Clarifying the relationship between mental illness and recidivism using machine learning: A retrospective study.

OBJECTIVE: There is currently inconclusive evidence regarding the relationship between recidivism and mental illness. This retrospective study aimed to use rigorous machine learning methods to understand the unique predictive utility of mental illness for recidivism in a general population (i.e.; not only those with mental illness) prison sample in the United States. METHOD: Participants were adult men (n = 322) and women (n = 72) who were recruited from three prisons in the Midwest region of the United States. Three model comparisons using Bayesian correlated t-tests were conducted to understand the incremental predictive utility of mental illness, substance use, and crime and demographic variables for recidivism prediction. Three classification statistical algorithms were considered while evaluating model configurations for the t-tests: elastic net logistic regression (GLMnet), k-nearest neighbors (KNN), and random forests (RF). RESULTS: Rates of substance use disorders were particularly high in our sample (86.29%). Mental illness variables and substance use variables did not add predictive utility for recidivism prediction over and above crime and demographic variables. Exploratory analyses comparing the crime and demographic, substance use, and mental illness feature sets to null models found that only the crime and demographics model had an increased likelihood of improving recidivism prediction accuracy. CONCLUSIONS: Despite not finding a direct relationship between mental illness and recidivism, treatment of mental illness in incarcerated populations is still essential due to the high rates of mental illnesses, the legal imperative, the possibility of decreasing institutional disciplinary burden, the opportunity to increase the effectiveness of rehabilitation programs in prison, and the potential to improve meaningful outcomes beyond recidivism following release.

Mental Health Units in Correctional Facilities in the United States.

BACKGROUND: The prevalence of severe mental illness (SMI) in correctional settings is alarmingly high. Some correctional facilities have developed mental health units (MHUs) to treat incarcerated individuals with SMI. OBJECTIVE: To identify existing MHUs in the United States and collate information on these units. DATA SOURCES: A systematic review using Criminal Justice Abstracts, ERIC, PsycINFO, PubMed, and SocINDEX, plus an exploratory review using the Google search engine were conducted. MHUs were included if they were located within an adult correctional facility in the United States, specifically catered to SMI populations, and were in active operation as of June 2019. RESULTS: Eleven articles were identified through the peer-reviewed literature, but there were still major gaps in the information on MHUs. The Google search identified 317 MHUs. The majority of units were located within prisons (79.5%) and served only men (76%). The Google search found information indicating that 169 (53.3%) offered groups or programming to inmates; 104 (32.8%) offered individual therapy; and 89 (23%) offered both. One hundred sixty-six units (52.4%) had dedicated mental health staff, and 75 (23.7%) provided mental health training to correctional officers. Information on funding and outcomes of the MHUs is presented. LIMITATIONS: Use of the Google search engine and sources that have not been peer reviewed limits the robustness of conclusions about the MHUs. CONCLUSIONS: Standards for developing and implementing MHUs are not widespread. The shortcomings of current MHUs are discussed in the context of desired criteria for size, staffing, and programming.

Cognitive and clinical predictors of community functioning across the psychoses.

Schizophrenia, bipolar disorder, and related psychotic illnesses are common, serious mental disorders that are often associated with functional impairments and poor quality of life, even after clinical recovery. Cognitive dysfunction is a strong predictor of functional impairment; however, findings regarding relative impairments in functioning and cognition across diagnoses have been mixed, as have reports of the contribution of clinical symptoms and other illness features to functioning across diagnostic boundaries. We assessed 211 patients with psychotic disorders and 87 healthy controls using the MATRICS Consensus Cognitive Battery, clinical measures of state mood and psychotic symptoms, and an interview measure of community functioning. Diagnostic groups were compared on MATRICS composite and domain scores, and clinical and functional measures. We then examined cognitive, clinical, and demographic predictors of community functioning using stepwise hierarchical linear regression. All three patient groups exhibited deficits in most cognitive domains relative to controls, and significantly poorer community functioning. While scores on most cognitive domains did not differ by diagnosis, when groups did differ patients with schizophrenia performed worse than patients with bipolar disorder. Cognition was correlated with functioning across the sample. The final regression model included negative symptoms, mania, social cognition, and processing speed, and explained 47% of the variance in community functioning scores across patient groups. Residual negative symptoms, residual mania, and social cognition significantly and independently predicted community functioning. These findings indicate that, while all cognitive domains are associated with community outcomes, when considered together and with clinical symptoms, negative symptoms, mania, and social cognition are the strongest predictors across diagnoses. Development of interventions targeting negative symptoms and social cognition may be effective in improving community functioning for patients across diagnoses.