Introduction

The COVID-19 pandemic in 2020-21 led to significant excess mortality globally, particularly impacting older populations1. Age has been identified as the predominant risk factor for severe outcomes from COVID-19, with risks escalating notably in older age groups. Data from the National Vital Statistics System (NVSS) reveal that individuals over the age of 65 are at the greatest risk, with more than 81% of COVID-19 related deaths occurring in this demographic, and mortality rates for those over 65 being 97 times higher than those between 18 and 29 years of age2,3. Moreover, the risk is not limited to older adults, as individuals with certain pre-existing conditions and those over 50 also face increased dangers, with substantial risk elevation beyond the age of 654,5.

The susceptibility of the elderly to severe illness and increased mortality rates from COVID-19 highlights the critical role of vaccination as a protective measure. Yet, vaccination willingness among older adults is influenced by a myriad of factors including demographic traits, socioeconomic status, vaccine-related knowledge, confidence in the vaccine′s efficacy, and exposure to vaccine-related communications6,7. Factors such as income levels and residency in socioeconomically disadvantaged or rural areas significantly correlate with vaccine acceptance among older adults8,9. Concerns regarding safety, a presumable risk of adverse effects, and doubts about vaccine efficacy are related to willingness to vaccinate10. Increasing vaccine acceptance in older adults will require knowledge of the factors that positively influence their vaccination intentions11.

Vaccine hesitancy and insufficient knowledge about the drivers of vaccine acceptance pose major challenges to pandemic preparedness, particularly within vulnerable and specific sociodemographic groups. Similar findings have been found in several studies, with people of lower educational attainment, lower employment status, and residing in rural areas being more vaccine-hesitant12,13.

The Health Belief Model (HBM) has been found to be significant in predicting acceptance of the vaccine14. Theoretical models of health beliefs are essential tools for understanding the factors behind decision-making by assessing what motivates people to adopt health-related behavior. HBM is one of the most widely used models for examining the relationship between health behavior and the use of health services15. This strategy seeks to explain and predict preventive health behavior in terms of certain belief patterns. Specifically, HBM suggests that an individuals’ engagement in health-promoting behavior can be explained by their beliefs about health problems, perceived benefits of action and barriers to action, and self-efficacy. Meanwhile, the cues to action trigger them to adopt health-promoting behaviors15.

This study aimed to explore potential factors that may influence their vaccination intentions among older adults, including socioeconomic characteristics, health beliefs, and behavioral cues. Therefore, based on these factors in the elderly, we can propose better vaccination strategies to increase their vaccination rates to better protect this COVID-19 susceptible group. Finding vaccination strategies that are more acceptable to older adults to increase vaccine acceptance and expand vaccine coverage to improve survival in older adults in the face of a prolonged global epidemic of COVID-19.

Methods

Participants

This was a community-based cross-sectional study conducted among people aged 60 years and older in Sichuan Province located in southwest China. A multi-stage stratified sampling method was used in this study. Briefly, 7 districts or counties were all selected in Luzhou City, and three neighborhoods or townships were randomly selected in each district or township, resulting in a total of 21 towns or communities were randomly selected.

Based on previous studies indicating a 92.3% willingness to receive the COVID-19 vaccine among hypertensive patients, along with a projected non-response rate of 20%, a confidence level of 95% (corresponding to a z-score of 1.96), the design effect of deff = 2 and a margin of allowable error of d = 0.05, we calculated that a minimum sample size of 273 participants was necessary for this study. Finally, 1527 participants were invited to participate in the study, and 1305 of them were included in this study, resulting in a response rate of 85.5%. Inclusion criteria were: (1) aged 60 years and above; (2) able to communicate effectively and voluntarily participate in the survey. Exclusion criteria were: (1) have any contraindications to Covid-19 vaccination; (2) unable to understand the content of the questionnaire; (3) patients with high blood pressure or diabetes that cannot be controlled by medication.

Procedure

Research design

This study was grounded in the Health Belief Model (HBM) to explore the determinants of vaccination willingness among older adults. The independent variables in this research included knowledge about vaccine, Health Beliefs and Cues to Action. The knowledge about vaccines primarily encompasses the following items included in the questionnaire: type and dose of vaccine, vaccine contraindications, and adverse reactions after vaccination. Health Beliefs included awareness of infection, awareness of severity and awareness of vaccine effectiveness. Cues to Action included recommendation from family members, recommendation from people around, recommendation from medical staff and official publicity of the health system.

Mediated effects analysis

The framework for mediated effects analysis utilized COVID-19-related knowledge as an exposure factor, with health beliefs and vaccination cues serving as mediators.

The Exposure-Mediator-Outcome Model is a framework used in statistical and causal analysis to explore how an exposure influences an outcome, emphasizing the mediating role of variables along the causal path. It postulates indirect effects through mediators and possible direct effects. Exposure is the influencing factor, Mediator bridges the gap, transmitting the exposure’s effect, while Outcome is explained by the model. Assessing the impact of exposure factors on outcomes through mediators by calculating the mediation proportion (PM) of indirect effects mediated by mediators. The formula for the calculation of PM (mediation proportion):

$$\:PM=\frac{NIE}{NIE+NDE}$$

(Abbreviations: NIE natural indirect effects, NDE natural direct effects.)

This approach aimed to assess their impact on the vaccination intentions of older adults through causal analysis, and the mediation analysis framework is illustrated in Fig. 1.

Fig. 1
Fig. 1
Full size image

Mediation analysis framework based on the health belief model.

Questionnaire development

The questionnaire was devised using the Health Belief Model and insights from previous studies on vaccination intentions15,16,17. The finalized questionnaire was divided into three sections, each rigorously evaluated for validity and reliability:

Validity Analysis: Expert evaluations from infectious disease and vaccinology specialists were solicited to score and revise the questionnaire. The overall content validity index (CVI) was 0.99, with individual section figures exceeding 0.78. Reliability Analysis: The overall Cronbach’s α for the Health Belief scale was 0.77, with each subsection scoring above 0.60, confirming the reliability of the instrument.

Mediating effects exploration

This segment focused on examining the influence of background variables, health beliefs about COVID-19, and cues to vaccination action on the vaccination willingness of older adults. The study aimed to determine the presence of mediating effects between the knowledge about the COVID-19 vaccine and the intention to vaccinate, thereby identifying the mediators influencing older adults’ willingness to receive the vaccine.

Data analysis

Data were analyzed using statistical software R 4.2.2 and IBM SPSS 26.0. The analytical methods included: (1) Comparative Analysis: Chi-square tests were employed to discern differences between groups. (2) Associative Analysis: Multivariable logistic regression analyzed the associations between study factors and outcomes. (3) Mediating Effects Analysis: The study used an Exposure-Mediator-Outcome model, applying the Monte Carlo method for mediator effect analysis. This involved fitting a parametric mediator model, simulating model parameters based on the sampling distribution, and calculating causal mediation effects through simulated mediator and outcome values within a counterfactual framework18,19,20.

Results

General characteristics

1305 participants were included in this study, with 50.8% being male and 49.2% being female. The results showed that 92.9% (1212 people) of participates were willing to be vaccinated against COVID-19, while 7.1% (93 people) preferred the opposite. Vaccination willingness differs by age among seniors (χ²=25.07, P < 0.01), with the 75 + group showing lower acceptance. Rural seniors had a higher acceptance rate (94.8%) than urban ones (89.6%, χ²=12.30, P < 0.01). Married seniors had a 94.8% acceptance rate, 14.2% higher than unmarried/divorced/widowed participates. Occupation also influenced acceptance (χ²=10.08, P < 0.05). See Table 1.

Table 1 Baseline characteristics of the study participants.
Full size table

Multiple logistic regression analysis

The results showed that participants who lived in urban areas (compared to rural areas, OR = 0.47, 95% CI: 0.26–0.83), were in the ≥ 75 age group (compared to 60–64 age group OR = 0.41, 95% CI: 0.18–0.98), were unmarried/divorced/widowed (compared to married, OR = 0.36, 95% CI: 0.20–0.64), or had occupations in commerce/services (compared to compared to farming, OR = 0.33, 95% CI: 0.13–0.83) had lower willingness to receive the COVID-19 vaccine. Additionally, older individuals with higher self-assessed scores of COVID-19 vaccine efficacy (OR = 1.23, 95% CI: 1.10–1.39), more cues to action regarding COVID-19 vaccination (OR = 1.31, 95% CI: 1.21–1.42), and higher scores of knowledge about COVID-19 vaccine (OR = 1.06, 95% CI: 1.02–1.10) were associated with a higher willingness to receive the vaccine.

Mediating effects

The participants’ knowledge about the Covid-19 vaccine was observed as the exposure factor in this study, meanwhile, Covid-19 Health Beliefs and Clues to vaccination action were studied as potential mediators to explore their mediating role between the participants’ knowledge and vaccination intention, in the control of sociodemographic characteristics (gender, age, urban or rural area, education, marital status, occupation, average monthly income). Based on the Health Belief Model, our mediation analysis framework was shown in Fig. 1.

As shown in Table 2, reliability and validity test showed factor loadings of each independent variables in this model were strong (Knowledge: 0.790, 0.679 and 0.760, Health beliefs: 0.756, 0.818 and 0.695; Cues to Action: 0.931, 0.942, 0.953 and 0.955), which could effectively measure the latent variable. The Cronbach’s alpha coefficients of the observation variables ranged from 0.621 to 0.975.

Table 2 Reliability and validity analysis in mediation model based on HBM (Health Belief Model).
Full size table

After screening the exposure factor (Knowledge about COVID-19), potential mediator (COVID-19 Health Beliefs or Cues to Vaccination Action), and outcome (Intention to COVID-19 Vaccination) one by one in our established model, the results of the single mediator analysis indicated that two following pathways had significant mediating effects (p-values<0.05 both in the natural indirect effect and the natural direct effect) in this study: (Table 3)

Table 3 Single mediating effect in the association between knowledge of COVID-19 vaccine and vaccination intention.
Full size table

(1) ″Knowledge about COVID-19 vaccine → Health Beliefs → Vaccination intention″ pathway.

The natural direct effect ranged from 0.012 to 0.016 while the natural indirect effect ranged from 0.001 (Awareness of infection/severity) to 0.003 (Awareness of vaccine effectiveness), and the mediator ratio ranged from 4% (Awareness of infection) to 18% (Awareness of vaccine effectiveness).

(2) ″Knowledge about COVID-19 vaccine → Cues to Action → COVID-19 Vaccination intention″ pathway.

The natural direct effect ranged from 0.012 to 0.016 while the natural indirect effect ranged from 0.002 to 0.003, and the mediator ratio ranged from 9% (recommendation from family members) to 20% (recommendation from medical staff).

Discussion

This is a population-based survey involving 1305 elderly individuals from southwestern China, which studies how sociodemographic factors and COVID-19-related knowledge influence their COVID-19 vaccination intention. The results revealed that the willingness to get vaccinated against COVID-19 among the elderly was 92.9%, surpassing figures from previous surveys among older adults in China and Thailand21,22, and aligning closely with those in the United States23. Perceived effectiveness of the COVID-19 vaccine, Knowledge related to COVID-19, and Cues to action for vaccination were all positively correlated with the intention to get vaccinated. Among these factors, Cues to action had the most significant impact on the willingness of the elderly to receive the COVID-19 vaccine. Participants aged ≥ 75 years were less inclined towards vaccination24, with increasing age inversely affecting their willingness, aligning with several studies from China25,26. Older adults often face higher risks due to underlying medical conditions and potential contraindications to vaccination27,28, and concerns about vaccine impacts on existing health issues may drive conservative vaccination approaches. Occupation and residence also significantly influenced vaccination willingness, with varying acceptance rates among different professional groups and geographic locations29,30.

Furthermore, this study explored the mediating effect of elderly people’s vaccine-related knowledge on their willingness to be vaccinated against COVID-19 based on the counterfactual framework, revealing that COVID-19 health beliefs and vaccination action cues played a mediating role, and identified two mediation pathways: (1) From Knowledge of COVID-19 to Health Beliefs to Vaccination Intention: Here, awareness of vaccine effectiveness played a pivotal role, suggesting that enhancing confidence in the vaccine’s efficacy could substantially increase willingness to vaccinate26,31. (2) From Knowledge of COVID-19 to Cues of vaccination action to Vaccination intention: Recommendations from medical staff and acquaintances proved crucial, indicating that policy efforts should focus on elevating vaccine awareness among healthcare workers and through interpersonal networks to boost vaccination intentions32.

Among these pathways, the “Cues to Vaccination Action” pathway mediates a stronger mediating effect, exerting a greater influence on the vaccination willingness of the elderly. Therefore, strengthening these related action cues can significantly enhance the willingness to be vaccinated, which is crucial for protecting the elderly during the long-term global pandemic.

Implications for public health

Employing the Health Belief Model, this study analyzed the mediating effects between vaccine-related knowledge and vaccination willingness among older adults. Identifying characteristics (such as sociodemographic factors, knowledge related to COVID-19, and cues to action for vaccination) associated with their vaccine acceptance can help tailor public health strategies to increase vaccine acceptance and coverage.

Strengths and limitations

The main strength of this study lies in its focus on the elderly population, examining the impact of sociodemographic factors and COVID-19 health beliefs on their willingness to be vaccinated against COVID-19. Furthermore, it employs a mediation effect framework based on causal inference to dissect the underlying mechanisms through which COVID-19-related knowledge influences vaccine acceptance among the elderly.

However, several limitations existed in our study. First, as this study is a cross-sectional study, its ability to infer causality is limited. Second, the analysis primarily focused on single mediation effects without considering potential interactions between mediators. Future research could incorporate multiple mediation analyses to better control for confounders between health beliefs and vaccination cues, refining our understanding of factors driving vaccination willingness among older adults.

Conclusion

The intention to vaccinate against COVID-19 among older adults varies significantly based on socioeconomic and health belief factors. Health beliefs and cues to action not only as direct contributors but also as crucial mediators in the pathway from knowledge to vaccination intention among older individuals. These discoveries can enhance our comprehension of how sociodemographic variables and health beliefs influence the willingness of older adults to receive the COVID-19 vaccine, ultimately contributing to better safeguarding the elderly during the prolonged global prevalence of COVID-19.