Assessing and managing the suicidal patient: forget the Reverend Bayes and try game theory.

Abstract

Probability-based estimates of the future suicide of psychiatric patients are of little assistance in clinical practice. This article proposes strategic management of the interaction between the clinician and the patient in the assessment of potentially suicidal patients, using principles derived from game theory, to achieve a therapeutic outcome that minimises the likelihood of suicide. Further developments in the applications of large language models could allow us to quantify the basis for clinical decisions in individual patients. Documenting the basis of those decisions would help to demonstrate an adequate standard of care in every interaction.