Association for Public Policy Analysis & Management

Despite what has been conceptualized as a revolving door between hospitals and homelessness, surprisingly little research has used data from both systems to examine the full healthcare use of this vulnerable population. They are more likely to use high cost acute hospital-based health care, including emergency department visits and inpatient hospitalizations. But without more comprehensive, system-level research, policymakers are ill-equipped to develop interventions that could result in better coordination of services between shelter and hospital systems to improve both health care and housing outcomes for this population.

To address this gap, we examine hospital use immediately preceding and following spells of homelessness among a cohort of approximately 120,000 adults in New York City who entered shelter for the first time between 2008 and 2015. We match shelter records from the City’s Department of Homeless Services with an all-payer claims database of hospital visits in New York State. We assess rates and correlates of emergency department visits and inpatient hospitalizations in the year prior to shelter entry and the year following shelter exit, with particular attention to differences by age, gender, and whether the adults were in a household with children. In addition, we examine the most common hospital presenting diagnoses to assess patient needs.

Results suggest high rates of hospital use both in the year prior to shelter entry and the year following shelter exit, with nearly 40% of shelter entrants and exiters using the hospital before or after homelessness. Of those using the hospital, nearly all had an emergency department visit and approximately one-third had an inpatient stay. Rates of adult hospital use were consistent regardless of the presence of a child in the household, but women were more likely than men to use the hospital and risk of inpatient hospitalization increased with age. Notably, risk of hospital use was highest in the time periods immediately before and after shelter use, suggesting a confluence of simultaneous healthcare and housing crises.

These findings demonstrate the interdependence of disparate systems and the need for collaboration across government entities and between government agencies and private providers. The spike in hospital use immediately prior to shelter entry is especially noteworthy, as hospital visits may present potential opportunities to identify risk for future homelessness and prevent entry into shelter. In addition, the study demonstrates the importance of integration of data across systems so researchers and policymakers can assess rates of and risk factors for multi-system use and develop programmatic tools as dictated by observed trends.