Grant: DRA : Canada Graduate Research Scholarship - Doctoral Program (CGRS D)

Eating disorders (EDs) are serious mental health conditions that often begin during adolescence. Many patients face a complex clinical course due to poor response to treatment or relapse after recovery. Currently, there are no reliable ways to predict which patients are at risk, partly because ED patient data collection has been historically limited and powerful machine learning techniques have not been widely applied in the ED research field. This project aims to address this gap by developing a unique patient dataset from 367 youth previously treated in CHEO's ED program. The dataset will combine clinical characteristics, psychological assessments, and routine biological tests from patients' first treatment episode, as well as their treatment episode trajectory into a "precision medicine" dataset. Machine learning models will be developed to predict and understand which patients are at risk for a complex clinical course based on their unique precision medicine profiles. The study will focus on: (1) Testing whether machine learning modelling with precision medicine data improves prediction accuracy compared to clinical data alone. (2) Identifying the key patient factors that most strongly predict clinical outcomes. (3) Exploring whether augmenting the dataset with artificial intelligence-generated synthetic data can improve model accuracy and support analysis of patient minority groups. This innovative study is the first to apply machine learning to precision medicine data for predicting ED clinical outcomes. This work could help develop a decision-support tool to identify high-risk youth early, enabling more personalized treatment and better use of limited program resources. Additionally, identifying key predictors may enhance understanding of ED causes, reveal important biomarkers, and guide future treatment targets. Finally, evaluating synthetic data methods could guide analysis of small psychiatric datasets in the future, advancing precision psychiatry.