Multi-platform social media use and trust, health beliefs, and HPV vaccination willingness among Chinese university students

Background Human papillomavirus (HPV) vaccination coverage among Chinese university students remains low despite effective vaccines and national elimination targets. How multi-channel social media shapes vaccination decisions in China's digital environment is not well understood. This study investigated determinants of HPV vaccination willingness among university students using the Health Belief Model (HBM), focusing on media exposure, trust, and gender-specific barriers. Methods A cross-sectional survey was conducted among 250 university students in Guangdong, China. The survey measured sociodemographic factors, HPV knowledge, HBM constructs (perceived susceptibility, severity, benefits, barriers), perceived vaccine safety, browsing and active search frequency on five social media platforms, trust in different information sources, and vaccination willingness. Univariate and multivariate logistic regression identified independent predictors. Gender-stratified analyses examined specific barriers. Results Of the 250 students, 78.0% (195/250) expressed willingness to get the HPV vaccine. Greater HPV vaccination willingness was associated with being female (aOR = 10.24, 95% CI: 4.05–25.92), age 18-26 years (aOR = 4.83, 95% CI: 1.41–16.50), higher HPV knowledge (aOR = 1.18, 95% CI: 1.03–1.35), greater perceived vaccine safety (aOR = 4.50, 95% CI: 2.01–10.09), greater perceived susceptibility (aOR = 4.48, 95% CI: 2.08–9.64), and greater social media trust (aOR = 2.12, 95% CI: 1.20–3.75). Passive browsing and active information searching were not retained in the final multivariable model. Social media platforms (WeChat, Xiaohongshu, Weibo) were the most frequently used information channels but received the lowest trust ratings; healthcare providers were trusted most. Willing students reported higher browsing frequencies on WeChat, Xiaohongshu, and Weibo than unwilling students ( p < 0.05), but these differences did not persist in the multivariate model. Male and female students had different barrier profiles: infertility myths were more common among males (14.7 vs. 4.5%, p < 0.001); worry about menstrual changes was reported by 20.0% of males and 27.7% of females ( p = 0.094). Conclusions Social media was frequently used but received lower average trust than healthcare providers; social media use frequency was not independently associated with willingness. Instead of just putting more content on social media, efforts should focus on building knowledge, using trusted sources such as healthcare providers, and on correcting gender-specific misunderstandings.