J Med Internet Res. 2025 Mar 7. doi: 10.2196/69351. Online ahead of print.
ABSTRACT
BACKGROUND: Escalating mental health demand exceeds existing clinical capacity, requiring scalable digital solutions. However, engagement remains challenging. Conversational agents enhance engagement by making digital programs more interactive and personalized but have not been widely used. This study evaluated a digital program for anxiety against external comparators. The program used an AI-driven conversational agent to deliver clinician-written content via machine learning, with clinician oversight and user support.
OBJECTIVE: This study aimed to evaluate the engagement, effectiveness, and safety of this structured, evidence-based digital program with human support for mild, moderate and severe generalized anxiety. Statistical analyses aimed to determine whether the program reduced anxiety more than a propensity-matched waiting control and was statistically non-inferior to real-world propensity-matched face-to-face and typed cognitive behavioral therapy (CBT).
METHODS: Prospective participants (N=299) were recruited from the NHS or social media in the UK and given the digital program to use for up to 9 weeks (study conducted from October 2023 to May 2024). Endpoints were collected before, during and after the digital program, and at one-month follow-up. External comparator groups were generated through propensity-matching of the digital program sample with NHS Talking Therapies (NHS TT) data from ieso Digital Health (typed-CBT) and Dorset Healthcare University NHS Foundation Trust (DHC) (face-to-face CBT). Superiority and non-inferiority analyses were conducted to compare anxiety symptom reduction (change on GAD-7 scale) in the digital program group and the external comparator groups. The program included human support and clinician time spent per participant was calculated.
RESULTS: Participants used the program for a median of 6 hours over 53 days, with 78% (n=232) engaged (i.e. completed a median of 2 hours over 14 days). There was a large clinically meaningful reduction in anxiety symptoms for the digital program group (per-protocol (PP; n=169): change on GAD-7 = -7.4, d = 1.6; intention-to-treat (ITT; n=299): change on GAD-7 = -5.4, d=1.1). The PP effect was statistically superior to the waiting control (d = 1.3), and non-inferior to the face-to-face CBT group (p <.001) and the typed-CBT group (p <.001). Similarly, for the ITT sample, the digital program showed superiority to waiting control (d=0.8) and non-inferiority to face-to-face CBT (p=.002) with non-inferiority to typed-CBT approaching significance (p=.06). Effects were sustained at one-month follow-up. Clinicians overseeing the digital program spent a mean of 1.6 hours (31 – 200 minutes) of clinician time in sessions per participant.
CONCLUSIONS: By combining AI and human support, the digital program achieved clinical outcomes comparable to human-delivered care while significantly reducing the required clinician time by up to 8 times relative to global care estimates. These findings highlight the potential of technology to scale evidence-based mental healthcare, address unmet need, and ultimately impact quality of life and economic burden globally.
CLINICALTRIAL: ISRCTN id: 52546704.
PMID:40152000 | DOI:10.2196/69351
