tetano
Editor, Senior Moderator
Hum Vaccin Immunother
. 2025 Dec;21(1):2536898.
doi: 10.1080/21645515.2025.2536898. Epub 2025 Jul 30. Influencing factors and dynamic changes of COVID-19 vaccine hesitancy in China: From the perspective of machine learning analysis
Lei Li[SUP] 1 [/SUP], Hui Jing[SUP] 2 [/SUP], Yuqi Zhao[SUP] 3 [/SUP], Shenghua Wu[SUP] 1 [/SUP], Boyu Zhu[SUP] 4 [/SUP]
Affiliations
Exploring the influencing factors of COVID-19 vaccine hesitancy and summarizing countermeasures is of great significance for effectively addressing potential public health crises. Based on survey data from China, we employed a Gradient Boosting Decision Tree (GBDT) model and conducted SHAP interpretability analysis. The results show that in the primary series of COVID-19 vaccines, the important factors include social norms, vaccine knowledge, anticipated regret, age, vaccine safety, social responsibility, education level, religious belief, vaccine effectiveness, and perceived severity. While for booster shots, the important variables include age, vaccination experience, vaccine knowledge, vaccine effectiveness, gender, perceived severity, concerns about the epidemic, social norms, anticipated regret, and sense of social responsibility. The differences in the composition and significance of these core predictive factors suggest that COVID-19 vaccine hesitancy is dynamically evolving. This pattern of evolution is manifested as a shift in the decision - making basis from social norms to subjective experiences, in the focus of vaccines from safety - first to effectiveness - priority, and in the decision - making mechanism from emotion - dominated to cognition - driven. The research findings inspire us that when formulating vaccination strategies, multiple factors need to be comprehensively considered. Moreover, strategies should be adjusted in a timely manner according to changes in the vaccination stages to align with the shift in public concerns and decision - making mechanisms.
Keywords: COVID-19 vaccine; dynamic transformation; machine learning; vaccine hesitancy.
. 2025 Dec;21(1):2536898.
doi: 10.1080/21645515.2025.2536898. Epub 2025 Jul 30. Influencing factors and dynamic changes of COVID-19 vaccine hesitancy in China: From the perspective of machine learning analysis
Lei Li[SUP] 1 [/SUP], Hui Jing[SUP] 2 [/SUP], Yuqi Zhao[SUP] 3 [/SUP], Shenghua Wu[SUP] 1 [/SUP], Boyu Zhu[SUP] 4 [/SUP]
Affiliations
- PMID: 40735971
- DOI: 10.1080/21645515.2025.2536898
Exploring the influencing factors of COVID-19 vaccine hesitancy and summarizing countermeasures is of great significance for effectively addressing potential public health crises. Based on survey data from China, we employed a Gradient Boosting Decision Tree (GBDT) model and conducted SHAP interpretability analysis. The results show that in the primary series of COVID-19 vaccines, the important factors include social norms, vaccine knowledge, anticipated regret, age, vaccine safety, social responsibility, education level, religious belief, vaccine effectiveness, and perceived severity. While for booster shots, the important variables include age, vaccination experience, vaccine knowledge, vaccine effectiveness, gender, perceived severity, concerns about the epidemic, social norms, anticipated regret, and sense of social responsibility. The differences in the composition and significance of these core predictive factors suggest that COVID-19 vaccine hesitancy is dynamically evolving. This pattern of evolution is manifested as a shift in the decision - making basis from social norms to subjective experiences, in the focus of vaccines from safety - first to effectiveness - priority, and in the decision - making mechanism from emotion - dominated to cognition - driven. The research findings inspire us that when formulating vaccination strategies, multiple factors need to be comprehensively considered. Moreover, strategies should be adjusted in a timely manner according to changes in the vaccination stages to align with the shift in public concerns and decision - making mechanisms.
Keywords: COVID-19 vaccine; dynamic transformation; machine learning; vaccine hesitancy.