COVID-19 Vaccine–Related Information on the WeChat Public Platform: Topic Modeling and Content Analysis (Preprint)

Abstract

BACKGROUND The COVID-19 vaccine is an effective tool in the fight against the COVID 19 outbreak and is the best way to restore social functioning. As the main channel of information dissemination in the context of the epidemic, social media influences public trust and acceptance of the vaccine. The rational application of health behavior theory is a guarantee of effective public health information dissemination. However, little is known about the application of health behavior theory in online COVID-19 vaccine messages, especially from Chinese social media posts. OBJECTIVE This study aimed to understand the main topics and communication characteristics of hot articles related to COVID-19 vaccine on the WeChat platform and to assess the health behavior theory application with the aid of health belief model. METHODS A systematic search was conducted on the Chinese social media platform WeChat to identify COVID-19 vaccine-related articles. A coding scheme was established based on the health belief model, and the sample was managed and coded using NVivo12 (QSR International) to assess the application of health behavior theory. And the main topics of the articles were extracted through the Latent Dirichlet Allocation (LDA) algorithm. Finally, temporal analysis was used to explore trends in the evolution of themes and health belief structures in the articles. RESULTS A total of 757 articles were analyzed. Almost all (671/757, 89%) of the articles were not have an original logo. By topic modeling, five topics were identified which were vaccine development and effectiveness (267/757, 35%), disease infection and protection (197/757, 26%), vaccine safety and adverse reactions (52/757, 7%), vaccine access (136/757, 18%), and vaccination science popularization (105/757, 14%). All articles identified at least one structure in the extended health belief model, but only 29 articles included all of the structures. Descriptions of solutions to obstacles (585/757, 77%) and benefit (468/757, 62%) were the most emphasized components in all samples. Relatively few elements of susceptibility (208/757, 27%) and the least were descriptions of severity (135/757, 18%). The heat map visualization shows the change in health belief structure before and after vaccine entry into the market. CONCLUSIONS To the best of our knowledge, this is the first study to assess the structural expression of health beliefs in information related to the COVID-19 vaccine on the WeChat public platform based on a health belief model. The study also identified topics and communication characteristics before and after the market entry of vaccines. Our findings can inform customized education and communication strategies to promote vaccination not only in this pandemic but also in future pandemics.