THE ANALYTICAL FRAMEWORK OF CLINICAL TRIALS EVALUATING CLINICAL OUTCOMES OF ARTIFICIAL INTELLIGENCE-BASED DIGITAL HEALTH INTERVENTIONS FOR MENTAL HEALTH DISORDERS: A SYSTEMATIC LITERATURE REVIEW
Author(s)
Vlad Zah, PhD, Dimitrije Grbic, PhD (c), Filip Stanicic, PhD (c);
ZRx Outcomes Research, Inc., Mississauga, ON, Canada
ZRx Outcomes Research, Inc., Mississauga, ON, Canada
OBJECTIVES: This systematic literature review (SLR) evaluated methodological approaches for trials assessing clinical outcomes of artificial intelligence-based digital health interventions (AI-DHI) among patients with mental health disorders.
METHODS: This SLR was designed in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines. PubMed and Embase databases were used for the main search, while Google Scholar platform for hand search. Patients with mental health disorders using AI-DHI were the target population in clinical trials that evaluated clinical outcomes. Only studies written in English were considered, without publication time restrictions. Quality appraisal was performed based on the National Institute for Health and Care Excellence checklist.
RESULTS: After the stepwise screening process, 16 clinical trials were included in the evidence synthesis. Only two studies (12.5%) were published before 2020, while half of the studies were conducted only in US. The most common indications were depression (43.8%), addiction (25.0%), and anxiety (18.8%). Clinical trials were mostly designed as controlled, parallel-group studies with 2 or more arms (94.0%). Standard-of-care and waitlist controls were the most common comparators. The double-blind design was mostly not feasible (only 18.8%), probably because of the AI-DHI type. Most studies were open-label (56.3%). Only a single study was performed in multiple countries. The use of power analysis to derive the required sample size was reported in 37.5% of trials. Dropout rates <20% were reported in 62.5% of trials. The most common outcomes pointing out the AI-DHI’s clinical value were abstinence rate in addiction, State-Trait Anxiety Inventory in anxiety, and General Anxiety Disorder-7 questionnaire in depression trials. Regression analyses were mainly used to assess statistical significance of outcome measures.
CONCLUSIONS: The SLR summarized methodological characteristics of published trials investigating clinical value of AI-DHI in mental health disorders. The results will serve as a guidance for planning future research.
METHODS: This SLR was designed in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines. PubMed and Embase databases were used for the main search, while Google Scholar platform for hand search. Patients with mental health disorders using AI-DHI were the target population in clinical trials that evaluated clinical outcomes. Only studies written in English were considered, without publication time restrictions. Quality appraisal was performed based on the National Institute for Health and Care Excellence checklist.
RESULTS: After the stepwise screening process, 16 clinical trials were included in the evidence synthesis. Only two studies (12.5%) were published before 2020, while half of the studies were conducted only in US. The most common indications were depression (43.8%), addiction (25.0%), and anxiety (18.8%). Clinical trials were mostly designed as controlled, parallel-group studies with 2 or more arms (94.0%). Standard-of-care and waitlist controls were the most common comparators. The double-blind design was mostly not feasible (only 18.8%), probably because of the AI-DHI type. Most studies were open-label (56.3%). Only a single study was performed in multiple countries. The use of power analysis to derive the required sample size was reported in 37.5% of trials. Dropout rates <20% were reported in 62.5% of trials. The most common outcomes pointing out the AI-DHI’s clinical value were abstinence rate in addiction, State-Trait Anxiety Inventory in anxiety, and General Anxiety Disorder-7 questionnaire in depression trials. Regression analyses were mainly used to assess statistical significance of outcome measures.
CONCLUSIONS: The SLR summarized methodological characteristics of published trials investigating clinical value of AI-DHI in mental health disorders. The results will serve as a guidance for planning future research.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
MT23
Topic
Medical Technologies
Topic Subcategory
Digital Health