Ethno-Racial Differences in the Effectiveness of Housing First on Healthcare Utilization Among Homeless Adults with Mental Illness: At Home Chez-Soi, a Pragmatic Randomized Trial

This study evaluates the impact of the Housing First (HF) intervention on healthcare utilization among homeless adults with mental illness, comparing ethno-racial (ER) and non-ER groups with moderate mental health support needs. We linked data from the At Home/Chez-Soi (AH/CS) randomized controlled trial with health administrative data from Ontario, including the Ontario Health Insurance Plan and several healthcare databases (e.g., Discharge Abstract Database, Ontario Mental Health Reporting System). We analyzed outcomes such as primary care visits, hospitalizations, emergency department (ED) visits, and Ambulance arrival Incidents. Applying the intention-to-treat principle, we used generalized estimating equation models using the Poisson distribution which included the intervention (HF vs. TAU), the time period (0-7 years post-randomization vs 1-year pre-randomization) and their interaction. Incidence rate ratios (IRRs) were estimated to assess the differences in outcome rate changes between HF vs TAU comparing the 0 to 7 year follow-up period with the pre-randomization year for each stratified ethno-racial group. The results revealed a mitigated HF effect among non-ER participants, with increased utilization rates for both primary care and acute care services. [Primary care visits [IRR 2.17 95% (1.03-4.57)], acute care visits (Mental health hospitalization [IRR 2.68 95% CI (1.17-6.20)], Ambulance arrivals to ED [IRR 1.94 95% (1.05-3.56)] and number of Ambulance arrivals to ED [IRR 3.67 95% (1.05-13.02)]). However, no statistically significant effect was found among the ER participants in any of the healthcare services examined. The differential impact of HF on healthcare utilization suggests persistent systemic barriers to care, such as stigma and discrimination among ER participants.Trial Registration: The study is registered with the ISRCTN (#42520374).