29.4 IMPLEMENTING PRECISION PSYCHIATRY FOR THE EARLY RECOGNITION OF ADVERSE OUTCOMES IN PSYCHOSES: FINDINGS FROM THE PRONIA STUDY

Abstract

Background The clinical high-risk state for psychosis does not only confer an elevated risk for developing psychotic disorders but is associated with pluripotent risks for adverse clinical and functional outcomes. Establishing generalizable tools that provide quantitative risk estimates for these outcomes is a key step toward the implementing personalized preventive intervention that scale beyond the Methods The talk will present recent findings from the PRONIA study (Personalized Prognostic Tools for Early Psychosis Management) demonstrating the feasibility of predicting functional and clinical outcomes in adolescents and young adults in a clinical high-risk state for psychosis (CHR) or with recent-onset depression (ROD). The talk will highlight the use of machine learning and multivariate data mining concepts and link those applications to potential clinical utility of these models for an improvement of early recognition and prevention Results I will present and discuss the performance and decision rules generated by the machine learning analysis of clinical, imaging-based, genetic and combined data for the individualized prediction of (1) social and role functioning outcomes, (2) transition to psychosis, and (3) remission vs. non-remission from symptomatic states in CHR and ROD patients. Conclusions The recent findings generated by the PRONIA consortium suggest that generalizable and clinical useful prediction pathways can be established to support the early recognition of adverse outcomes in CHR and ROD patients. External and prospective validation of these prognostic pathways is needed across healthcare systems to benchmark the clinical and health economic utility of these precision psychiatry methods.