Stopping the Revolving Door: Understanding the Connection between Mental Illness and Recidivism

Background: The United States criminal legal system has in many ways become the default mental health system. Prevalence rates of mental illness (MI) in U.S. prisons are estimated to be between 3-12 times greater than the general population. Despite the numerous indicators that MI is a common issue among those incarcerated, it remains unclear what the nature of the relationship between having a MI and recidivism (defined in this study as rearrest within three years) is. Deepening our understanding of these relationships is vital to developing strategies that are responsive to individuals’ needs, effectively addressing public safety and health concerns, and supporting successful reentry to society post-incarceration. The primary aims of the study are to test the extent MI is related to increased recidivism and examine the degree factors commonly associated with recidivism (i.e., criminogenic risk factors) have differential impacts when comparing individuals who have been identified as having a MI to those without a MI.

Methods: The dataset contains over 25,000 individuals released on parole in the state of Georgia between January 1, 2013, and December 31, 2015. Recidivism is tracked over a three-year period. About 15% of the sample was assigned mandated mental health treatment as a condition of their parole. Entropy balancing is utilized to facilitate comparison between those identified as having a MI in need of mandated treatment compared to those without this parole conditions. Logistic regression was utilized to examine if individuals with mandated mental health treatment had a greater likelihood of being rearrested within three years of release. Additional logistic regressions were run to test for interactions between having a MI and common criminogenic risk factors.

Results: For all individuals on parole, higher risk scores and increased percentage of positive drug tests were associated with increased likelihood of being rearrested. Individuals with MI had an increased risk of recidivism (OR=1.18, 95% CI [1.08, 1.28]) compared to those without mandated mental health treatment. Employment served as a protective factor for both groups, however, it was a greater protective factor for individuals without MI. When comparing the interaction of positive drug test and MI, persons with MI who did not test positive for substances had a higher likelihood of being arrested compared to those without MI who also had no positive tests. As the percentage of positive drug tests increased the two groups no longer become significantly different in their likelihood of rearrest.

Conclusion: Th findings highlight a need to avoid a one size fits all approach for individuals under community supervision, particularly for those with MI. Community supervision agencies can better support this population by reviewing how their practices and policies can be more responsive to this populations specific needs. Understanding employment may serve as a protective factor by varying degrees depending on if a person has a MI is an important consideration for professionals engaging in treatment planning and case management. Additionally, these findings support continued efforts to incorporate both a mental health and criminogenic rehabilitative focus.