Use and Efficacy of Conversational AI in Health Management: An Integrative Review

Abstract

Conversational artificial intelligence (AI) is increasingly being integrated into the field of health communication. Existing research largely evaluates these technologies in terms of usability, technical performance, or behavioral outcomes, with less attention paid to the communication processes that drive these outcomes. To fill this research gap, this review provides a comprehensive analysis of empirical research on conversational AI from a health communication perspective. Based on a theory-oriented integrative review of 65 studies, this study focuses on the mechanisms by which conversational AI plays a role in health interactions and the reasons for differences in communication outcomes across different contexts. This review uses a thematic synthesis approach to identify four core communication roles of conversational AI in health communication: information communicator, behavioral interventionist, emotional supporter, and service mediator. Beyond identifying these roles, the findings reveal significant differences in reported communication outcomes. To explain these differences, this study constructs a Conversational AI-Mediated Health Communication (CA-MHC) framework. This framework conceptualizes conversational AI as a mediated communication agent whose impact depends on the fit between its communication role and four dimensions of communication configuration: interaction pattern, relationship strength, temporal orientation, and institutional embedding. This study also provides a conceptual foundation for understanding when and why conversational AI can support or hinder effective health communication. Furthermore, the CA-MHC framework provides guidance for the future theoretical development and empirical research of AI-mediated health interactions.