Housing is a fundamental determinant of health, and growing housing instability across North America has prompted jurisdictions to employ measures such as municipal encampment sweeps. These sweeps often involve forcibly displacing unhoused people from public spaces and, in many cases, confiscation of personal belongings. Prior research has documented health and safety harms linked to sweeps, but the impact of these practices on housing outcomes—specifically whether they affect transitions out of homelessness—remained unclear in the study setting.
This study evaluated whether experiences of municipal government–enforced confiscation of belongings or forced displacement were associated with a reduced likelihood of stopping to report homelessness among unhoused people who use drugs in Vancouver, Canada.
The analysis used data from two harmonized prospective cohort studies of people who use drugs in Vancouver (VIDUS and ACCESS). Participants were followed with semi-annual interviews between December 1, 2021 and May 31, 2025. The primary exposure for the main analysis was self-reported municipal confiscation of personal belongings by city workers during the overall study period. A secondary analysis examined municipal government–enforced forced displacement (which may or may not have involved confiscation) during June 1, 2023 to May 31, 2025.
The analytic sample comprised 235 unhoused participants for the primary analysis and 154 participants for the secondary analysis. The outcome was defined as no longer reporting homelessness in the past six months during follow-up (i.e., transitioning from homelessness to not reporting homelessness).
To account for potential time-varying confounding, the investigators used proportional hazards marginal structural modeling (MSM). Models adjusted for a range of demographic, behavioral, health status, and social/structural covariates collected during follow-up. The study reports incidence rates and adjusted hazard ratios (AHRs) with 95% confidence intervals and p-values.
In the primary analytic sample of 235 participants, 59 (25.1%) reported confiscation of belongings at baseline. During follow-up, 128 participants (54.5%) stopped reporting homelessness, which corresponded to an incidence rate of 55.6 cases per 100 person-years (95% CI 46.4–66.1).
In the secondary analysis of 154 participants with data on forced displacement between June 2023 and May 2025, 50 (32.5%) reported forced displacement at baseline and 62 (40.3%) stopped reporting homelessness during follow-up.
Using proportional hazards MSM, experiencing municipal confiscation of belongings was associated with a lower hazard of transitioning out of homelessness (AHR 0.35; 95% CI 0.13–0.94; p = 0.036). Municipal government–enforced forced displacement showed an even stronger negative association with transitioning from homelessness (AHR 0.07; 95% CI 0.02–0.20; p < 0.001).
These results indicate that participants who experienced confiscation or forced displacement were less likely to stop reporting homelessness during the study period compared with those who did not report such experiences.
The study provides prospective cohort evidence that municipal practices associated with encampment sweeps—specifically confiscation of personal belongings and forced displacement—were negatively associated with transitions out of homelessness among unhoused people who use drugs in Vancouver. The findings suggest that such interventions may undermine pathways to housing rather than facilitate them.
Authors emphasize that policy approaches should prioritize access to dignified and affordable housing over practices that displace people and remove personal property. The results are positioned as a call for public health and policy responses that focus on housing-first and dignity-preserving strategies rather than routine displacement.
Key limitations reported by the authors include:
Observational study design: causal inference is limited and residual confounding cannot be excluded.
Potential overestimation of rates of transitioning from homelessness: the cohort may have disproportionately included unhoused people who were more connected with services, which could bias estimates of the incidence of exiting homelessness.
Lack of data on housing quality: the study did not capture details about the quality or stability of housing placements once participants stopped reporting homelessness, limiting interpretation of whether placements were adequate or durable.
Measurement and reporting: exposures and outcomes were self-reported, and some analyses covered different time windows (confiscation across 2021–2025, forced displacement in 2023–2025), which should be considered when interpreting comparisons.
The authors note these limitations while presenting adjusted estimates derived from MSM to address time-varying confounding.
In this cohort of unhoused people who use drugs in Vancouver, municipal confiscation of belongings and forced displacement were both associated with reduced likelihood of transitioning out of homelessness during follow-up. The authors conclude that encampment sweeps may undermine pathways to housing and advocate for policy approaches that center access to dignified, affordable housing rather than displacement and property confiscation.
They recommend future research to examine the quality and stability of housing placements after transitions out of homelessness and to further investigate causal pathways linking encampment sweeps to housing outcomes.
The VIDUS and ACCESS cohorts and the current study were supported by the US National Institutes of Health (U01DA038886, U01DA021525), the Canadian Institutes of Health Research (PJT-196009), and other Canadian research grants noted by the authors. The funders did not influence study design, data collection, analysis, interpretation, or publication.
Data cannot be shared publicly under the institutional research ethics approval; however, anonymized data can be made available to qualified researchers under conditions specified by the Simon Fraser University Research Ethics Board. A minimal dataset for some summary statistics is available in the study's supporting information (S1 Data).