Clinical Practice Data Research Project - Abstract Numerous studies have demonstrated the effectiveness of coordinated specialty care (CSC) for the treatment of people with a first episode of psychosis (FEP). However, there is a major problem of patient disengagement, which has adversely affected the impact of CSC on the long-term course of FEP. The lack of an individualized, empirically developed approach to assess risk of disengagement for a given patient has compromised the ability of the field to address this issue. Consistent with the transformative movement towards precision medicine, individualized risk calculators for many medical conditions have been developed, validated and applied in clinical settings. On the heels of these successes, the mental health field has seen a proliferation of research aiming to develop and validate risk- prediction models for various mental health conditions, however research that fosters ethical, practical and clinically useful implementation of risk prediction in applied settings has lagged behind. Our primary objective is to develop, validate, and lay the groundwork for implementing a risk calculator for CSC program outcomes. A precision psychiatry approach to enhancing retention in CSC programs is a logical and necessary next step to advance this field and optimize CSC client outcomes. Our central hypothesis is that a calculator to predict personalized CSC program outcome will allow us to identify participants most likely to disengage from care. Our approach rests on our extensive experience: (a) implementing and analyzing the hub-wide and national Core Assessment Battery (CAB), (b) developing and validating risk-prediction models in related contexts, (c) demonstrating feasibility via the preliminary analyses of our admission CAB data that provide support for risk prediction with moderate accuracy, and (d) weighing ethical and practical issues related to risk prediction. Our long-term goal is to develop a personalized medicine approach to identify and rapidly address risk for disengagement and retain individuals in CSC to optimize its benefit. This Prospective Practice-Oriented Research Project entitled Developing and Validating Models to Predict Risk of CSC Disengagement proposes to use CLHS CAB data to develop and validate longitudinal models for predicting risk of CSC disengagement and to integrate stakeholder input on feasibility, acceptability, utility, facilitators, and barriers to using risk information in clinical practice. We will develop several versions of a risk calculator predicting length of time in program, program completion, and disengagement (Aim 1) and establish longitudinal validity (2-3 years) to compare different calculators’ predictive accuracy (Aim 2). We will leverage the AC’s participatory research resources to develop implementation strategies for using risk information in practice (Aim 3).