National characteristics and predictors of neurologic 30-day readmissions.

Abstract

To examine national 30-day readmission rates for patients hospitalized with neurologic disorders. Using the University HealthSystem Consortium (UHC) database, we identified 554,399 index neurologic admissions from October 2011 through January 2015. We collected information regarding age, race, insurance payer, Medicare severity diagnosis-related group, and severity of illness. We examined readmission by diagnosis and performed multivariable logistic regression to determine predictors of readmission. The unplanned readmission rate was 11.0%. Rates of unplanned readmission were highest for patients with peripheral nerve disorders (21.9%), CNS neoplasms (21.0%), nonhypertensive encephalopathy (15.5%), arterial stenosis (15.4%), and bacterial CNS infections (14.5%). In multivariable analysis, higher severity of illness and public health insurance coverage predicted higher rates of 30-day readmission. We found significant variation in readmission rates for different neurologic disorders. These data provide insight into management of neurologic disease nationally, offering policymakers realistic goals for standards of care and challenging health care providers to develop systems-based solutions that will improve transitions of care for those at highest risk of readmission with neurologic disease.