Sentiments and themes in the public posts related to influenza vaccines in China: Insights from Weibo data in 2023

Vaccination is a key measure for preventing influenza; however, the influenza vaccination rate remains very low in China. Understanding the public’s sentiment polarity in the topics related to influenza vaccines and monitoring the hot topics discussed by the public can not only help identify potential public opinion risks but also provide a scientific basis for designing targeted health communication strategies to promote influenza vaccination. This study aimed to identify the sentiment polarities and corresponding themes in the posts regarding influenza vaccines from the public on the Chinese social media, and to explore the evolutionary patterns of public sentiments and concerns over time. Within the framework of descriptive infodemiology, the original posts related to influenza vaccines published by the public on Sina Weibo were collected from January 1, 2023 to December 31, 2023. SnowNLP was utilized to perform sentiment analysis on the included posts and the Latent Dirichlet Allocation (LDA) was employed to identify themes within posts corresponding to each sentiment category. The monthly post volume, the monthly average sentiment score, and the distribution of posts across different sentiments and themes were calculated. A total of 12,045 original posts were included. The number of posts with positive sentiment (7,454; 61.88%) was higher than that with neutral or negative sentiment. With the exception of June, 2023 (0.60), the monthly average sentiment scores of posts remained consistently above 0.60. Five overarching themes were identified within the positive, neutral, and negative posts, respectively. Across all sentiment categories, the most concerned theme was discussions on the association between relevant infectious diseases/vaccines and influenza vaccines. This study characterizes the sentiment polarity of influenza vaccine-related posts by the Chinese public on Sina Weibo and identifies the corresponding themes of concern. During the study period, the discussions revolved around themes such as co-administration of vaccines, cost of the influenza vaccine, vaccination for children, and vaccine hesitancy. The algorithm used in this study classified a majority of included posts as positive, and post volume exhibited a monthly fluctuation pattern that roughly coincided temporally with influenza activity. This study suggests that health authorities should leverage sentiment analysis of social media posts to monitor public concerns, thereby providing a basis for developing targeted communication strategies.