Identifying influential factors distinguishing recidivists among offender patients with a diagnosis of schizophrenia via machine learning algorithms

• Published in: Forensic science international Vol. 315; p. 110435
• Main Authors: Kirchebner, Johannes, Günther, Moritz Philipp, Lau, Steffen
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier B.V 01-10-2020; Elsevier Limited
• Subjects: Adult; Aggression; Algorithms; Bayesian analysis; Crime; Criminal justice; Criminal recidivism; Criminals; Datasets; Diagnosis; Drug abuse; Female; Forensic psychiatry; Forensic science; Forensic sciences; Hospitalization; Humans; Hypothesis testing; Learning algorithms; Legal issues; Machine Learning; Male; Medical diagnosis; Mental disorders; Methods; Offender; Opioids; Psychotropic drugs; Recidivism; Researchers; Retrospective Studies; Risk factors; Schizophrenia; Sex crimes; Subgroups; Substance abuse treatment; Support vector machines; Variables; Switzerland; Criminal recidivism; Offender; Schizophrenia; Criminal justice; Machine learning
• ISSN: 0379-0738; 1872-6283
• DOI: 10.1016/j.forsciint.2020.110435

•A new approach to machine learning provided new insight on criminal recidivism.•653 variables were explored in 344 offender patients with schizophrenia.•The final algorithm offers an accuracy of 81.7 % and an AUC of 0.89.•Legal, criminological and clinical factors were identified in a European setting.
There is a lack of research on predictors of criminal recidivism of offender patients diagnosed with schizophrenia.
653 potential predictor variables were anlyzed in a set of 344 offender patients with a diagnosis of schizophrenia (209 reconvicted) using machine learning algorithms. As a novel methodological approach, null hypothesis significance testing (NHST), backward selection, logistic regression, trees, support vector machines (SVM), and naive bayes were used for preselecting variables. Subsequently the variables identified as most influential were used for machine learning algorithm building and evaluation.
The two final models (with/without imputation) predicted criminal recidivism with an accuracy of 81.7 % and 70.6 % and a predictive power (area under the curve, AUC) of 0.89 and 0.76 based on the following predictors: prescription of amisulpride prior to reoffending, suspended sentencing to imprisonment, legal complaints filed by relatives/therapists/public authorities, recent legal issues, number of offences leading to forensic treatment, anxiety upon discharge, being single, violence toward care team and constant breaking of rules during treatment, illegal opioid use, middle east as place of birth, and time span since the last psychiatric inpatient treatment.
Results provide new insight on possible factors influencing persistent offending in a specific subgroup of patients with a schizophrenic spectrum disorder.