Predicting general criminal recidivism in mentally disordered offenders using a random forest approach

Abstract

BackgroundPsychiatric expert opinions are supposed to assess the accused individual’s risk of reoffending based on a valid scientific foundation. In contrast to specific recidivism, general recidivism has only been poorly considered in Continental Europe; we therefore aimed to develop a valid instrument for assessing the risk of general criminal recidivism of mentally ill offenders.MethodData of 259 mentally ill offenders with a median time at risk of 107 months were analyzed and combined with the individuals’ criminal records. We derived risk factors for general criminal recidivism and classified re-offences by using a random forest approach.ResultsIn our sample of mentally ill offenders, 51% were reconvicted. The most important predictive factors for general criminal recidivism were: number of prior convictions, age, type of index offence, diversity of criminal history, and substance abuse. With our statistical approach we were able to correctly identify 58-95% of all reoffenders and 65-97% of all committed offences (AUC = .90).ConclusionsOur study presents a new statistical approach to forensic-psychiatric risk-assessment, allowing experts to evaluate general risk of reoffending in mentally disordered individuals, with a special focus on high-risk groups. This approach might serve not only for expert opinions in court, but also for risk management strategies and therapeutic interventions.

Machine‐learning prediction of adolescent alcohol use: a cross‐study, cross‐cultural validation

“…Machine-learning studies in the domain of substance use are based on the use of regularized regression models (i.e. For instance, random forests (RF), support vector machines (SVMs) and artificial neural networks (ANN) are being used for prediction in other areas of health sciences [31][32][33]. These algorithms provide explicit control over the complexity of the fitted model and deliberately promote simpler and more interpretable solutions compared to non-regularized models.…”

“…However, previous substance use machine-learning studies fail to consider algorithms other than regularized regression models. For instance, random forests (RF), support vector machines (SVMs) and artificial neural networks (ANN) are being used for prediction in other areas of health sciences [31][32][33]. Considering the complexity of pathological mechanisms and non-linearity of the predictive effects, these techniques might improve the chances of prediction of pathological trajectories, treatment response and matching patients to the most effective treatments [34].…”

Aspectos criminogénicos de la reincidencia y su problema

“…Desde el mismo se promulga la existencia de ocho variables que explicarían de manera significativa el riesgo de reincidir, entre estas aparecen: a) historia individual de conducta antisocial, b) patrón de personalidad antisocial, c) cognición antisocial, d) redes o vínculos antisociales, e) ámbito familiar relacionado con bajo nivel de satisfacción, f) dificultades en el ámbito escolar y laboral, g) manejo del tiempo de ocio, h) abuso de sustancias psicoactivas En esta línea, los resultados de estudios, como el realizado por Mastrorilli, et al 46 , encontraron que la carrera pro-criminal emerge como un fuerte predictor. Ratificado por Pflueger, et al 47 , cuando en los resultados de su estudio de corte longitudinal "Predicting general criminal recidivism in mentally disordered offenders using a random forest approach" concluyeron que las variables con mayor nivel de predicción de la reincidencia corresponden a historial pro-criminal, edad y versatilidad delictiva. Esta última variable, retomada de forma general en el estudio de Olson, et al 48 , cuya investigación "Comparing male and female prison releasees across risk factors and postprison recidivism" concluye, a partir de una muestra de 26.534 infractores entre los 35 y 50 años, que el historial criminal extenso aumenta la probabilidad de reincidir.…”

Functioning of psychopathy and trait aggression as predictive variables of criminal recidivism

“…Establishing a negative relationship between psychiatric diagnosis and recidivism would help mitigate the stigma surrounding persons with a mental health disorder and the supposed dangers inherent in their condition (Fazel & Yu, 2011). However, it is important to note that other studies on the relationship between the presence of mental health problems and recidivism have found that inmates with a psychiatric disorder are more prone to reoffending and repeat incarceration compared with individuals with a psychiatric disorder but no previous incarceration (Baillargeon, Binswanger, Penn, Williams, & Murray, 2009;Pflueger, Franke, Graf, & Hachtel, 2015;Segeren, de Wit, Fassaert, & Popma, 2017).…”