Abstracts

Rationale: Patients with epilepsy are at high risk for hospitalization. Identifying patients vulnerable to admission, long hospital stays, and adverse outcomes would be beneficial to clinicians. Our objectives were to investigate the causes and rates of hospital admission, factors associated with discharge status, and prevalence of in-hospital mortality in patients with epilepsy. Methods: We analyzed the 2013 National Inpatient Sample, the largest publicly available US all-payer inpatient health care database which consists of information on patient and hospital level factors for 35 million hospitalizations nationally. Diagnosis of epilepsy was based on the presence of validated ICD-9-CM codes in any diagnostic position. 1:2 propensity score matching was performed and epilepsy cases (n = 207,767) were matched by age, sex, race and number of chronic conditions to those without epilepsy (n = 415,534). Primary outcome of interest was the discharge disposition of the patients (used as a proxy for morbidity) categorized as: 1) Home (Routine/home health care) vs. 2) Other facilities (transfers, against medical advice, died, destination unknown). Other outcomes of interest were predictors of epilepsy admission and top ten conditions for hospitalization. Chi-square tests and multivariable logistic regression were conducted using sampling weights to examine the patient and hospital level factors associated with discharge disposition. Results: Out of 207,767 epilepsy admissions, 52% were female, mean age was 52.2 years (range 0-90), with 5,627 (2.7%) in-hospital deaths. Mean length-of-stay was 5.8 days (range 0-365) and mean number of comorbidities was 7 (range 1-25). The top ten causes of admission by frequency were: epilepsy/convulsions (18%), septicemia (6.2%), mood disorders (3.3%), pneumonia (3.1%), acute cerebrovascular disease (2.3%), complication of device/implant (2.2%), schizophrenia and other psychotic disorders (2.2%), urinary tract infections (2.2%), respiratory failure (2.1%), aspiration pneumonitis (1.7%). Compared to controls, epilepsy admissions had lower odds of discharge home (OR 0.64, 95% Confidence Interval (CI): 0.63-0.65, p<0.0001). Patients with non-Medicare insurance had higher odds of discharge home (Medicaid vs. Medicare, OR 2.72, CI: 2.63-2.82, p<0.0001; Other insurance vs. Medicare, OR 3.15, CI: 3.07-3.22, p<0.0001). Factors associated with discharge to a facility included small hospital size (OR 0.86, CI: 0.83-0.89, p<0.0001), longer length-of-stay (OR 0.95, CI: 0.95-0.95, p<0.0001), rural hospitals (OR 0.74, CI: 0.71-0.77, p<0.0001), and urban non-teaching hospitals (OR 0.8, CI: 0.78-0.83, p<0.0001). Conclusions: Hospitalized epilepsy patients are at greater odds of not being discharged home vs. non-epilepsy patients. Treatment at an urban teaching hospital, an elective admission and shorter in-hospital length-of-stay show significantly greater odds of being discharged home. Increased awareness of reasons for admission in epilepsy such as psychiatric, infectious and respiratory causes may help prevent some hospital admissions. Funding: None