Graphical Modelling in Mental Health Risk Assessments

Abstract

Probabilistic models can be a combination of graph and probability theory that provide numerous advantages when it comes to the representation of domains involving uncertainty. In this paper, we present the development of a chain graph for assessing the risks associated with mental health problems, which is a domain that has high amounts of inherent uncertainty. The Galatean mental health Risk and Social care Tool, GRiST, has been developed to support mental-health risk assessments by using a psychological model to represent the expertise of mental-health practitioners. It is a hierarchical knowledge structure based on fuzzy sets for reasoning with uncertainty. This paper describes how a chain graph can be developed from the psychological model to provide a probabilistic evaluation of risk that complements the one generated by GRiST’s clinical expertise.