Artificial intelligence-based chatbots to enhance medication adherence among patients with non-communicable chronic diseases: Systematic review and meta-analysis.

Medication adherence remains a major public health challenge among patients with non-communicable diseases (NCDs) worldwide. Artificial intelligence (AI)-based chatbots may enhance adherence by automating reminders, education, and real-time support. This systematic review and meta-analysis evaluated the effectiveness of AI-based chatbots in improving medication adherence among patients with NCDs. This review (CRD420251151031) is reported in accordance with the PRISMA guidelines. Relevant studies were identified from PubMed, MEDLINE, Embase, Web of Science, Global Health, CINAHL, Cochrane Library, APA PsycINFO, and APA PsycArticles up to August 2025. Eligible designs included randomized controlled trials (RCTs), quasi-experimental studies, and single-arm pre-post studies. Seven studies published between 2017 and 2025 were included, of which six RCTs contributed to the meta-analysis. The pooled standardized mean difference (SMD) for medication adherence was 0.69 (95% confidence interval: 0.17 to 1.22; p = 0.01), indicating a significant medium effect with high heterogeneity (I2 = 97%). Subgroup analyses revealed greater effects for cardiovascular diseases and short-term interventions (<6 months). Daily and weekly to monthly interventions were effective, whereas on-demand or episodic interventions were not. Significant improvements were also observed when adherence was assessed using self-reported measures or pill counts. AI-based chatbots that incorporated three functions (medication reminders, education and health coaching, and real-time question-answer features) showed significant improvement in medication adherence, whereas those with dual functions did not. These findings indicate that AI-based chatbots significantly improve medication adherence in NCDs. Chatbots have the potential to supplement existing interventions to support medication adherence.