Forensic Summary by Kiana DelGrosso
In the Decision Support Systems volume 80 journal, Anthony Constantinou, Mark Freestone, William Marsh, and Jeremy Coid used Bayesian inference in order to construct diagrams for pertaining to violence risk management and decision support situations in forensic psychiatry in their article Casual Inference for Violence Risk Management and Decision Support in Forensic Psychiatry. An introduction of Bayesian applications was given so as to inform the reader how this inference system can be used in real-world applications. In order to use this inference system, data must be given so this study used datasets given by two questionnaires, interviews, and assessment results. One dataset focused on risk assessment of 386 released patients and the other was a prisoner cohort study which had 953 inpatients. Both datasets were then analyzed and narrowed down to only include mentally ill individuals. Multiple diagrams were created after each piece of data was analyzed in order to create the most precise real-world application of this Bayesian inference.
Their results were separated into two specific models; data-driven and expert-drive. There was significant improvement in predicting violence among inpatients and outpatients that allows decision support to be well-established. This helps forensic psychiatrists in determining which patients are to be released based upon a questionnaire, an interview, and assessment of the patient’s time within the facility. Using the Bayesian system can increase the predictability of a patient’s future behavior in a real-world model which can assist forensic psychiatrists in their evaluations.
Anthony Costa Constantinou, Mark Freestone, William Marsh, Jeremy Coid, Causal inference for violence risk management and decision support in forensic psychiatry, Decision Support Systems, Volume 80, December 2015, Pages 42-55, ISSN 0167-9236, http://dx.doi.org/10.1016/j.dss.2015.09.006.