AID-ME cluster RCT: Aifred Health deep-learning CDSS tested for personalized depression treatment selection across psychiatric sites

Depression treatment selection remains largely trial-and-error, with clinicians lacking data-driven tools to personalize antidepressant selection for individual patients. This leads to delayed remission, unnecessary medication switches, and poor patient outcomes across psychiatric settings.

A deep-learning-enabled clinical decision support system developed by Aifred Health was deployed across multiple psychiatric institutions in a cluster randomized trial. The system (AID-ME) ingested patient-level data to generate personalized treatment recommendations for depression medication selection and ongoing management.

The abstract on this PubMed page is truncated and specific outcome metrics are not visible. The trial was published in the Journal of Clinical Psychiatry (2025) and involved sites including Douglas Mental Health University Institute, Jewish General Hospital, CAMH, Emory University, Yale, University of Michigan, and the Salem VA Medical Center.

• • Multi-site cluster RCT design allows real-world generalizability across diverse psychiatric settings and patient populations.
• • Industry-academia partnerships (Aifred Health + McGill/Douglas) enable rigorous clinical validation of AI decision support tools.
• • Personalized treatment selection via deep learning targets a high-unmet-need area where standard-of-care remains largely empirical.