Changes in HPV vaccine hesitancy and influencing factors among female university students in China: A national repeated study Authors: Peng Zhang, Xichen Tong, Xing Chen, Yijia Feng, Luying Zhang Published: 2026 ABSTRACT ABSTRACTHPV vaccine hesitancy remains a critical barrier to cervical cancer prevention in China despite a series of strategic policy initiatives. Our study conducted two large-scale online cross-sectional surveys in 2022 and 2025 among nationally representative samples of female university students, applying propensity score matching to enhance comparability. HPV vaccine hesitancy, assessed on a 0–10 scale (including graded hesitancy and refusal), increased significantly over time, with prevalence rising from 8.8% to 21.6%, and mean hesitancy scores also increasing. Although HPV knowledge improved and exposure to negative information declined, the perceived necessity of vaccination fell markedly, and trust eroded across several domains, including trust in health professionals’ advice and in both domestic and imported vaccines. Multivariable analyses identified perceived necessity, healthcare professional recommendation, and trust in professionals and domestic vaccines as protective factors, while prior self-payment for other vaccines was also associated with lower hesitancy. These determinants remained stable over time, underscoring the need for targeted strategies to counter rising hesitancy. Introduction Cervical cancer, primarily attributable to persistent infection with high-risk human papillomavirus (HPV) types, represents a major cause of cancer-related mortality among women worldwide, with a disproportionate burden in developing countries.1,2 In China, cervical cancer ranks second among female malignancies in both incidence (17.69 per 100,000) and mortality (5.52 per 100,000), accounting for over 11.9% of global cervical cancer deaths.3,4 Timely vaccination of age-appropriate women is therefore a critical strategy toward achieving the global goal of cervical cancer elimination.5 Long-term follow-up studies have demonstrated that, 5–8 y post-vaccination, the prevalence of HPV types 16 and 18 decreased by approximately 83% among females aged 13–19 y and by 66% among those aged 20–24 y.6 Nevertheless, in China, the HPV vaccination rate for women is less than 0.05%, who are aged 9–45 y.7 As one of the most accessible and preventable cohorts, female university students in China nonetheless exhibit persistently low HPV vaccination coverage, with uptake rates generally remaining below 15%.8–10 Aiming at expanding HPV vaccination coverage, Chinese authorities have implemented a series of strategic policy measures over the past decade.11 Since 2016, the “Healthy China Scheme” has incorporated cervical cancer prevention, including HPV vaccination, as a crucial element of national health planning.12 In 2020, the CFDA (now NMPA) lowered the minimum eligible vaccination age for the quadrivalent HPV vaccine to 9 y, facilitating earlier vaccination entry.13 In January 2023, the National Health Commission released the Action Plan for Accelerating the Elimination of Cervical Cancer (2023–2030), which promotes school-based vaccination with financial subsidies, and multiple pilot programs across ten provinces, such as in Jiangsu, Guangdong, Fujian, have provided free or subsidized HPV vaccines to school-aged girls.14 However, these efforts have yielded limited success, with vaccine hesitancy emerging as a key barrier to improving vaccination uptake.11 Vaccine hesitancy is defined as the delay in acceptance or refusal of vaccination despite the availability of vaccination services,15 which has been identified as an increasing challenge to promoting vaccination coverage and implementing immunization programs worldwide by the World Health Organization (WHO) Strategic Advisory Group of Experts (SAGE).16 In recent years, it has emerged as a leading contributor to suboptimal or declining HPV vaccination rates in several countries, including Japan, Brazil, and India.17–21 In China, the prevalence of HPV vaccine hesitancy among female university students has been reported to be as high as 37.57% in 202022 and 44.75% in 202223 in different studies, indicating a rising trend that poses a significant barrier to HPV vaccination uptake. Previous studies have demonstrated that HPV vaccine hesitancy is associated with multiple factors.24 According to the Vaccine Hesitancy Matrix proposed by the WHO SAGE on Immunization, the determinants of HPV vaccine hesitancy can be categorized into three domains: contextual influences, individual and group influences, and vaccine- or vaccination-specific factors.25 Consistent evidence among female university students indicates that insufficient knowledge about the disease and vaccine, low-risk perception, and distrust in vaccines or medical advice contribute to HPV vaccine hesitancy.26–29 Conversely, improvements in related knowledge, recommendations from healthcare professionals, and improved accessibility due to reduced vaccine costs significantly increase vaccination willingness among university students.30–32 However, most of these studies have been conducted as single cross-sectional surveys in limited regions. There remains a lack of nationally representative research in China that compares changes in HPV vaccine hesitancy among female university students over time, as well as studies exploring temporal trends and their joint influencing factors. Based on this context, this study aims to focus on HPV vaccine hesitancy among female university students in China and conduct repeated cross-sectional surveys nationwide in 2022 and 2025, to describe the temporal trends in vaccine hesitancy, examine whether significant changes in vaccine hesitancy occur over time, and identify its influencing factors classified by the Vaccine Hesitancy Matrix. The findings are expected to provide valuable evidence to update the trend of HPV hesitancy in China and inform targeted strategies for improving HPV vaccination coverage. Methods Study design and participants This study employed a repeated cross-sectional design based on two national online surveys conducted in June 2022 and January–February 2025, respectively. Both surveys in 2022 and 2025 utilized anonymous, self-administered questionnaires with consistent core items to ensure comparability (shown in supplementary materials Table S1). The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for cross-sectional studies. The study population is female university students in mainland China at both undergraduate and graduate levels, spanning a range of academic majors. A prior sample size was estimated by the following formula:(1)N=Z1−α/22×P×1−Pδ2 where Z equals 1.96 (when α is 0.05, the 95% confidence interval) and δ is 0.05. Based on findings from our previous study conducted in 2022, which reported the willingness of female college students in China to receive the HPV vaccination (P) was 84.4%.19 Assuming a 95% confidence level, 5% margin of error, and allowing for non-response and poor data quality (+50%), the minimum required sample size per region was calculated to be 304 participants for each year. Each participant was permitted to submit only one response, and electronic IP restrictions were applied to minimize duplication. Data collection and preprocessing This study employed a three-stage stratified sampling method to conduct a repeated cross-sectional survey. First, three regions (eastern, central, and western China) were selected according to economic development level and geographic location. The combined sample covered universities from eastern, central, and western regions, spanning multiple provinces and municipalities, including both economically developed and less-developed areas. Second, 16 universities were randomly selected within each region. The sampling frame was limited to public universities, as over 95% of Chinese university students attend public institutions. The randomly selected universities within each region covered a range of institution types (e.g., comprehensive, medical, normal, and engineering), reflecting the diversity of academic majors among female students. This diversity reflects the structure of China’s higher education system and allows for the inclusion of students with varied academic backgrounds. Third, within each selected university, one faculty or class was randomly sampled for questionnaire distribution, and students who received the survey link voluntarily chose whether to participate. Finally, based on the minimum sample size of 304 for a single region and the proportional distribution of university student populations across regions (4:3:3),33 at least 406 questionnaires were required from the eastern region (i.e., a minimum of 26 from each university), and at least 304 questionnaires were required from both the central and western regions (i.e., a minimum of 19 from each university). 2261 and 3600 participants were recruited in 2022 and 2025 through an online survey platform (Questionnaire Star), which was far more than the calculated sample size. The platform was configured to require a response to every question before submission; consequently, there were no item-level missing data in the collected questionnaires. Following data collected from 2022 as the baseline wave and from 2025 as the follow-up wave, the datasets were merged. Then, a multi-step data quality control process was implemented. Responses were excluded if they met any of the following criteria: (1) the questionnaire was completed in less than 180 seconds, indicating insufficient engagement; (2) internal logic checks revealed inconsistencies or implausible answers; or (3) the respondent self-reported a university location outside mainland China. To enhance comparability and control for potential confounding variables, propensity score matching (PSM) was performed after data collection. Based on 7 key demographic and academic characteristics, including age, education level, major, urban or rural residence, whether insured for basic medical insurance, economic level, and region, this study compared the results of different PSM methods. The 1:1 nearest-neighbor matching algorithm with replacement was the best, and it was applied to establish two well-balanced groups. To ensure the validity of the PSM, we tested both the overlap and balance assumptions by examining the bias and t-test results of the difference for all covariates between the two groups and visually inspecting the distribution of propensity scores before and after matching. Finally, 3321 participants were matched, allowing for robust comparison of HPV vaccine hesitancy and its associated factors over the three-year interval. Measures HPV vaccine hesitancy HPV vaccine hesitancy was measured using a two-step approach. First, participants were asked, “Are you willing to receive the HPV vaccine?” with three response options: willing, hesitant, or not willing. Those who selected “hesitant” were further prompted to indicate the degree of their hesitancy on a 9-point Likert-type scale ranging from 1 (minimally hesitant) to 9 (extremely hesitant). To standardize the hesitancy score across all respondents, participants who indicated ‘willing’ were assigned a score of 0, while those who indicated ‘not willing’ were assigned a score of 10. Crucially, ‘not willing’ (score 10) is treated as a distinct behavioral category—outright refusal—which is qualitatively different from even the highest level of hesitancy (score 9, ‘extremely hesitant’). An extremely hesitant individual remains theoretically open to persuasion or behavior change, whereas a not-willing individual has made a firm decision against vaccination. Thus, the 0–10 coding scheme comprises three conceptually distinct segments: willingness (0), graded hesitancy (1–9, with higher scores indicating greater intensity but still within the hesitancy spectrum), and refusal (10). Potential influencing factors In this study, the Vaccine Hesitancy Matrix developed by SAGE was applied as a conceptual framework.25 All variables were mapped into the three domains of the matrix based on their theoretical relevance and empirical evidence from prior studies. Contextual influences Contextual influences included demographic characteristics and exposure to negative information about vaccines. Sociodemographic variables included age; education level (college vs. postgraduate); academic major (medical vs. non-medical), indicating differences in health-related knowledge; type of residence (urban or rural) to capture environmental and resource-related disparities; type of medical insurance (insured vs. uninsured), serving as a proxy for healthcare accessibility; economic level, categorized into four groups based on monthly living expenses percentiles (≤25th, 25th–50th, 50th–75th, >75th percentile); and region location (eastern, central, or western China), reflecting national economic stratification. These sociodemographic variables were used as the PSM factors and controlled in the regression. In addition, participants were asked whether they had been exposed to negative information about HPV vaccines (yes or no), which may influence vaccine-related attitudes and decision-making. Individual and group influences Individual and group influences were evaluated in accordance with the Vaccine Hesitancy Matrix and included cognitive, attitudinal, and social variables. HPV-related knowledge was measured using a 5-item questionnaire, with total scores ranging from 0 to 5, where higher scores indicated greater knowledge. Perceived risk and perceived severity of HPV infection were assessed separately, ranging from 0 to 10, with higher scores indicating stronger perceptions of risk or severity. Perceived necessity of HPV vaccination was assessed through a single-item question, categorized into five levels from “Not at all necessary” to “Very necessary.” In addition, participants were asked whether they had received self-paid vaccinations other than HPV (yes or no), as a proxy for general vaccination acceptance. Lastly, HPV vaccine recommendations from social circles (including family, peers, or healthcare providers) were recorded as a binary variable (yes or no), reflecting the influence of interpersonal support on vaccine-related decisions. Vaccine/vaccination-specific issues Vaccine- and vaccination-specific issues were assessed based on participants’ perceptions and trust related to the HPV vaccine itself and its sources of recommendation. Perceived efficacy of the HPV vaccine was evaluated using a single item, with responses ranging from “Do not believe at all” to “Strongly believe.” Participants were also asked whether they had received HPV vaccine recommendations from health professionals (yes or no). Whether they trusted health professionals’ vaccination advice, the domestically produced HPV vaccine, and the imported HPV vaccine was assessed separately, and recorded as five levels from “Do not trust at all” to “Strongly trust.” No composite index was constructed; instead, individual variables from each domain were retained and entered simultaneously into the regression models to estimate their independent effects on HPV vaccine hesitancy. This approach enables a domain-based interpretation of determinants while preserving the granularity of individual factors. Statistical analysis Statistical description and statistical test were used to examine changes across two cross-sectional surveys. Continuous variables were summarized using means and standard deviations (SD), while categorical variables were presented as frequencies and percentages. Independent-sample t-tests were employed to assess differences in continuous variables, and chi-square tests were used for categorical variables. To examine temporal changes in HPV vaccine hesitancy and identify its potential determinants, regression analyses were conducted. Given that the dependent variable, vaccine hesitancy, is an ordinal categorical variable ranging from 0 to 10, an ordered logit model was employed. The ordered logit model, which assumes a latent continuous propensity partitioned by thresholds but does not require equal spacing between adjacent categories, was chosen because the dependent variable (0–10) represents ordered categories.34–36 Additional potential influencing factors were included in the model to account for their effects with odds ratios (OR) and corresponding 95% confidence intervals (CI) reported. Finally, exploratory interaction analyses were conducted by including cross-product terms (e.g., year × key determinants) in the multivariable ordered logistic models to assess whether the effects of influencing factors on HPV vaccine hesitancy differed between 2022 and 2025. Statistical significance was set at a two-sided α level of 0.05. All statistical analyses were performed using Stata MP version 17. The study protocol received ethical approval from the Institutional Review Board, School of Public Health, Fudan University (IRB#2023-TYSQ-02-261 and IRB#2022-08-0992). This study was conducted in accordance with the principles of the Declaration of Helsinki. Respondents could only begin the survey after providing electronic informed consent by clicking a consent button on the first page of the online questionnaire and did not receive compensation. This electronic consent procedure was approved by the Institutional Review Board. The study did not use biological samples from the survey subjects. Results Sample characteristics The characteristics of participants in 2022 and 2025 before and after PSM are shown in the supplementary materials Table S2. Before matching, significant differences were observed in age, education, major, insurance, and economic level. After 1:1 nearest neighbor matching, all covariates were well balanced, with standardized biases below 10% and no statistically significant differences between groups. Specifically, post‑matching standardized biases for all covariates ranged from −3.0% to 6.4% (see Table S2 in the supplementary materials), confirming that all values satisfied the conventional <10% balance threshold. Figure 1 displays the kernel density plots of the propensity scores before and after matching, indicating improved balance in covariate distribution post-matching. Figure 1.Kernel density plots of propensity scores before and after matching.Two multi-line kernel density graphs of propensity score at baseline and follow-up before and after matching. Kernel density plots of propensity scores before and after matching. The matched samples included 1,430 participants in 2022 and 1,891 in 2025. Table 1 presents the demographic and socioeconomic characteristics of participants from the 2022 and 2025 surveys after PSM without weighting. The mean age was similar between the two groups. An insignificantly lower proportion of participants attained postgraduate education and majored in medical-related fields in 2025. The distribution of residence (urban vs. rural) and the coverage of medical insurance remained relatively stable across both time points. Regarding self-reported economic status, there was a slight increase in participants from the upper quartiles in 2025, while those in the lower quartile decreased. The regional distribution remained relatively consistent between the two cohorts.Table 1.Characteristics of participants in 2022 and 2025 after PSM without weight.Characteristic2022 (n = 1,430)2025 (n = 1,891)Age, mean and SD 20.982.1620.932.19Education level, n and %College1,20984.55%1,70890.32% Postgraduate22115.45%1839.68%Major, n and %Medical65745.94%54929.03% Others77354.06%1,34270.97%Residence, n and %Rural75752.94%95250.34% Urban67347.06%93949.66%Medical insurance, n and %No886.15%1668.78% Yes1,34293.85%1,72591.22%Economic level, n and %≤25th17512.24%1698.94% 25th–50th91463.92%1,09457.85% 50th–75th28119.65%50126.49% >75th604.20%1276.72%Regional location, n and %East area76253.29%1,02254.05% Mid area34924.41%49826.34% West area31922.31%37119.62% Table 1.Characteristics of participants in 2022 and 2025 after PSM without weight.Characteristic2022 (n = 1,430)2025 (n = 1,891)Age, mean and SD 20.982.1620.932.19Education level, n and %College1,20984.55%1,70890.32% Postgraduate22115.45%1839.68%Major, n and %Medical65745.94%54929.03% Others77354.06%1,34270.97%Residence, n and %Rural75752.94%95250.34% Urban67347.06%93949.66%Medical insurance, n and %No886.15%1668.78% Yes1,34293.85%1,72591.22%Economic level, n and %≤25th17512.24%1698.94% 25th–50th91463.92%1,09457.85% 50th–75th28119.65%50126.49% >75th604.20%1276.72%Regional location, n and %East area76253.29%1,02254.05% Mid area34924.41%49826.34% West area31922.31%37119.62% Characteristics of participants in 2022 and 2025 after PSM without weight. Comparisons of HPV vaccine hesitancy and related factors Descriptive analyses in Table 2 revealed significant changes in HPV vaccine hesitancy and related factors between 2022 and 2025. The proportion of participants expressing hesitancy or refusal toward the HPV vaccine increased markedly. The HPV hesitancy significantly increased from 8.81% in 2022 to 21.63% in 2025, and the proportion of people willing to receive HPV vaccination decreased from 89.79% in 2022 to 72.98% in 2025. The degree of HPV vaccine hesitancy was coded on a scale from 0 to 10, with 0 representing willingness, scores between 1 and 9 indicating varying levels of hesitancy, and 10 representing refusal. The average degree of hesitancy increased significantly from 0.61 in 2022 to 1.67 in 2025.Table 2.Comparisons of HPV vaccine hesitancy and related factors in 2022 and 2025.Factors2022 (n = 1,430)2025 (n = 1,891)χ2 / tP valueVaccine hesitancy Hesitancy, n (%)Not willing20 (1.40%)102 (5.39%)167.98<.001 Hesitant126 (8.81%)409 (21.63%) Willing1284 (89.79%)1380 (72.98%) Hesitancy degree (0–10), mean (SD) 0.61 (1.95)1.67 (2.98)−11.61<.001Contextual influences Exposure to negative information about vaccines, n (%)No405 (28.32%)1036 (54.79%)232.16<.001 Yes1025 (71.68%)855 (45.21%) Individual and group influences HPV knowledge (0–5), mean (SD) 2.87 (1.48)3.34 (1.48)−8.95<.001Perceived risk of HPV (0–10), mean (SD) 4.84 (2.58)4.56 (2.27)3.39<.001Perceived severity of HPV (0–10), mean (SD) 7.96 (2.13)7.01 (2.11)12.76.244Perceived necessity of HPV vaccination, n (%)Not at all necessary1 (0.07%)26 (1.37%)354.87<.001 Not very necessary4 (0.28%)49 (2.59%) Neutral113 (7.90%)393 (20.78%) Somewhat necessary463 (32.38%)879 (46.48%) Very necessary849 (59.37%)544 (28.77%) Self-paid other vaccinations, n (%)No1081 (75.59%)1225 (64.78%)44.87<.001 Yes349 (24.41%)666 (35.22%) HPV vaccine recommended by social circles, n (%)No242 (16.92%)785 (41.51%)230.48<.001 Yes1188 (83.08%)1106 (58.49%) Vaccine/ vaccination-specific issues Perceived efficacy of vaccine, n (%)Do not believe at all0 (0.00%)20 (1.06%)210.29<.001 Slightly doubt9 (0.63%)100 (5.29%) Neutral162 (11.33%)498 (26.34%) Somewhat believe957 (66.92%)1005 (53.15%) Strongly believe302 (21.12%)268 (14.17%) HPV vaccine recommended by health professionals, n (%)No920 (64.34%)1026 (54.26%)34.09<.001 Yes510 (35.66%)865 (45.74%) Trust in health professionals’ vaccination advice, n (%)Do not trust at all1 (0.07%)9 (0.48%)95.79<.001 Slightly doubt9 (0.63%)75 (3.97%) Neutral271 (18.95%)550 (29.09%) Somewhat trust857 (59.93%)941 (49.76%) Strongly trust292 (20.42%)316 (16.71%) Trust in domestic HPV vaccine, n (%)Do not trust at all2 (0.14%)22 (1.16%)43.15<.001 Slightly doubt36 (2.52%)98 (5.18%) Neutral435 (30.42%)542 (28.66%) Somewhat trust768 (53.71%)897 (47.44%) Strongly trust189 (13.22%)332 (17.56%) Trust in imported HPV vaccine, n (%)Do not trust at all1 (0.07%)31 (1.64%)107.11<.001 Slightly doubt10 (0.70%)94 (4.97%) Neutral342 (23.92%)565 (29.88%) Somewhat trust843 (58.95%)866 (45.80%) Strongly trust234 (16.36%)335 (17.72%) Table 2.Comparisons of HPV vaccine hesitancy and related factors in 2022 and 2025.Factors2022 (n = 1,430)2025 (n = 1,891)χ2 / tP valueVaccine hesitancy Hesitancy, n (%)Not willing20 (1.40%)102 (5.39%)167.98<.001 Hesitant126 (8.81%)409 (21.63%) Willing1284 (89.79%)1380 (72.98%) Hesitancy degree (0–10), mean (SD) 0.61 (1.95)1.67 (2.98)−11.61<.001Contextual influences Exposure to negative information about vaccines, n (%)No405 (28.32%)1036 (54.79%)232.16<.001 Yes1025 (71.68%)855 (45.21%) Individual and group influences HPV knowledge (0–5), mean (SD) 2.87 (1.48)3.34 (1.48)−8.95<.001Perceived risk of HPV (0–10), mean (SD) 4.84 (2.58)4.56 (2.27)3.39<.001Perceived severity of HPV (0–10), mean (SD) 7.96 (2.13)7.01 (2.11)12.76.244Perceived necessity of HPV vaccination, n (%)Not at all necessary1 (0.07%)26 (1.37%)354.87<.001 Not very necessary4 (0.28%)49 (2.59%) Neutral113 (7.90%)393 (20.78%) Somewhat necessary463 (32.38%)879 (46.48%) Very necessary849 (59.37%)544 (28.77%) Self-paid other vaccinations, n (%)No1081 (75.59%)1225 (64.78%)44.87<.001 Yes349 (24.41%)666 (35.22%) HPV vaccine recommended by social circles, n (%)No242 (16.92%)785 (41.51%)230.48<.001 Yes1188 (83.08%)1106 (58.49%) Vaccine/ vaccination-specific issues Perceived efficacy of vaccine, n (%)Do not believe at all0 (0.00%)20 (1.06%)210.29<.001 Slightly doubt9 (0.63%)100 (5.29%) Neutral162 (11.33%)498 (26.34%) Somewhat believe957 (66.92%)1005 (53.15%) Strongly believe302 (21.12%)268 (14.17%) HPV vaccine recommended by health professionals, n (%)No920 (64.34%)1026 (54.26%)34.09<.001 Yes510 (35.66%)865 (45.74%) Trust in health professionals’ vaccination advice, n (%)Do not trust at all1 (0.07%)9 (0.48%)95.79<.001 Slightly doubt9 (0.63%)75 (3.97%) Neutral271 (18.95%)550 (29.09%) Somewhat trust857 (59.93%)941 (49.76%) Strongly trust292 (20.42%)316 (16.71%) Trust in domestic HPV vaccine, n (%)Do not trust at all2 (0.14%)22 (1.16%)43.15<.001 Slightly doubt36 (2.52%)98 (5.18%) Neutral435 (30.42%)542 (28.66%) Somewhat trust768 (53.71%)897 (47.44%) Strongly trust189 (13.22%)332 (17.56%) Trust in imported HPV vaccine, n (%)Do not trust at all1 (0.07%)31 (1.64%)107.11<.001 Slightly doubt10 (0.70%)94 (4.97%) Neutral342 (23.92%)565 (29.88%) Somewhat trust843 (58.95%)866 (45.80%) Strongly trust234 (16.36%)335 (17.72%) Comparisons of HPV vaccine hesitancy and related factors in 2022 and 2025. For contextual influences, a sharp decline was observed in exposure to negative vaccine information (71.68% in 2022 vs. 45.21% in 2025). Regarding individual and group influences, HPV knowledge scores increased significantly, while perceived severity of HPV infection declined. The perceived necessity of HPV vaccination also shifted, with fewer participants rating it as “very necessary” in 2025 (59.37% in 2022 vs. 28.77% in 2025). More participants reported having received other self-paid vaccines (24.41% vs. 35.22%, p < .001), while recommendations from social circles decreased substantially (83.08% to 58.49%, p < .001). Vaccine- and vaccination-specific influences also changed significantly. Belief in the efficacy of the HPV vaccine decreased, with a notable drop in somewhat and strongly believe responses (88.04% in 2022 vs. 67.32% in 2025). While recommendations from health professionals increased (35.66% to 45.74%), trust in their vaccination advice declined (combined “somewhat” and “strongly” trust: 80.35% in 2022 vs. 66.47% in 2025). Trust in both domestic and imported HPV vaccines also showed significant shifts, with higher proportions of neutrality or doubt reported in 2025 (combined “neutral,” “slightly doubt,” and “do not trust at all”: 33.08% in 2022 vs. 35.01% in 2025 for domestic vaccines; 24.69% in 2022 vs. 36.49% in 2025 for imported vaccines). The change of HPV vaccine hesitancy and influencing factors Based on the ordered logit regression results in Table 3, vaccine hesitancy was significantly higher in 2025 compared to 2022, with individuals being more than twice as likely to exhibit greater hesitancy (OR = 2.044, p = .001). Among contextual influences, none of the sociodemographic factors, including age, education, major, residence, medical insurance, economic level, or region, were significantly associated with vaccine hesitancy. Similarly, exposure to negative vaccine information was positively but non-significantly associated with hesitancy.Table 3.Ordered logit regression model of vaccine hesitancy and related factors.FactorsORSEzP-value95% CITime (ref. 2022)2.0440.4293.410.0011.355to3.083Contextual influences Age0.9230.058−1.290.1970.817to1.043Education level (ref. college)1.3910.6650.690.4900.545to3.552Major (ref. medical)0.7060.156−1.580.1150.458to1.088Residence (ref. rural)1.0650.2110.320.7490.723to1.570Medical insurance (ref. No)1.0920.4190.230.8180.515to2.317Economic level (ref. ≤25th)0.7820.132−1.460.1450.561to1.088Region (ref. East area)0.8960.128−0.770.4390.677to1.184Exposure to negative information about vaccines (ref. No)0.8000.130−1.370.1720.581to1.101Individual and group influences HPV knowledge0.9940.059−0.100.9190.884to1.117Perceived risk of HPV0.9860.039−0.360.7200.913to1.065Perceived severity of HPV1.0240.0440.550.5820.941to1.115Perceived necessity of HPV vaccination (ref. Not at all)0.2640.038−9.350.0000.200to0.349Self-paid other vaccinations (ref. No)0.6540.137−2.030.0430.434to0.986HPV vaccine recommended by social circles (ref. No)0.8330.168−0.910.3640.560to1.237Vaccine/vaccination-specific issues Perceived efficacy of vaccine (ref. Not at all)0.8820.173−0.640.5210.601to1.294HPV vaccine recommended by health professionals (ref. No)0.4960.102−3.400.0010.331to0.743Trust in health professionals’ vaccination advice (ref. Not at all)0.5410.089−3.730.0000.391to0.747Trust in domestic HPV vaccine (ref. Not at all)0.6520.096−2.900.0040.488to0.870Trust in imported HPV vaccine (ref. Not at all)1.1980.2380.910.3630.812to1.768*The dependent variable is an ordinal measure of HPV vaccine hesitancy ranging from 0 (willing) to 10 (refusal), with higher scores indicating greater hesitancy. An odds ratio (OR) greater than 1 indicates higher odds of being at a higher hesitancy level (i.e., increased hesitancy), whereas an OR less than 1 indicates lower odds of higher hesitancy (i.e., decreased hesitancy). Table 3.Ordered logit regression model of vaccine hesitancy and related factors.FactorsORSEzP-value95% CITime (ref. 2022)2.0440.4293.410.0011.355to3.083Contextual influences Age0.9230.058−1.290.1970.817to1.043Education level (ref. college)1.3910.6650.690.4900.545to3.552Major (ref. medical)0.7060.156−1.580.1150.458to1.088Residence (ref. rural)1.0650.2110.320.7490.723to1.570Medical insurance (ref. No)1.0920.4190.230.8180.515to2.317Economic level (ref. ≤25th)0.7820.132−1.460.1450.561to1.088Region (ref. East area)0.8960.128−0.770.4390.677to1.184Exposure to negative information about vaccines (ref. No)0.8000.130−1.370.1720.581to1.101Individual and group influences HPV knowledge0.9940.059−0.100.9190.884to1.117Perceived risk of HPV0.9860.039−0.360.7200.913to1.065Perceived severity of HPV1.0240.0440.550.5820.941to1.115Perceived necessity of HPV vaccination (ref. Not at all)0.2640.038−9.350.0000.200to0.349Self-paid other vaccinations (ref. No)0.6540.137−2.030.0430.434to0.986HPV vaccine recommended by social circles (ref. No)0.8330.168−0.910.3640.560to1.237Vaccine/vaccination-specific issues Perceived efficacy of vaccine (ref. Not at all)0.8820.173−0.640.5210.601to1.294HPV vaccine recommended by health professionals (ref. No)0.4960.102−3.400.0010.331to0.743Trust in health professionals’ vaccination advice (ref. Not at all)0.5410.089−3.730.0000.391to0.747Trust in domestic HPV vaccine (ref. Not at all)0.6520.096−2.900.0040.488to0.870Trust in imported HPV vaccine (ref. Not at all)1.1980.2380.910.3630.812to1.768*The dependent variable is an ordinal measure of HPV vaccine hesitancy ranging from 0 (willing) to 10 (refusal), with higher scores indicating greater hesitancy. An odds ratio (OR) greater than 1 indicates higher odds of being at a higher hesitancy level (i.e., increased hesitancy), whereas an OR less than 1 indicates lower odds of higher hesitancy (i.e., decreased hesitancy). Ordered logit regression model of vaccine hesitancy and related factors. *The dependent variable is an ordinal measure of HPV vaccine hesitancy ranging from 0 (willing) to 10 (refusal), with higher scores indicating greater hesitancy. An odds ratio (OR) greater than 1 indicates higher odds of being at a higher hesitancy level (i.e., increased hesitancy), whereas an OR less than 1 indicates lower odds of higher hesitancy (i.e., decreased hesitancy). In terms of individual and group influences, a stronger perceived necessity of HPV vaccination was significantly associated with lower odds of hesitancy (OR = 0.264, p < .001), while individuals who had previously paid out-of-pocket for other vaccinations were less likely to report hesitancy (OR = 0.654, p = .043). Other variables, including HPV knowledge, perceived risk, and perceived severity, were not significantly associated with hesitancy. Regarding vaccine- and vaccination-specific factors, receiving an HPV vaccine recommendation from health professionals was strongly associated with reduced hesitancy (OR = 0.496, p = .001). Greater trust in health professionals’ vaccination advice (OR = 0.541, p < .001) and in domestic HPV vaccines (OR = 0.652, p = .004) were both significantly associated with lower hesitancy. In contrast, trust in imported vaccines and belief in vaccine efficacy were not significantly associated with hesitancy in the multivariable model. Analysis of interaction effects between influencing factors and time The ordered logistic regression analysis examining the interaction effects of time and key influencing factors on HPV vaccine hesitancy is presented in Table 4. The insignificant interactions indicate that while the perceived necessity of HPV vaccination, HPV vaccine recommended by health professionals, trust in health professionals’ vaccination advice, and trust in the domestic HPV vaccine are key protective factors against HPV vaccine hesitancy, their influence remained stable over time, with no robust moderation detected. Perceived necessity of HPV vaccination was strongly associated with reduced hesitancy. Regarding self-paid other vaccinations, neither the main effect nor the interaction with time reached statistical significance; however, time independently showed a significant positive association with hesitancy. The recommendation of HPV vaccine by health professionals was significantly associated with decreased hesitancy, while its interaction with time was insignificant. Trust in health professionals’ vaccination advice was inversely related to hesitancy without a significant interaction effect with time. Similarly, trust in the domestic HPV vaccine showed a significant protective effect against hesitancy, with no significant interaction with time.Table 4.Interaction effects between time and key influencing factors on HPV vaccine hesitancy.Key FactorsORSEzP-value95% CIPerceived necessity of HPV vaccination*Time1.4850.3671.601.1100.914to2.413Perceived necessity of HPV vaccination0.2080.046−6.989.0000.134to0.323Time0.4590.449−0.792.4270.067to3.129Self-paid other vaccinations*Time0.5850.266−1.181.2390.239to1.428Self-paid other vaccinations0.9330.385−0.171.8660.415to2.096Time2.2540.5333.439.0011.418to3.584HPV vaccine recommended by health professionals*Time2.0900.9541.622.1060.855to5.111HPV vaccine recommended by health professionals0.2980.1173.079.0020.138to0.644Time1.7850.4122.512.0121.135to2.807Trust in health professionals’ vaccination advice*Time1.2050.3770.589.5520.652to2.226Trust in health professionals’ vaccination advice0.4910.133−2.631.0090.289to0.834Time1.0701.2030.061.9520.118to9.700Trust in domestic HPV vaccine*Time1.1070.3520.321.7490.594to2.063Trust in domestic HPV vaccine0.6130.152−1.967.0480.377to0.996Time1.4371.6100.323.7460.160to12.914 Table 4.Interaction effects between time and key influencing factors on HPV vaccine hesitancy.Key FactorsORSEzP-value95% CIPerceived necessity of HPV vaccination*Time1.4850.3671.601.1100.914to2.413Perceived necessity of HPV vaccination0.2080.046−6.989.0000.134to0.323Time0.4590.449−0.792.4270.067to3.129Self-paid other vaccinations*Time0.5850.266−1.181.2390.239to1.428Self-paid other vaccinations0.9330.385−0.171.8660.415to2.096Time2.2540.5333.439.0011.418to3.584HPV vaccine recommended by health professionals*Time2.0900.9541.622.1060.855to5.111HPV vaccine recommended by health professionals0.2980.1173.079.0020.138to0.644Time1.7850.4122.512.0121.135to2.807Trust in health professionals’ vaccination advice*Time1.2050.3770.589.5520.652to2.226Trust in health professionals’ vaccination advice0.4910.133−2.631.0090.289to0.834Time1.0701.2030.061.9520.118to9.700Trust in domestic HPV vaccine*Time1.1070.3520.321.7490.594to2.063Trust in domestic HPV vaccine0.6130.152−1.967.0480.377to0.996Time1.4371.6100.323.7460.160to12.914 Interaction effects between time and key influencing factors on HPV vaccine hesitancy. Discussion In this repeated cross-sectional study of HPV vaccine hesitancy among Chinese female university students, propensity score matched samples from 2022 to 2025 revealed a substantive and statistically significant increase in HPV vaccine hesitancy over time. The proportion of participants reporting hesitancy rose from 8.81% in 2022 to 21.63% in 2025, and the mean hesitancy score also significantly increased. In the ordered logistic model, after controlling for other factors, participants in 2025 had more than double the odds of greater hesitancy relative to those in 2022 (OR = 2.044, p = .001). This trend was supported by other studies on female university students in China, with HPV vaccine hesitancy reported to be 37.57% in 202022 and 44.75% in 2022.23 Compared with these earlier single-wave studies in China, our study adds temporal evidence on increased hesitancy, employing a nationwide, repeated cross-sectional design and rigorous propensity score matching to ensure sample comparability. The observed rise in hesitancy could be explained by the notable changes in several related factors, including declines in perceived vaccination necessity and in multiple domains of trust. Despite improvements in HPV knowledge and a reduction in exposure to negative vaccine information between 2022 and 2025, perceived risk and perceived severity of HPV declined. Consequently, the perceived necessity of HPV vaccination fell markedly, with the proportion rating “very necessary” declining from 59.4% to 28.8%. Simultaneously, belief in vaccine efficacy weakened, with the proportion of participants who indicated either “somewhat believe” or “strongly believe” decreasing by 20.72% in 3 y. Although the proportion of participants reporting that health professionals recommended the HPV vaccine increased, the confidence in advice and vaccine declined: combined “somewhat/strongly” trust in health-professional advice decreased by 13.88% points, in domestic HPV vaccines fell by 1.93% points, and in imported HPV vaccines declined by 11.79% points. Taken together, these patterns suggest that simple increases in factual knowledge or lower exposure to negative content do not necessarily translate into greater acceptance37; however, perceived necessity and institutional trust appear to have a stronger influence on vaccination hesitancy.38 Other studies have reported that greater knowledge can paradoxically correlate with higher vaccine hesitancy.38,39 For instance, Li et al. found that increased HPV knowledge was associated with vaccine hesitancy among university students, suggesting that information overload may adversely affect decision-making processes.38 Conversely, Licata et al. observed that being enrolled in medical and life sciences major—a proxy for higher HPV-related knowledge—was positively associated with vaccination willingness among unvaccinated students.39 Taken together, these findings indicate that the relationship between knowledge and hesitancy is context-dependent and may be mediated by factors such as information complexity and trust. In other words, a reduced perceived risk of HPV, growing doubts about vaccine efficacy, and an erosion of trust likely offset gains in knowledge and the decline in exposure to negative information, thereby contributing to the net increase in vaccine hesitancy.40 Beyond these individual-level factors, broader contextual changes may have also played a role. The 2023 Action Plan for Accelerating the Elimination of Cervical Cancer promoted school-based HPV vaccination programs and intensified media coverage. While these policy efforts improved vaccine accessibility, they also exposed the public to diverse—and sometimes conflicting—information about HPV vaccines, which may have inadvertently heightened hesitancy. In particular, social media platforms such as Douyin have become important channels for HPV vaccine information, but fragmented and unverified content on these platforms may have further complicated public understanding and eroded trust.41 Furthermore, the post-pandemic context, characterized by general vaccine fatigue and eroded trust in immunization programs, could have further shaped the observed trends.42 These explanations were further proved by the results of multivariable ordered logit regression. Our results identified several consistent protective factors against higher hesitancy, which also echo and extend prior research. A stronger perceived necessity of HPV vaccination was associated with substantially lower odds of greater hesitancy. By emphasizing the preventative benefits of the HPV vaccine, health communicators can improve perceived efficacy to motivate vaccination uptake among this population.40 Receipt of an HPV vaccine recommendation from health professionals, greater trust in health professionals’ vaccination advice, and trust in domestic HPV vaccines were independently associated with lower hesitancy. Previous research has similarly demonstrated that university students place considerable weight on recommendations from healthcare providers, with over 70% indicating willingness to receive HPV vaccine if advised by a doctor or nurse.43 These results align with systematic reviews and national surveys among Chinese students, indicating the trust in the efficacy of vaccines, health professional endorsement, and institutional trust as central drivers of vaccination acceptance.26–28 Notably, participants reporting prior out-of-pocket uptake of other vaccines showed lower odds of hesitancy, suggesting an important role of financial experience in shaping vaccination attitudes. It may reflect greater prior engagement with preventive services, higher vaccination acceptance in this subgroup, or residual selection bias.44 In the Chinese context, where HPV vaccination is largely self-financed and not yet fully covered by the national immunization program, financial cost remains a critical structural determinant of vaccine uptake.31,32 Therefore, prior self-paid vaccination behavior may act as a proxy for both economic capacity and acceptance of paying for preventive services. The results proved that the out-of-pocket cost of HPV vaccines may influence both perceived affordability and willingness to vaccinate. These findings highlight the importance of reducing financial barriers, for example, through subsidies, insurance coverage, or integration into public health programs, as a potential strategy to mitigate HPV vaccine hesitancy. Unlike previous research findings, in our study, sociodemographic characteristics and knowledge-related factors were not significant determinants of HPV vaccine hesitancy. The intention rate of HPV vaccination among female university students varies among different populations with different sociodemographic characteristics (e.g., major and region),28,45 while in our adjusted model, contextual sociodemographic variables, including age, education, major, residence, insurance, economic level, and region, were not associated with hesitancy. As for cognitive measures, higher perceived risk and perceived severity of HPV were proven to be negatively associated with hesitancy in a former study,28 but our results didn’t support these factors having a significant influence. Consistent with the finding that trust and risk perception mediate the relationship between knowledge and HPV vaccine hesitancy, this pattern indicates that cognitive awareness alone may be insufficient to reduce hesitancy when trust or perceived importance remains low.19 After identifying perceived importance, self-payment for other vaccines, healthcare professionals’ recommendations and trustworthiness, and trust in domestic vaccines as key influencing factors, we conducted an exploratory interaction analysis to examine whether the influence of key predictors changed between 2022 and 2025. The results show no robust, statistically significant moderation effects. Thus, while population-level hesitancy increased, the relative importance of those key determinants appears largely stable, implying that interventions targeting these factors should retain relevance over time. International evidence on COVID-19 vaccine hesitancy supports that the key determinants or modifiable risk factors (e.g., trust, risk perception, and preference for natural immunity) remained stable over time.46,47 Our results have some policy implications. First, reframing risk perception and vaccination necessity through tailored public education is critical. Although HPV knowledge improved, declining perceptions of risk and severity eroded the perceived necessity of vaccination. Educational strategies should therefore go beyond fact dissemination to emphasize HPV as a relevant and preventable threat, employing narratives, peer-led initiatives, and social media interventions to highlight the consequences of non-vaccination. Second, strengthening institutional and professional trust is essential. Recommendations from healthcare providers and confidence in their advice were among the strongest protective factors against hesitancy, underscoring the need to embed HPV vaccination counseling into routine student health services and to train providers in persuasive, evidence-based communication that can restore confidence in both domestic and imported vaccines. Third, expanding structural accessibility by integrating HPV vaccination into preventive service packages can normalize uptake. Our results suggest that prior out-of-pocket vaccination experiences reduce hesitancy, indicating that embedding HPV vaccination within student health insurance or preventive care bundles could reduce financial and logistical barriers while promoting HPV vaccination as part of routine preventive care. Several limitations should be mentioned. First, the repeated cross-sectional design precludes causal inference about individual-level change, and the sample was not strictly nationally representative. The multicenter stratified sampling design and subsequent PSM were applied to improve the representativeness and comparability of the study population, although the sampling frame was limited to public universities (which enroll > 95% of female students in China). Private universities, accounting for <5% of female enrollment, were excluded; thus, findings are most directly generalizable to female students in the public higher education system. Students in private institutions may differ in socioeconomic background or health literacy, and caution is warranted when generalizing findings to that subgroup. Therefore, caution should be exercised when generalizing the findings to populations beyond female university students or to broader national populations. In the future, longitudinal national data would be required to examine long-term change at the within-person level. Second, online recruitment may introduce selection bias toward students with higher digital access. In our study, we minimized such issues through nationwide sampling, and given the high internet penetration among Chinese university students, selection bias is likely to be largely controlled. Additionally, all key measures (knowledge, trust, perceived necessity, etc.) were self‑reported, which is subject to social desirability and recall bias. Although the anonymous online format may have mitigated social desirability effects, residual bias cannot be ruled out. Finally, policy or supply changes occurring between surveys, such as local pilot programs, may have influenced attitudes in ways not fully captured by our covariates. Future work should prioritize longitudinal cohort studies to map local policy implications and causal mechanisms, and qualitative or mixed-methods research to unpack the drivers of rising hesitancy. Conclusion In summary, HPV vaccine hesitancy (including graded hesitancy and refusal) among Chinese female university students increased substantially between 2022 and 2025. Perceived vaccination necessity, health professional recommendation, and trust in professional advice and domestic vaccines were consistently protective against higher hesitancy, and these relationships remained broadly stable over time. To arrest and reverse the rising hesitancy trend among Chinese female university students, policymakers and health practitioners should combine tailored education to reframe risk and vaccination necessity, strengthen trust in healthcare providers, and expand structural accessibility. Supplementary Material 260107Supplementary.docx