Respiratory complications and 30-day unplanned hospital readmissions in patients with epilepsy.

ABSTRACT

BACKGROUND: To help mitigate the burden of health care on US economy, public policymakers and health care legislation have been focusing on reducing hospital readmissions. Respiratory complications have been identified among the commonest of adverse events in neurologic patients. The goal of our study was to better understand respiratory complications and their contribution to rehospitalizations in patients with seizures. METHODS: We used the 2013 Nationwide Readmission Database to analyze unplanned 30-day readmission rate (30RR). The study population comprised of patients with index hospital discharge diagnosis of generalized convulsive epilepsy and status epilepticus. Patients under 18 years of age, who died during hospitalization or who had missing demographic data, were excluded. Patients hospitalized in December were also excluded due to lack of 30-day follow-up. The primary outcome of interest was 30-day readmission. The causes of readmission were determined by corresponding International Classification of Diseases, Ninth Revision, Clinical Modification codes. RESULTS: The 30RR was highest in patients with index hospitalization discharge diagnosis of status epilepticus, followed by generalized convulsive epilepsy (intractable), followed by generalized convulsive epilepsy (nonintractable). While seizure was the most common reason for readmission, contribution of respiratory complications to readmissions was 7.85%, 12.39%, and 6.93%, respectively. Pneumonia/aspiration pneumonitis and respiratory insufficiency accounted for the majority of the readmissions in all subgroups. CONCLUSIONS: Respiratory complications are the leading nonseizure cause of 30-day unplanned readmissions in patients with generalized convulsive epilepsy and status epilepticus. Further research on identifying appropriate interventions to reduce readmissions from respiratory causes may improve outcomes for patients in these epilepsy subgroups.