Design and participants

The study is embedded in a two-arm randomized controlled trial (RCT) that evaluated the effectiveness of a pay-it-forward intervention on HPV vaccine uptake among adolescent girls in Chengdu, China. For the purposes of this study, the term “caregivers” is used to refer to adults who are responsible for making healthcare decisions for adolescents, including but not limited to biological parents. This broader term encompasses individuals such as guardians, grandparents, or other family members involved in caregiving roles. The detailed methodology and design have been previously described in detail in the protocol publication [24]. The RCT was conducted at four community health service centers in Chengdu, China between July 6, 2022 and June 9, 2023. The four sites were selected to reflect varying socio-economic contexts based on the relative level of individual disposable income in 2021. This approach ensures a broad representation of different socio-economic environments within Chengdu. The selected sites are as follows: Site A: Located in the most developed urban area, this site benefits from higher healthcare access and represents a higher socio-economic status with better public health infrastructure. Site B: Situated in a higher middle-income suburban area, this site experiences relatively good healthcare access but is influenced by suburban socio-economic factors. Site C: This site represents a lower middle-income suburban area with more limited healthcare resources and socio-economic challenges. Site D: The least developed area in terms of infrastructure, healthcare access, and socio-economic conditions. This site faces the greatest challenges related to economic factors and healthcare delivery. Eligibility criteria for participation included (1) girls aged 15–18 years; (2) residents of the communities served by the selected health centers; (3) no prior history of HPV vaccination; (4) no history of vaccine allergy; and (5) caregivers’ consent to participate in the study. Girls who were under 15 or over 18 years of age, as well as those who did not consent to complete the questionnaire, were excluded.

We obtained a de-identified contact list of adolescent girls aged 15–18 years from the residential committee through community health center staff. This list included all girls within this age range who have registered their names and contacts information at the community health center. We then performed simple random sampling to select eligible girls by generating a random number for each individual using Microsoft Excel. These individuals were sorted in ascending order based on their assigned random numbers. Recruitment proceeded by contacting individuals at the top of the list until the desired sample size was reached at each study site.

Upon arrival at the study site, participants were randomly assigned to either the standard-of-care arm or the pay-it-forward arm using sealed opaque envelopes to ensure allocation concealment. Each envelope contained a random numerical ID corresponding to the assigned arm (standard-of-care or pay-it-forward). The allocated intervention was concealed from both study participants and research assistants onsite prior to assignment until the envelope was opened, thereby maintaining blinding throughout the assignment process.

During the intervention process, all participants received an educational pamphlet about cervical cancer and HPV vaccines. In addition, participants in the pay-it-forward arm received personalized postcard messages containing additional information about HPV vaccines and well-wishes from others, designed based on caregiver preferences gathered during focus group discussions. Caregivers who chose to vaccinate their daughters received a subsidy of CNY 330(approximately USD 46.21) towards the HPV vaccine. The total cost for a full three-dose HPV vaccine series in the local market ranged from approximately CNY 990 to CNY 3,900 (USD 139–546), depending on the type and manufacturer. After vaccination, they had the opportunity to donate and/or write postcard messages to encourage other to get vaccinated early. Participants in the standard-of-care arm who chose to receive the HPV vaccine had to pay the standard market price for the HPV vaccine without any subsidies or additional messaging. All participants completed an online questionnaire at study site (Appendix A1) after the intervention. The sample size was estimated based on a pilot study conducted in Chengdu from January to February 2022. In this pilot study, vaccine uptake was observed to be 98% in the pay-it-forward arm and 82% in the standard-of-care arm, with a 16% difference. However, due to the low HPV vaccination rate (<10%) in the target population, we conservatively estimate 20% uptake in the standard-of-care and 36% in the pay-it-forward arm for the trial. To detect this difference with 80% power, 120 participants are required per arm. Considering a 10% non-response or dropout rate based on results from the pilot, 133 participants per arm are needed. To allow for secondary and subgroup analyses, the desired sample size will be increased by 20%, resulting in a total of 320 participants. A total of 321 participants were enrolled in the study, with 160 in the standard-of-care arm and 161 in the pay-it-forward arm.

Data sources

Data for this study were derived from the randomized controlled trial, including participant questionnaires and vaccination records maintained by the study staff. The primary data source was a structured online questionnaire completed by caregivers after providing informed consent. The questionnaire encompassed sections on sociodemographic information, willingness to vaccinate within a month, vaccine confidence, vaccine hesitancy, and intention to delay vaccination if 4v- and/or 9v- vaccines were unavailable. Adolescents’ vaccination records were retrieved from an electronic vaccination registry system within three months after the pay-it-forward intervention. Only participants with complete questionnaire data were included.

Variable measurements

Guiding Conceptual framework.

The study was guided by Anderson’s Health Service Utilization Behavioral Model, which has been widely used to understand factors influencing health service utilization, including sexual health services [25,26]. The model posits that health service use is determined by a combination of predisposing, enabling, need, and environmental factors [27]. Predisposing factors include demographic characteristics, social structure variables, beliefs, attitudes, and knowledge about health services [28]. Enabling factors include financial resources, access to services, and availability of healthcare facilities [28]. Need factors include perceived or evaluated health conditions requiring medical attention [28]. Environmental factors include external influences such as social networks and exposure to information [28]. An adapted version of the model was used to select independent variables and interpret the results (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Adapted Andersen’s Health Service Utilization Behavioral Model of potential factors associated with HPV vaccine uptake.

https://doi.org/10.1371/journal.pone.0324260.g001

Statistical analysis

Descriptive statistics, including means and frequencies, were calculated for demographic variables and HPV vaccine uptake. Univariable logistic regression analyses were performed to assess the association between each independent variable and HPV vaccine uptake, expressing the results as odds ratios (ORs) with 95% confidence intervals (CIs). Variables with a p-value < 0.20 in the univariable analyses were considered for inclusion in the multivariable logistic regression model. The multivariable regression model was developed using a backward stepwise strategy, adjusting for potential confounders such as age and gender. This method was chosen for its systematic elimination of non-significant variables and ability to simplify the model. However, it carries risks of overfitting, exclusion of important interactions, and dependence on the initial variable set. The logistic regression models in this study assumed the independence of observations, linearity of the logit, and the absence of multicollinearity. To ensure the independence of observations, participants were randomized at the individual level, and no repeated measures were taken from the same individual. Multicollinearity was assessed using the variance inflation factor (VIF), and variables with a VIF > 5 were excluded. A p-value < 0.05 was considered statistically significant. All analyses were conducted using IBM SPSS Statistics version 26.0.

Ethical approval

The Ethics Committee of West China Fourth Hospital and West China School of Public Health approved the study (Gwll2021057/ Gwll2023125). Informed written consent was obtained from all participants before participation. This study was registered on the Chinese Clinical Trial Registry (ChiCTR2200055542).

Caregiver characteristics and HPV vaccination status of adolescents

All 321 caregivers were sent the questionnaire, with a 100% completion rate in both the pay-it-forward arm and the standard-of-care arm. Therefore, all 321 caregivers were included in the final analysis. The mean age of caregivers was 44.7 years (SD = 8.6). The majority of caregivers (80.1%, 157/321) were female. Regarding occupation, 43.3% of caregivers were employed, and 32.1% were farmers. Most (91.9%) caregivers had education level above primary school and the majority (80.4%) of caregivers were married. Of all participating in the study, 25.9% (82/321) of adolescent girls received their first HPV vaccine dose within three months, with 17.5% in the standard-of-care arm and 34.2% in the pay-it-forward arm. (Table 1)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Caregiver characteristics and HPV vaccination status of adolescents (N = 321).

https://doi.org/10.1371/journal.pone.0324260.t001

Factors influencing caregivers’ action to vaccinate their daughters

The univariable logistic regression results are shown in the Table 2. Caregivers who had previously heard of HPV or the HPV vaccine were more likely to vaccinate their daughters. Caregivers with intention to delay vaccination also had higher vaccination rates for their daughters. However, caregivers with an annual household income exceeding USD 1,120 were less likely to have their adolescents vaccinated. Adolescents whose caregivers were farmers, unemployed, or retired had higher vaccination rates compared to those whose caregivers were employed. Additionally, adolescents who received the pay-it-forward intervention had a higher HPV vaccine uptake than those did not received any intervention. Conversely, girls whose peers had been vaccinated against HPV were less likely to be vaccinated.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Associated factors for caregivers’ action to vaccinate adolescent girls against HPV (N = 321).

https://doi.org/10.1371/journal.pone.0324260.t002

Variables with a P-value less than 0.2 after adjusting for age and gender of caregivers were included in the multivariable model. The multivariable logistic regression analyses revealed that knowledge of HPV vaccine, intention to delay vaccination, caregiver occupation, the pay-it-forward and prior refusal of the vaccine remained statistically significantly associated with HPV vaccine uptake. Caregivers who had not heard of the HPV vaccine prior to participating in our program were more likely to vaccinate their daughters than those who were already aware of it (OR=3.01,95%CI:1.27–7.14, P = 0.013). Adolescents whose caregivers did not intend to delay vaccination were nearly eight times higher vaccine uptake than those whose caregivers did (OR=8.26,95%CI:4.36–15.67, P < 0.001). Adolescents who received the pay-it-forward intervention had twice the HPV vaccine uptake compared to those who did not (OR=2.22,95%CI:1.19–4.15, P = 0.013). Furthermore, adolescents whose caregivers who were unemployed or retired had higher vaccination rates than those whose caregivers were employed (OR=3.97,95%CI:1.81–8.72, P = 0.001). Lastly, adolescents who had previously refused the HPV vaccine for some reason (other than allergy or illness) had lower vaccine uptake (OR=0.23,95%CI:0.64–0.81, P = 0.022). (Table 2)

Impact of the pay-it-forward intervention

Our findings revealed that caregivers participating in the pay-it-forward intervention had a substantially higher vaccination rate (34.2%) compared to those receiving standard care (17.5%). This suggests that the pay-it-forward model is an effective approach to increase HPV vaccine uptake, corroborating previous studies demonstrating the efficacy of such models in improving health service utilization [22,23,31]. The effect of the pay-it-forward intervention was quantified by an odds ratio (OR) of 2.22 (95% CI: 1.19–4.15), meaning that caregivers in the pay-it-forward group were 2.22 times more likely to vaccinate their daughters compared to those in the standard-of-care group. While this effect is significant, it is less pronounced than some other factors, such as the intention to delay vaccination (OR = 8.26), which emerges as a key barrier to vaccination. Several mechanisms may explain this success. First, the intervention provided tailored educational materials, such as specially designed postcards. These materials may have addressed specific knowledge gaps among caregivers, enabling more informed decision-making. Second, the financial subsidies inherent in the pay-it-forward model likely reduced economic barriers. Prior research shown that lowering vaccine costs increases parental willingness to vaccinate [32]. Third, the altruistic component of the model may have fostered a sense of community engagement and trust, positively influencing health behaviors [33,34]. Interestingly, the standard-of-care arm exhibited a higher uptake (17.5%) than previous observational studies (4.66%) [7]. This highlights the impact of active engagement strategies.

These results underscore the potential of the pay-it-forward intervention to expand HPV vaccine coverage, particularly among populations not currently served by government-funded programs. While some regions in China have initiated free or subsidized HPV vaccination initiatives[35], many adolescents remain unvaccinated due to geographic and socioeconomic disparities. Implementing the pay-it-forward model could bridge this gap by combining financial assistance with community engagement. However, due to the strong intention to delay vaccination (OR = 8.26), future research should dissect the individual components of the intervention and focus on strategies to address this barrier in particular. Additionally, our observation that unemployed or retired caregivers were nearly 4 times more likely to vaccinate their daughters compared to employed caregivers, and that those previously unaware of the HPV vaccine were 3 times more likely to do so. These results may indicate a higher level of trust in the healthcare system among these population and the importance of informational campaigns.

Barriers to HPV vaccine uptake

Caregivers who intended to delay HPV vaccination were more than 8 times as likely to postpone vaccination compared to those who did not intend to delay. This factor represents a significant barrier in the Chinese context and should be prioritized in interventions. Despite a high willingness to vaccinate within a month among caregivers (98.8%) (317/321) (S1 Fig), actual vaccine uptake was markedly lower (25.9%) (83/321). This highlights a significant gap between willingness and action. This disparity may be attributed to vaccine availability issues, particularly the preference for 4v- and 9v- HPV vaccines [36,37]. Although the domestically produced 2v-HPV vaccines were available, many caregivers opted to wait for the higher-valency vaccines. These vaccines often face supply constraints when our study was conducted. The demand for higher-valency vaccines, coupled with limited availability, may lead to vaccination delays or refusals [38,39]. Furthermore, caregivers who had previously refused the HPV vaccine were less likely to vaccinate their daughters, possibly due to persistent misconceptions or lack of confidence in the vaccine’s safety and efficacy [40]. This finding is particularly concerning because those with prior refusals (OR = 0.23) are less likely to change their position. Addressing vaccine hesitancy and building confidence in vaccine safety may require more intensive efforts. Misinformation propagated through social media regarding vaccine cost and availability may exacerbate these concerns [37]. Both 2v- and 9v-HPV vaccines offer protection against HPV types 16 and 18. These types are responsible for the majority of cervical cancer cases [41,42]^. Therefore, it is imperative to emphasize the benefits of early vaccination with any available vaccine. Enhancing public education on vaccine efficacy and addressing supply chain issues for higher-valency vaccines are critical steps toward improving uptake. In addition to misinformation and vaccine availability, cultural stigma significantly impacts HPV vaccine uptake [43,44]. In China, cultural concerns about adolescent sexuality and the HPV vaccine’s association with sexual behavior contribute to vaccine hesitancy. Similar stigma is seen in other Asian countries, like Hong Kong and South Korea, though its intensity varies. In India, seeking the HPV vaccine can be stigmatized due to concerns over family honor and social prestige, further fueling vaccine hesitancy [44]. This stigma can deter caregivers from vaccinating adolescents due to concerns about societal judgment or misinterpretation of their intentions [44]. To address this, public health campaigns should emphasize the vaccine’s role in preventing cervical cancer while incorporating culturally sensitive messaging while incorporating culturally sensitive messaging. These efforts can help to reduce stigma, promote acceptance, and improve vaccination uptake.

Global relevance and application of the findings

HPV vaccination rates remain low in many countries, particularly in low-income regions where vaccine access is limited and healthcare systems may lack the infrastructure to promote the vaccine effectively [45]. The significant role of economic barriers and lack of knowledge about the vaccine mirrors challenges observed in other countries, such as India and several sub-Saharan African nations [46,47]. Although this study was conducted in China, the findings are highly relevant to other regions facing similar barriers to HPV vaccination uptake. Countries in Sub-Saharan Africa, Southeast Asia, and rural areas of Latin America, where financial and informational barriers to vaccination are prevalent, could benefit from strategies that combine financial support with community-driven health education initiatives. For example, in Southeast Asia, where cultural stigmas around adolescent sexuality are prevalent, the “pay-it-forward” model could be adapted by involving local leaders, such as religious figures or community health workers, to help reshape societal perceptions around the HPV vaccine. Future research could explore how the “pay-it-forward” model might be adapted to improve HPV vaccination rates in diverse settings. This would involve tailoring the intervention to suit local contexts, cultural nuances, and available resources.

Limitations

Our study has several limitations. As a study embedded within a randomized controlled trial, the sample size was relatively small and limited to an urban sample, which may restrict the generalizability of the findings to rural areas or other regions with different socio-cultural contexts. Additionally, the intervention at one site coincided with the COVID-19 pandemic, which may have reduced participants’ willingness to visit clinics due to restricted access and increased anxiety about public spaces. The new guidelines imposed on healthcare settings may have led to longer waiting times, restricted clinic hours, and limited patient capacity, which further dissuaded caregivers from seeking vaccination services for their children. These factors likely contributed to lower vaccination rates at Site D compared to other sites. Another limitation is the self-reported nature of the data, which may be affected by recall bias and social desirability bias, as caregivers might not accurately remember past vaccination events or may report more positive vaccine confidence to align with social expectations. Frequent stockouts of the 4v- and 9v- HPV vaccines may have also led to underestimation of vaccine uptake due to this structural barrier. Nonetheless, alternative HPV vaccine options were available across all sites, ensuring access to vaccination for participants considering other formulations.