Long-Term Behavioral Health and Recidivism Outcomes for Individuals Participating in Jail-Based Mental Health Intervention Programs

Background: There is an increase in individuals struggling with serious mental illness (SMI), many of whom are engaged with the criminal legal system (Paccone, 2020). Studies show that around 44% of individuals in jails meet the criteria for a mental health disorder (Bronson & Berzofsky, 2017). Research on incarceration rates and recidivism often fail to highlight the importance of providing mental health jail interventions (MHJI) within these facilities (Stringer, 2019). Failure to identify, track, and treat those suffering with SMI can contribute to relapses in care and possibly recidivism (Stringer, 2019). Interventions, when implemented early-on, could lead to positive outcomes. Thus, this study asks: How do individuals with SMI who participate in MHJI compare in demographic, community-based treatment engagement, and recidivism to individuals with SMI in jail who do not participate?

Methods: This exploratory study assesses pre/post outcomes of 1,348 individuals in 9 jails in a Midwestern state in 2019: 157 (11.6%) of whom screened positive for SMI and participated in a MHJI (intervention group), compared to 1,191 individuals (88.4%) in jail who screened positive for SMI but did not participate (comparison group). Demographic and behavioral health data were collected from a screening instrument distributed during booking, pre/post treatment data was collected through Medicaid encounters, and pre/post jail stays were provided by each county sheriff’s office. Bivariate analyses (chi-square, independent samples and paired samples t-tests) assessed differences between the intervention and comparison groups. A logistic regression model, controlling for demographic differences between the samples, assessed the impact that intervention participation had on recidivism in the post-jail period.

Results: The intervention group differed from the non-intervention group in that they were more likely to: be older (t(1,338)= -2.00, p<.05), housing insecure (χ²(1)=4.00, p<.05), a consumer of the community mental health system (χ²(1)=14.76, p<.001), have a prior jail stay (χ²(1)=20.30, p<.001), and be in jail on a technical violation (χ²(2)=11.67, p<.01). Paired samples t-tests showed that the intervention group had a statistically significant drop in the number of jail stays in the post-period (M=1.02, SD=1.32), compared to the pre-period (M=1.29, SD=1.47; t(131)= -1.92, p=.057); while the non-intervention group saw no significant change in jail stays. After controlling for ways the samples differed, a logistic regression model (χ²(7, n=932)=60.117, p<.001) showed that participants in the intervention group did not significantly differ from the non-intervention group in recidivism.

Conclusions: Individuals with SMI in the intervention sample had greater risk factors coming into jail, yet saw significant improvements in post-jail recidivism. While it is preferred that individuals with SMI be diverted from the criminal/legal system into treatment, it is not always possible given their similar criminogenic risk profiles as individuals without SMI (Epperson, et al., 2014). Thus, the provision of mental health services and connection to care post-release are needed interventions for this population, which may lead to a reduction in cycling in and out of jail facilities. Future studies should continue assessing various jail-based programming across different areas of the U.S. to see if findings can be replicated and become generalizable to this population.