AI Chatbots and Conversational Agents for Health Communication in Low- and Middle-Income Countries: A Scoping Review of Applications, Outcomes, and Implementation Barriers

<title>Abstract</title> <p>Background Conversational agents or AI-enabled chatbots are scalable digital health tools capable of delivering personalized information, supporting self-management, promoting healthy behaviors, and facilitating engagement with healthcare services. However, evidence regarding their application in low- and middle-income countries (LMICs) remains fragmented. We synthesized available evidence on the utilization of AI chatbots and rule-based conversational agents for health communication and health-related outcomes in LMICs to identify their functional roles and implementation barriers. Methods Following PRISMA-ScR guidelines, a systematic search was conducted across PubMed, ScienceDirect, Cochrane Library, IEEE Xplore, and Google Scholar. Primary empirical and mixed-methods studies set in LMICs that assessed the use of conversational agents or chatbots in health communication, patient education, or health-related outcomes were eligible for inclusion. Results Of 1,803 records identified, nine studies involving over 141,000 users were included. Chatbots or conversational agents were utilized across HIV prevention, sexual and reproductive health, tuberculosis care, diabetes, hypertension, oral health, and childhood immunization. Eight studies reported improvements in health knowledge, while seven reported positive behavioral changes such as enhanced medication adherence. Chatbot interventions demonstrated high acceptability, satisfaction, and user engagement. Although some interventions improved self-management, objective clinical outcomes were infrequently assessed and findings were mixed. Common implementation barriers included digital access inequities, language and literacy challenges, declining user engagement, and minimal health-system integration. Conclusions AI chatbots hold diverse potential for enhancing health literacy and behavioral outcomes across LMICs. However, more rigorous evaluations are needed to establish long-term clinical effectiveness, sustainability, and equitable implementation.</p>