Abstract: Online health communities (OHCs) are an important platform for internet-based medical services and knowledge dissemination in health education in China. However, previous studies often treat online consultation and health promotion as two separate research questions, with limited research on their potential associations. Given this research gap, this study constructs a research model by integrating the signaling perspective within the framework of social capital theory and employs text mining and statistical modeling methods to explore the multidimensional effects of physicians’ social capital in the context of health knowledge popularization in OHCs, and examine how it affects physicians’ online service performance. For sample selection, this study takes China’s most representative online health community platform (HaoDF Online) as the research context and obtains abundant research data from it. The researchers first train text classification models using machine learning methods to automatically identify the topics of doctors’ popular science articles, which greatly improves the accuracy and coverage of sample selection. Next, this study employs word vector techniques to compare the similarity between doctors’ health popular science articles and patients’ consultation text, and quantitatively extracts the “doctor-patient shared language” variable that reflects the readability of physicians’ health popular science articles. Based on the above work, this study constructs an empirical dataset with accurate mapping relationships between popular science article topics and online consultation diseases, which lays the data foundation for subsequent fine-grained causal relationship verification. Finally, this study adopts multi-level regression models to simultaneously examine the interaction effects among independent, mediating and moderating variables. The results show that doctors’ use of doctor-patient shared language in health popular science writing has a positive impact on their online consultation volume, and doctor-patient reciprocal feedback mediates this process. Furthermore, doctors’ professional qualifications can significantly moderate the effect of doctor-patient reciprocal feedback on online consultation volume. These findings extend the theoretical connotation and interpretation of social capital theory. They also provide important references for doctors to provide high-quality popular science knowledge and promote efficient OHCs operation.

Key words: doctor-patient shared language, physicians’ online service performance, doctor-patient reciprocal feedback, professional qualifications, social capital theory