To identify patient characteristics associated with hospital readmissions across three inpatient neurology services at NYULH.

Reducing 30-day hospital readmissions improves patient outcomes and lowers healthcare costs. Identifying high-risk patient characteristics can help direct interventions to patients most at risk. Prior studies mostly focused on stroke or epilepsy patients, and generally used large national databases, which may miss specific patient or service-level characteristics.

All 30-day readmissions from the General Neurology, Stroke, and Epilepsy services at NYULH Brooklyn and Manhattan campuses from 2016-2017 were reviewed and matched to a random sample of non-readmitted patients during this time who were not deceased, discharged to hospice, or discharged to another hospital system. Collected data included: patient demographics, comorbidities, disease specific metrics, index admission and discharge characteristics. Univariate statistical analyses were conducted to compare non-readmitted and readmitted groups within services. To examine specific variables with adjustment for confounders, we used binomial logistic regression.

A total of 599 patients were included, 289 (48%) of which were readmitted (87 general, 60 epilepsy, 142 stroke). Across all services, readmitted patients had more than 1 hospital visit the year prior, a longer LOS, and greater medications at discharge, with hospital visits being statistically significant after adjusting for confounders. General Neurology service readmissions had more seizure as admitting disorder and had an ICU stay. In the stroke service, readmitted patients were older, had more specific comorbidities (hypertension, malignancy, dementia), had dysphagia or a tracheostomy/percutaneous endoscopic gastrostomy tube placement, and had a mRS of 4 or 5. Variables associated with epilepsy readmissions were greater than 8 seizures per month, prior VNS placement, and more total comorbidities.

We identified patient characteristics associated with readmission among all neurology services and within each service at a multi-center academic center. These quantified associations may provide clinical utility in identifying high risk patients for readmission.