Symposia
Technology/Digital Health
Shiri Sadeh-Sharvit, Ph.D. (she/her/hers)
Chief Clinical Officer
Eleos Health
Palo Alto, California, United States
Eyal Grossman, MA (he/him/his)
Clinical Analyst
Eleos Health
Needham, Massachusetts, United States
Steven D. Hollon, Ph.D. (he/him/his)
Professor of Psychology and Human Development, Psychiatry
Vanderbilt University
Brentwood, Tennessee, United States
In response to the growing demand for scalable and effective mental health services, community-based treatment centers are exploring innovative solutions. This study examines the effects of implementing artificial intelligence (AI) in outpatient therapy at one such center. The randomized controlled trial investigated the feasibility, acceptance, and initial effectiveness of an AI platform in improving clinical outcomes for clients receiving cognitive behavioral therapy (CBT) for depressive or anxiety disorders. Conducted at the Ozark Center, a community-based clinic in Western Missouri, the study included 47 adults referred for outpatient, in-person individual therapy. Utilizing an AI platform that enabled clinicians to gather objective patient data, streamline workflow processes, and automate administrative tasks, we compared treatment-as-usual (TAU) to therapy supported by AI for the initial 2-month therapy period. The AI platform summarized therapy sessions, provided feedback to therapists on evidence-based practices, and integrated patient self-report data into session progress notes. Findings indicated that participants whose therapists utilized the AI platform attended 67% more therapy sessions compared to TAU counterparts. Moreover, depression and anxiety symptoms decreased by 34% and 29%, respectively, in the AI group, with therapists submitting progress notes 55 hours earlier on average than TAU therapists. Overall, therapy supplemented by the AI platform demonstrated superior outcomes in depression and anxiety reduction, patient retention, and session note efficiency compared to standard therapy. Therapists also reported they had benefited from AI-based feedback on their implementation of CBT techniques. These findings underscore the potential of integrating AI into community-based clinic settings to enhance behavioral treatment effectiveness and care delivery.