Patient-level predictors of emergency department utilization and recidivism among individuals with a history of incarceration

Background: Patients recently released from correctional facilities are high utilizers of emergency department (ED) services and frequently re-incarcerated. Little is known about which patient factors are associated with ED utilization or recidivism. We sought to identify, through readily available health systems information and risk assessments, patient determinants amenable to care management interventions aimed to reduce ED utilization and recidivism. Methods: We used data collected from a randomized trial of the Transitions Clinic in San Francisco, CA. Participants included 200 individuals recently released from prison who had at least one chronic illness or were 50 years or older and did not have a primary care provider at the time of release. Potential predictors, measured at study baseline, included: socioeconomic status (including housing status), incarceration history, health insurance, receipt of health care in prison, substance use, chronic diseases, and self-rated health status. The outcomes were ED use and repeated incarceration (each defined as ≥2 occurrences compared to <2) in the year following release and were obtained from administrative data. We used χ2 and student t-test to assess bivariate relationships between potential predictors and each of the outcomes, and ran multivariable logistic regression models to identify predictors for each outcome, including predictors with bivariate p<.2. Results: The mean age of participants was 43.2±9.0 years old, 64% were Black, and 41% had a chronic disease diagnosed during their last imprisonment. Mean length of imprisonment was 36.4±61.9 months. In the year after baseline interview, 29 (15%) of patients had ≥2 visits to the ED and 46 (23%) had ≥2 incarceration episodes. In multivariable models adjusted for total time incarcerated, past suicide attempt, having diabetes, and clinic assignment (transitions versus community), poor/fair self-rated health was associated with multiple ED visits (Adjusted Odds Ratio [AOR]=3.0[1.3-6.9]). Poor/fair self-rated health (AOR=3.9[1.7-9.2]) was also strongly associated returning to jail, as were having less than a high school education (AOR=2.7 [1.1-6.7]), seizure disorder (AOR=4.4, [1.2-15.5]) or incarceration before 18 years [3.9 (1.8-8.3)] after adjustment for health insurance, total time incarcerated, past mental health hospitalization, past suicide attempt, treatment in the ED during incarceration, having a case manager, and having chronic pain. Conclusion: In a cohort of returning prisoners with chronic medical conditions, poor self reported health is a predictor of ED utilization and recidivism. Providers should assess all returning prisoners for self-rated health, and use this information to identify a high-risk subset of individuals for more intensive care management.