- AI chatbots are feasible and acceptable in chronic disease care, notably improving smoking cessation and mental health outcomes in some studies.
- Observed benefits include higher smoking cessation rates and reduced depression and anxiety, but overall evidence certainty is low to very low.
- Methodological limitations include small samples, short follow-up, high attrition, and absence of controls; large RCTs with objective clinical endpoints are required.
Cureus. 2026 Jun 22;18(6):e111334. doi: 10.7759/cureus.111334. eCollection 2026 Jun.
ABSTRACT
Chronic diseases impose a substantial burden on healthcare systems, requiring innovative approaches to support long-term self-management. AI-driven chatbots have emerged as a scalable digital health intervention, but the evidence base remains fragmented. This systematic review synthesizes current evidence on the role of AI chatbots in chronic disease care and evaluates their impact on clinical and behavioural outcomes. A comprehensive literature search was conducted in PubMed, Embase, Scopus, and Web of Science for studies published between 2021 and 2025. The review followed PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines, and the narrative synthesis was reported in accordance with the Synthesis Without Meta-analysis (SWiM) guidelines; the certainty of the evidence was rated using the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) approach. Risk of bias was assessed using the Cochrane RoB 2 tool for RCTs and the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for non-randomized studies (inter-reviewer agreement Cohen κ = 0.84-0.90). Eleven studies met the inclusion criteria, covering conditions such as diabetes, hypertension, asthma, cancer, mental health disorders, and smoking cessation. Chatbot technologies ranged from rule-based systems to advanced natural language processing (NLP) and large language models (LLMs). AI chatbots improved smoking cessation rates (26% vs. 18.8%; absolute difference 7.2 percentage points), reduced depression and anxiety in some studies, and demonstrated high usability and acceptability. No serious adverse events were reported. However, methodological limitations included small sample sizes, short follow-up periods, high attrition, and the absence of control groups, and the overall certainty of the evidence was low to very low across outcome domains. AI chatbots are feasible and acceptable for chronic disease care, particularly for smoking cessation and mental health support; however, current evidence is insufficient to confirm effects on hard clinical endpoints. Future research requires large-scale RCTs with objective clinical endpoints and longer follow-up.
PMID:42495491 | PMC:PMC13392408 | DOI:10.7759/cureus.111334
Share Evidence Blueprint
Save to Google Notes

Search Google Scholar
⭐ My Revision List

