Background Homelessness is associated with poor health. A policy approach aiming to end homelessness across Europe and North America, the ‘Housing First’ (HF) model, provides rapid housing, not conditional on abstinence from substance use. We aimed to systematically review the evidence from randomised controlled trials for the effects of HF on health and well-being.

Methods We searched seven databases for randomised controlled trials of interventions providing rapid access to non-abstinence-contingent, permanent housing. We extracted data on the following outcomes: mental health; self-reported health and quality of life; substance use; non-routine use of healthcare services; housing stability. We assessed risk of bias and calculated standardised effect sizes.

Results We included four studies, all with ‘high’ risk of bias. The impact of HF on most short-term health outcomes was imprecisely estimated, with varying effect directions. No clear difference in substance use was seen. Intervention groups experienced fewer emergency department visits (incidence rate ratio (IRR)=0.63; 95% CI 0.48 to 0.82), fewer hospitalisations (IRR=0.76; 95% CI 0.70 to 0.83) and less time spent hospitalised (standardised mean difference (SMD)=−0.14; 95% CI −0.41 to 0.14) than control groups. In all studies intervention participants spent more days housed (SMD=1.24; 95% CI 0.86 to 1.62) and were more likely to be housed at 18–24 months (risk ratio=2.46; 95% CI 1.58 to 3.84).

Conclusion HF approaches successfully improve housing stability and may improve some aspects of health. Implementation of HF would likely reduce homelessness and non-routine health service use without an increase in problematic substance use. Impacts on long-term health outcomes require further investigation.

Trial registration number CRD42017064457