JMIR Publications

ABSTRACT

Background:

Conversational agents (CAs) are increasingly used to deliver scalable interventions for depression and anxiety, but existing evidence has often treated these tools as a single category. The comparative effectiveness of different agent types, intervention doses and levels of human support remains unclear.

Objective:

This review systematically evaluated the comparative effectiveness of various CA modalities and levels of human support, as well as potential dose-response relationships across different CA types.

Methods:

A comprehensive search of PubMed, Embase, Web of Science, Scopus, Ovid, CINAHL, and PBSC databases was conducted from inception to December 2025. Eligible studies were randomized controlled trials evaluating CA-based interventions. We conducted random-effects network meta-analysis (NMA) using post-intervention outcomes and assessed inconsistency and risk of bias. Evidence certainty was assessed using the CINeMA framework. Where data permitted, model-based network meta-analysis (MBNMA) was performed to examine dose-response associations.

Results:

A total of 41 studies (n = 7,577) were included. NMA showed that AI-Enhanced Intensive Human Support appeared to produce greater symptom improvement for depression and anxiety compared with Pure Human Digital Guidance (g = -4.28, 95% CI: -5.56 to -2.99) and Open-Ended Multimodal-Enhanced Chatbot (g = -4.49, 95% CI: -6.31 to -2.66), whereas most other comparisons were not statistically significant. MBNMA indicated that Goal-Oriented CA and Highly Structured Modular Chatbot appeared to exhibit nonlinear effects, with optimal response at low-to-moderate doses. However, publication bias was detected for depression (Egger’s P = 0.018), and the overall certainty of evidence was very low, primarily due to risk of bias, indirectness, and imprecision.

Conclusions:

While hybrid CA models with enhanced human involvement appeared to show potential advantages over fully automated systems, the evidence base remains limited. The very low certainty of evidence constrains confidence in comparative rankings and dose recommendations. Future high-quality trials are needed to establish optimal CA architectures and intervention doses for scalable mental health care.