Background Social media has transformed health communication into a dynamic and interactive process, shifting from one-way dissemination by experts to user-driven content creation and sharing. However, this openness also facilitates the spread of misinformation, which poses a threat to public health behaviors. While prior research has explored dialogue paths and narrative analysis independently, there remain gaps in understanding their interactive effects and the contextual heterogeneity-such as platforms, user groups, and emotions-on health behaviors. This study addresses these gaps within China’s social media landscape. Methods A multi-method approach was employed. Data collection Between January 2023 and December 2024, over 50,000 data points related to health communication were collected from prominent Chinese social media platforms, including Weibo, WeChat, Xiaohongshu, and Douyin. From this comprehensive dataset, a subsample of 300 valid user cases was identified for structural equation modeling (SEM), to ensure statistical adequacy and model robustness. Survey 300 valid questionnaires assessed user perceptions and behaviors regarding health information. Analysis Drawing upon health communication theory and narrative persuasion frameworks, a structural equation modeling (SEM) approach was employed to test the proposed conceptual model. The SEM analysis comprised two essential stages: (1) the measurement model stage, where the reliability and validity of latent constructs were evaluated through confirmatory factor analysis (CFA), including assessments of convergent and discriminant validity; and (2) the structural model stage, which examined the hypothesized relationships among dialogue path, narrative structure, engagement, and health behavior outcomes. Results The results of the structural equation modeling (SEM) indicate that both dialogue pathways and narrative strategies significantly influence users’ health information behaviors. Specifically, dialogue depth exhibited a strong positive effect on information sharing behavior, while narrative consistency was significantly related to feedback intention. The measurement model confirmed good reliability and validity, with all factor loadings exceeding 0.7 and composite reliabilities surpassing 0.8. Normality tests indicated acceptable skewness and kurtosis for all observed variables. Furthermore, multi-group analysis revealed that platform type moderates the strength of these relationships, with Weibo users demonstrating more emotionally driven interaction patterns compared to WeChat users. Conclusion This study identifies dialogue coherence, emotional narrative structure, and user engagement as significant predictors of health-related behavioral intentions within social media contexts. Notably, engagement serves as a crucial mediating variable that links both narrative and dialogic features to behavioral outcomes. These findings enhance our understanding of the mechanisms that underpin digital health communication, specifically by elucidating how narrative processes and dialogue influence user behavior. Practically, the results indicate that health communication initiatives on social media can be improved by integrating emotionally resonant narratives and ensuring coherence in dialogic exchanges, both of which are essential for fostering positive behavioral change. By analyzing dialogic interaction and narrative structure within a unified Structural Equation Modeling (SEM) framework, this research underscores the importance of considering storytelling elements alongside interactive features, thus providing a more comprehensive perspective on how users engage with and are influenced in digital health environments.
Wang et al. (Fri,) studied this question.