Background:The advent of COVID-19 has hindered people's chances to meet in real life, but increased the frequency of communication in online communities. Patients with similar disease experiences and their families independently form health communities based on online platforms, and seek various social support through communication. However, the provision and seeking of social support in online health communities are not always in a balanced state, and not every user has the same chance to receive replies or social support in the form of replies. A large number of empirical studies have discussed the characteristics of users who seek social support to get more responses, but previous studies have overlooked two important factors. Firstly, its independent variable is usually the subjective data reported by users, while social network analysis can provide scientific objective data to analyze the role of user attributes. Second, they ignore the role of the request text, which is how the type of social support affects the number of responses a user receives.

Methods: The core issue of this study is the influence of users' network centrality on the number of replies to social support, and what kind of social support will get more replies in the community.Most importantly, how the types of social support interact with users' characteristics to affect the number of replies received by users. In this study, more than 300,000 words of chat logs of depressed communities over four months from 2021 to 2022 were collected for content analysis, and social network analysis was conducted based on their response relationships to measure users' network attributes. In this study, the number of responses received was taken as the dependent variable, and network attributes and types of social support were taken as the independent variables to discuss the influencing mechanism of social support in the community.

Conclusion: This study found that users'network centricity positively predict the number of social support received they reply.Users who seek for emotional support will get more replies.The interaction between social support seeking types and users' network characteristics was significant.In the information and tools support level, the number of them received reply less affected by the user attributes.In the aspect of emotional support,the number of replies received is greatly affected by user attributes.