Effect of an Emergency Department Closure on Homeless Patients and Adjacent Hospitals

Abstract

Introduction: Homeless and housed patients differ on several emergency department (ED) metrics (emergency medical services [EMS] use, chief complaints, admission rates, etc.). On January 1, 2018, Memorial Hospital (MH), a safety-net hospital in Pawtucket, Rhode Island, closed. We studied the impact of this closure by analyzing homeless patient utilization of the two closest EDs before and after MH closed. Methods: A retrospective chart review compared the ED records of The Miriam Hospital (TMH), (1.8 miles from MH) and Rhode Island Hospital (RIH), (4.3 miles from MH). We analyzed visits between January 1, 2017–December 30, 2018. (MH closed on 1/1/2018). Patients were identified as homeless if their address listed was either “homeless” or a shelter/ homeless service provider. All other patients were assumed to be housed. We removed from the analysis visits without an address listed or visits missing other key study variables (1.6% of the total). Results: A total of 113,925 unique patients visited the RIH and TMH EDs in 2017, as well as 117,167 in 2018. Homeless patients accounted for 1.18% of patients seen in 2017 and 1.32% in 2018. Between 2017 and 2018, this represents an increase of individual homeless patients of 15.46% (1553-1345), while the number of unique housed patients increased by 2.69% (115,614-112,580). The closer hospital, TMH, saw a 43.72% increase in homeless visits, while RIH saw an 8% increase. Homeless patients were discharged significantly more often than housed patients (74% vs 65%) and had significantly longer time to admission (466.0 vs 304.0 minutes) and discharge (397.9 vs 263.7 minutes) compared to housed patients. Homeless patients presented with suicidality (8.61% of visits) and alcohol-related concerns (29.88% of visits) significantly more than housed patients (1.43% and 2.94%, respectively). Conclusion: When a local ED closes, other EDs are impacted. We found visits made by homeless patients increased more than those made by housed patients and skewed significantly toward the closer hospital. We also found that homeless patients spend significantly more time in the ED and presented with behavioral health complaints more frequently. This impact of hospital closure on patterns of ED utilization by homeless patients has implications for ED management and homeless services both in the ED and the community.