Abstract Background Efforts to address high emergency department (ED) use among persons experiencing homelessness may be strengthened by identifying risk factors and clarifying whether the primary care experience influences subsequent ED utilization. Although homeless-tailored primary care clinics receive more favorable patient ratings, it remains unclear whether patient experience or clinic type predicts future ED use. Objective To identify clinical and social predictors of high ED utilization among homeless-experienced veterans (HEVs) and to test whether (a) unfavorable care ratings predict higher ED use and (b) care in homeless-tailored primary care clinics predicts lower ED use compared with mainstream clinics. Design A retrospective cohort study linking national survey data with electronic health record data was conducted to assess ED utilization during the 1-year following survey completion. Participants Five thousand seventy-nine HEVs engaged in Veterans Health Administration primary care who completed the 2018 Primary Care Quality–Homeless Services Tailoring (PCQ-HoST) survey. Main Measures Primary care type (homeless-tailored or mainstream), patient-reported primary care experience on a validated survey, and ED utilization (defined as high based on four or more visits). Key Results Among 5079 HEVs, 474 (9.3%) had high ED utilization. High utilization was associated with current and chronic homelessness, chronic pain, medical and psychiatric comorbidities, and substance use disorders. An unfavorable primary care experience was associated with higher odds of subsequent high ED utilization (OR 1.31; 95% CI 1.26–1.68). Primary care clinic type was not independently associated with high ED utilization. In post hoc count-based models, care in a homeless-tailored clinic was associated with 0.15 fewer ED visits annually. Conclusions Among HEVs engaged in primary care, high ED utilization is driven primarily by clinical complexity and housing instability. However, an unfavorable patient-reported primary care experience was independently associated with subsequent ED use, suggesting that improving primary care interactions may contribute to broader efforts to address high ED utilization in this population.
Shufflebarger et al. (Thu,) studied this question.