PROJECT SUMMARY/ABSTRACT The opioid crisis is the deadliest drug epidemic in American history and new approaches are needed. One novel approach includes predicting likelihood of opioid use disorder (OUD) treatment retention by assessing someone’s risk of early departure from treatment. Current methods to improve treatment retention rely on providers using their intuition to identify when an individual is at risk of leaving treatment early in order to intervene, which often happens too late. Mobile health and machine learning predictive analytics offer a new opportunity to personalize OUD treatment, improve retention in OUD care, and mitigate the risk of relapse and overdose episodes. Project Motivate will combine physiological and behavioral data from disparate sources in order to predict when an individual is at risk of early departure from OUD treatment. This data will be displayed in a user-friendly manner so that providers can more effectively support patients to remain in treatment with timely intervention and responses.