Artificial Intelligence in Depression – Medication Enhancement (AID-ME): A Cluster Randomized Trial of a Deep Learning Enabled Clinical Decision Support System for Personalized Depression Treatment Selection and Management

Abstract

Major Depressive Disorder (MDD) is a leading cause of disability and there is a paucity of tools to personalize and manage treatments. A cluster-randomized, patient-and-rater-blinded, clinician-partially-blinded study was conducted to assess the effectiveness and safety of the Aifred Clinical Decision Support System (CDSS) facilitating algorithm-guided care and predicting medication remission probabilities using clinical data. Clinicians were randomized to the Active (CDSS access) or Active-Control group (questionnaires and guidelines access). Primary outcome was remission (<11 points on the Montgomery Asberg Depression Rating Scale (MADRS) at study exit). Of 74 eligible patients, 61 (42 Active, 19 Active-Control) completed at least two MADRS (analysis set). Remission was higher in the Active group (n = 12/42 (28.6%)) compared to Active-Control (0/19 (0%)) (p = 0.01, Fisher’s exact test). No adverse events were linked to the CDSS. This is the first effective and safe longitudinal use of an artificial intelligence-powered CDSS to improve MDD outcomes.