Supplementary Material for: Commercial AI-based Mental Health Chatbots as Low-Intensity Ad

A 2025 systematic review and meta-analysis of 52 studies (N=110,594) covering 13 commercial AI mental health chatbots. The research, published on figshare, synthesizes evidence from randomized controlled trials conducted between 2015 and 2025 to evaluate effectiveness, adherence, and safety. It includes meta-analytic results for depression, anxiety, and loneliness outcomes.

Use Cases

• Benchmarking the efficacy of AI chatbots for depression based on pooled effect sizes (Hedges' g) from 18 RCTs.
• Analyzing user adherence and scalability based on the reported 79.5% pooled completion rate and attrition risk ratios.
• Evaluating safety reporting standards in digital health trials based on the finding that only 3 out of 22 RCTs met active-ascertainment criteria.
• Assessing the heterogeneity of outcomes for anxiety and affect based on the reported confidence intervals and certainty ratings.

Strengths

• Based on a substantial evidence base of 52 studies including 22 randomized controlled trials.
• Provides quantitative meta-analysis results, including specific effect sizes (e.g., g = -0.35 for depression) and confidence intervals.
• Explicitly evaluates multiple dimensions: effectiveness, adherence (79.5% completion), and safety reporting.

Limitations

• The dataset is a 16.9 MB PDF document; the underlying structured data (e.g., individual study data) is not provided.
• Column-level documentation is absent; field semantics must be inferred from the published article.
• The certainty of evidence for outcomes ranged from low to very low according to the review authors.

Provenance

Source

figshare admin karger

Collection Method

Systematic review and meta-analysis of peer-reviewed studies from eight databases.

Time Range

Studies published from January 1, 2015 to December 31, 2025.

Freshness

Last updated 2026-05-09 07:55:30.