Factors Influencing Hospital Recidivism among Patients Discharged from an Academic Hospitalist Service

Background:

Many patients that hospitalists care for have complex social situations and poor access to medical care. These vulnerable populations often do poorly after discharge and may repeatedly return to the hospital for unresolved medical issues. We hope that a more thorough understanding of factors that lead to hospital readmission will ultimately guide targeted discharge systems to improve patient safety. The goal of this study was to identify factors that place patients at higher risk for hospilal readmission 30 days after discharge from an academic hospitalist service.

Methods:

We conducted a prospective cohort study approved by our institution's IRB. A trained research assistant enrolled 168 patients consecutively from May through July 2009 from the UNC Chapel Hill (UNC‐CH) Hospitalist service. Patients” medical records were reviewed by a physician 4 weeks after their discharge, Univariate and multivariate hazard ratio (HRs) for hospital recidivism (rehospitalization or emergency department visits within 30 days after hospital discharge) were calculated.

Results:

The median age of enrol lees was 49 years (range 20–85) and 55% were female. Fifty‐eight percent were white and 33% African American. Twenty‐three percent had no health insurance, with the remainder having some combination of Medicare, Medicaid, and private insurance, Chest pain was the most common discharge diagnosis (24%) followed by Gl disease (23%). Alcohol or drug detoxification was the primary discharge diagnosis in 11%, Fifty‐five percent of patients admitted for alcohol or drug detoxifcation were readmitted to the hospital or returned to the ED within 30 days after discharge compared with 19% of patients with other diagnoses (P < 0.001). In multivariate analyses, patients with a diagnosis of alcohol or drug detoxification continued to have increased risk of hospital recidivism (HR = 3.94, 1.79–8.68). There were no statistically significant associations between sex, age, race/ethnicity, or insurance status and hospital recidivism.

Conclusions:

Patients admitted to the UNC‐CH hospitalist service for alcohol or drug detoxification had a markedly higher likelihood of hospital recidivism at 30 days compared with those with other discharge diagnoses. Development and implementation of evidence‐based discharge strategies to improve the coordination of care between in patient and outpatient providers could result in better outcomes after discharge. Determining ways of improving such communication with outpatient substance abuse support groups may aid the coordination of care for those at highest risk of hospital recidivism.

Author Disclosure:

M. Gilchrist, UNC Chapel Hill School of Medicine, employment; C. Moore, UNC Chapel Hill School of Medicine, employment; S. Holcombe, UNC Chapel Hill School of Medicine, medical student.