), and 584 patients (15.8%) accounted for 1420 visits (36%).  “Seizure” comprised 2513 visits, “Epilepsy NOS” 1096, “Withdrawal seizures” 135, “status epilepticus” 63, “PNES” 55, “syncope” 31, “epilepsy not intractable” 27, “misc” 16 visits, “epilepsy intractable” 9. While there were differences between sites in the proportions of each diagnosis, these did not follow a discernable pattern. Different diagnostic categories were associated with increased or decreased likelihood of inpatient admission. Overall, 36% of ED visits with seizure-related diagnoses resulted in admission, while 64% were discharged from ED. Significant (p<0.05) increased likelihood of admission was associated with diagnoses of status epilepticus, odds ratio (OR) 55 (95% confidence interval [CI] 13-227), withdrawal seizures OR 4.7 (3.2-6.9), or syncope OR 3.2 (1.5-6.7). Diagnosis of epilepsy intractable had increased OR 3.5 (0.9-14.0), but was not significant due to low number of cases. The inter-facility likelihood varied by location, ORs between 0.5 and 2.4, and requires further analysis.  Conclusions: Our study of ED visit data across our multihospital network identified a number of interesting findings. First, as with other studies of ED utilization, a small number of patients account for a large proportion of ED utilization.  We found that 1.4% of patients accounted for nearly 10% of ED visits, and 16% of patients accounted for 36% of ED visits. Looking across admission categories, patients identified with status epilepticus, withdrawal seizures, intractable epilepsy, or syncope were more likely than average to result in inpatient admission, and admission rates varied substantially by hospital. Further analysis of this data with chart review of individual high-frequency patients offer an opportunity to reduce patient seizure burden while possibly also reducing utilization of acute care resources. Funding: None