Photo caption: Nicole Obeid, Scientist, CHEO Research Institute.
By: Kelsea Towell, CHEO Research Institute
Eating disorders affect more than 1.4 million youth in Canada and are among the most complex and debilitating mental health conditions. For decades, treatment has relied on a one-size-fits-all approach, limited to small sample sizes and lack of data from minority communities – but CHEO researchers are changing that.
The CHEO Research Institute’s Eating Disorders Research Lab in collaboration with an interdisciplinary team of scientists, clinicians, and data science experts have created EDBioMAP (Eating Disorder Bio Registry and Multiaxial Precision Health Platform), the first large pediatric eating disorders bio-registry and multiaxial precision health platform.
Through leveraging existing data and artificial intelligence (AI), this project will identify a set of best indicators for detecting different eating disorder presentations in diverse youth. This is the first step to develop an AI tool for eating disorders that can support clinicians’ decision-making and personalize care.
Rising Incidence Requires an Innovative, Data-Driven and AI Solution
Eating disorders affect many parts of the body and can lead to serious medical problems, mental health challenges, and frequent relapse. Current treatment approaches are often similar and based on limited evidence, which contributes to relatively low remission rates of about 35 to 40 percent. Cases are increasing among adolescents, including boys and children as young as 12 to 14, highlighting the urgent need for new and more effective treatment options.
“No two youth experience an eating disorder the same way,” said Dr. Nicole Obeid, Scientist and Lead of the Eating Disorder Research Lab at the CHEO Research Institute. “Finally, through this data we have a diverse representation that reflects our population by linking clinical, psychosocial, and biological markers so we can detect risk earlier, match treatments smarter, and support recovery more personally for every child and family.”
Personalizing Eating Disorder Care for Better Diagnosis and Treatment
The EDBioMAP platform incorporates six years of retrospective data from the CHEO Eating Disorders Research Lab from 2018 to 2024, integrating clinical, psychosocial, and biological information, along with biomarkers generated through ongoing microbiome, genomic, and imaging research. Using AI‑driven predictive modeling, including clustering and data augmentation, the platform uncovers complex and clinically relevant patterns in the data, supporting more accurate predictions while enhancing fairness and ensuring the results can be generalized across diverse patient groups.
- Build a large-scale, multiaxial dataset combining clinical, psychosocial, and biological markers.
- Use AI-driven predictive modeling to identify phenotypic signatures and guide personalized interventions.
- Develop a platform that learns and evolves, tailoring care to each child’s unique biology, environment, and lived experience.
Impact Beyond Hospital Walls
This work has the potential to revolutionize the personalization of eating disorder treatment, improving outcomes for youth and their families across Canada. Leveraging data to deliver evidence-based, individualized care for pediatric eating disorders, means moving beyond symptom-focused treatment to whole-child care.
This work was made possible thanks to the CHEO Research Institute’s Precision Child and Youth Mental Health (PCYMH) Collaboratory , which uses comprehensive information about the individual ranging from the DNA code to the postal code, seeks to improve mental health in the current generation and beyond through collaborative precision mental health research and care.