Hospital Discharge to Long-Term Care Facilities Among Persons Experiencing Homelessness

Background and Purpose: People experiencing homelessness rely heavily on emergency departments (ED) for care and often have complex medical and behavioral health needs, yet there is limited understanding of how housing status is associated with discharge to long-term care (LTC) settings and under what conditions. This study examines the relationship between housing instability and discharge to LTC among adults treated in U.S. hospital-owned EDs.

Methods: We used cross-sectional data from the 2015-2022 Nationwide Emergency Department Sample (NEDS), a 20% stratified sample of U.S. hospital-owned EDs weighted to produce national estimates. The analytic sample included 140,943,235 observations (weighted N=601,623,661 ED visits). Housing status was documented with the use of ICD-10-CM codes to characterize whether patients were unhoused generally, living in shelters, on the streets, or were otherwise experiencing housing instability. Multivariate logistic regression models, accounting for the complex survey design, were estimated to assess the association between housing status and LTC discharge. Models adjusted for demographic characteristics, insurance status, mental health diagnoses, substance use disorders, and cognitive impairment. Additional models examined whether inpatient length of stay (LOS) modified these relationships.

Results: Housing status was strongly associated with discharge outcomes but not always in expected ways. Overall, 2.18% of ED visits resulted in LTC discharge, with higher rates among inpatient admissions (8.2%) than ED discharges (1.8%). In adjusted models, sheltered homelessness was associated with 27-42% lower odds of LTC discharge across samples (e.g., aOR=0.59, 95% CI:0.50-0.70 in the full sample), and unsheltered homelessness was also associated with reduced odds in inpatient models (aOR=0.74, 95% CI: 0.59-0.93). Inadequate housing was associated with higher odds of LTC discharge (full sample aOR=1.48, 95% CI: 1.23-1.78; ED discharge aOR=2.23, 95% CI: 1.76-2.96). Among inpatients, each additional hospital day was associated with a 9% increase in the odds of LTC discharge (aOR=1.09, 95% CI:1.09-1.09), though this relationship varied by housing status. Predicted probabilities showed that longer stays were associated with increases in LTC placement among stably housed patients, but smaller increases among those experiencing homelessness.

Conclusions and Implications: The findings demonstrate a potential mismatch between clinical need and access to post-acute care for people experiencing homelessness. While this population often has significant medical and behavioral health needs, they appear less likely to transition into institutional care settings and more likely to experience prolonged hospital stays or discharge to unstable environments. At the same time, individuals with marginal or inadequate housing may be better positioned to meet administrative or structural requirements for LTC placement. These patterns have implications for how health and social service systems respond to homelessness across institutional settings. Expanding alternatives such as medical respite care, strengthening discharge planning, and improving coordination between hospitals, behavioral health providers, and housing systems may help reduce avoidable hospital use and improve continuity of care for this population.