Who Succeeds in Mental Health Court? Identifying Predictors of Program Retention and Legal Recidivism
Methods: This exploratory study used structured interviews to collect demographic, socioeconomic, criminal history, psychiatric, substance use/misuse, health, motivation, and therapeutic alliance characteristic data from 148 MHC participants. Six month follow-up retention and recidivism data was assessed. Participants were recruited by clinicians in four West Coast MHCs who referred clients meeting selection criteria to the researcher. Participants were 148 available adults who: a) had recently enrolled in MHCs; 2) were diagnosed with Axis I schizo-spectrum disorder, bipolar disorder, or major depression; 3) were legally competent; and 4) provided informed consent. Interviews utilized standardized measures including the AUDIT (Saunders et al., 1993), DAST-20 (Skinner, 1982), SF-36 (Ware & Shelbourne, 1992), MARS (Thompson et al., 2000), BPRS (Overall & Gorham, 1962), MCSI (Conrad et al., 2001), URICA-SA (McConnaughy et al., 1983), and Measure of Judicial Alliance (MoJA) based on Skeem's (2004) Dual Role Relationship Inventory, and self-report questions to collect characteristic data. Diagnoses, GAF (APA, 2000), criminal history, and six-month post optin program retention and recidivism data were provided by the participating MHCs. Chi-square and independent samples t-test analyses were calculated to compare differences between retained/not retained and arrested/not arrested participants at six months follow-up. Characteristics that significantly differed between groups were retained as factors in logistic regression models.
Results: Participants ranged in age from 18 to 64 (M = 36.56, SD = 11.81) and most had less than a high school education (M = 11.46, SD = 2.49). They were more typically male (61%), white (58%), unmarried (93%), unemployed (92%), and diagnosed with schizophrenia (49%). Many had a substance use disorder (68%) and prior felony convictions (57%). At 6-month follow-up, 72% of the participants remained enrolled in MHC and 45% had recidivated. Results showed significant differences in education, GAF, contact treatment providers, and therapeutic alliance with the MHC judge between retained/not retained participants. A significant logistic regression model identified that education, mental health contact, and a stronger therapeutic alliance with the MHC judge were significant predictors of participants’ program retention (X2(df = 10) = 92.63, p < .001). Additionally, significant differences were found between arrested/not arrested participants’ age, ethnicity, education, income, housing, criminal history, GAF, BPRS, AUDIT, DAST, and substance use. This model identified that being younger, less educated, non-white, and reliant on SSI/SSDI were significantly predictive of recidivism (X2(df = 11) = 92.63, p <.001).
Implications: MHC programs may use these findings to provide better client assessments, facilitate more individualized and targeted treatments, and strengthen their working alliance with participants. Future research should attempt to understand how various MHC program conditions contribute to better participant outcomes in broader and more diverse samples.