Impact of criminal justice involvement profiles on the effectiveness and cost-effectiveness of Housing First among homeless people with mental illness

People who experience homelessness and have mental illness are at increased risk of involvement with the criminal justice system that comes at significant personal, social and economic cost. While there is strong evidence that Housing First increases residential stability, the literature suggests that Housing First does not, on average, have any impact on justice outcomes. This thesis seeks to explore if this absence of impact is due to a heterogenous intervention effect, both in terms of profiles of criminal justice involvement and in terms of the nature of the offense, using data from the multisite randomized controlled trial At Home/Chez soi. In a prior paper, we empirically identified five profiles: Non/Infrequent Offenders (very few offences, but some single violent offences), Moderate Offenders (higher mean number of offences, longest duration of homelessness), Antisocial Offenders (highest means of violent and nuisance offences), Versatile Subsistence Offenders (highest mean of all offences and highest means of subsistence-related offences – e.g., theft), and Drug Offenders (highest mean number of drug-related offences). The first manuscript explores whether the absence of impact of Housing First on criminal charges overall was due to heterogenous intervention effects. The Risk-Needs-Responsivity model of offender rehabilitation suggests that individuals with certain risk factors of criminal justice involvement would be more likely to experience a reduction in criminal justice involvement as a result of Housing First, while others could experience an increase. For example, people for whom offending is in part driven by poverty and homelessness (Moderate Offenders) or by substance abuse (Drug Offenders) before coming into the program would be more likely to experience a reduction in offending as a result of increased residential stability and a harm reduction approach to substance abuse than Antisocial Offenders and Versatile Subsistence Offenders, whose criminogenic needs might not be addressed as part of the intervention. For these profiles, the intervention could even have a deleterious effect. Results showed that the impact of Housing First on violent charges differs by profiles. Moderate and Drug Offenders were more likely to experience fewer charges for violent crimes from Housing First (-0.40 [95% CI: -0.70, -0.10] and -0.53 [95% CI: -1.01, -0.05] new charges over two years, respectively). There was no impact of Housing First on subsistence charges and charges related to the administration of justice for any profile.The second manuscript assessed, first, the association between CJI profiles and days stably housed as well as costs. Second, it tested whether the effect of Housing First and its net benefit differed according to profile. On average, controlling for group assignment and other variables, Antisocial Offenders spent 48 fewer nights annually (95% CI: -72.1, -23.0) in stable housing than Non/Infrequent Offenders. Versatile Subsistence and Moderate Offenders engendered greater costs (1.33 [95% CI: 1.11, 1.59] and 1.21 [95% CI: 1.07, 1.36] times, respectively). However, the effectiveness and cost-effectiveness of the intervention did not differ by profile. Findings suggest that Housing First, when provided to homeless individuals with mental illness who have an ongoing risk of offending, may need to target the criminogenic needs of the participants in order to observe an overall reduction in criminal justice involvement. Future research could, for example, develop and test voluntary adjunctive interventions for specific groups of offenders who want to reduce their criminal justice behaviour. However, neither the effectiveness, in terms of residential stability, nor the cost-effectiveness of Housing First compared to usual services varied by profile. Our findings suggest that there is no ground for selective admission into Housing First programs based on criminal history.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.