Practical application of machine learning in forensic psychiatric research and its clinical implications

Practical application of machine learning in forensic psychiatric research and its clinical implications

Paper presentation102Johannes Kirchebner, University Hospital of Psychiatry Zurich, Switzerland; Lena Antonia Hofmann, University Hospital of Psychiatry Zurich, Switzerland; Steffen Lau, University Hospital of Psychiatry Zurich, Switzerland; Andrea Aemmer, University Hospital of Psychiatry Zurich, Switzerland

Schadee ZaalSat 09:00 - 10:30

Research: The work of the research group was based on a comprehensive analysis of offender patients with a schizophrenia spectrum disorder (SSD) regarding different difficult treatment courses via machine learning (ML). Based on these findings, a prognostic tool for their occurrence is currently under development.Methodology: A database of 370 offender patients with SSD with over 500 variables was created and analyzed via ML for different adverse treatment events.Findings: The different difficult treatment courses were based on similar predictors: biographical information, severity of illness according to PANNS, and antisocial behaviors in the past. The models achieved notable AUC above 0.75.Practice and Relevance: Psychiatric illnesses and resulting behavioral problems are complex phenomena driven by various, often interdependent factors. ML offers the possibility of an according analysis. The identified predictors of difficult treatment courses in forensic psychiatry are currently compiled in a screening tool.

Augmented Reality, Virtual Reality)
machine learning, schizophrenia spectrum disorder, screening tool development
No Comments

Sorry, the comment form is closed at this time.