Optimization of health care through a generative artificial intelligence-based chatbot platform: A systematic review. (#828)

Authors

Córdova-Berona, Heli

Briones Zuñiga, Jose Luis

Nuñez Diaz, Luis Alberto

Vargas Calderon, Arnold Cesar

Abstract

Limited healthcare accessibility represents a major challenge for healthcare systems globally, with more than 70% of facilities in developing countries lacking adequate technological infrastructure. The limited evidence on the effective implementation of generative AI-based chatbots exacerbates this problem, limiting service optimization. This systematic review analyzes how generative AI chatbot platforms optimize healthcare. A review was conducted using PICO components and the PRISMA protocol, selecting 40 articles from Scopus, Web of Science, and EBSCOhost (2020–2024). The results indicate that limitations were addressed through relational agents, gamified chatbots with OMO strategies, culturally appropriate adaptive systems, and empathetic frameworks. Effectiveness was validated through experimental studies, demonstrating improvements of 95% in continuous availability, 90% in geographic coverage, 85% in response time, and 78% in adherence versus the traditional 60%. Mental health emerged as the most effective sector, with 30% of successful implementations. In conclusion, generative AI chatbots are effective tools for overcoming traditional access barriers, with variations by sector. The quality of implementation outweighs the specific type of chatbot, with cultural adaptation and customization determining sustainable success. Large-scale studies in low- and middle-income countries and integration with medical IoT technologies are suggested.