Heterogeneity in the psychological antecedents of influenza vaccination among Chinese older adults: A 5C model approach

ABSTRACT

Seasonal influenza remains a significant public health concern among older adults, who face an increased risk of hospitalization and mortality. Despite the availability of free vaccination programs in many regions of China, influenza vaccination coverage among older adults remains low. Psychological antecedents play a crucial role in shaping vaccination decisions, yet few studies have explored how these factors interact with sociodemographic characteristics. A cross-sectional survey was conducted among 13,363 community-dwelling adults aged ≥60 y in six major Chinese cities between December 2024 and January 2025. The 5C scale – Confidence, Collective Responsibility, Complacency, Constraints, and Calculation – was used to assess psychological determinants of influenza vaccination. Logistic regression models, likelihood ratio tests (LRTs), and stratified analyses were performed to examine the main and interaction effects between psychological and sociodemographic factors. Overall, 34.05% of participants reported receiving an influenza vaccine within the past year. Calculation was positively associated with vaccination uptake (OR = 1.075, 95% CI: 1.057–1.093), while Complacency (OR = 0.658, 95% CI: 0.643–0.673) and Constraints (OR = 0.849, 95% CI: 0.831–0.867) were negatively associated. Significant interaction effects were observed between 5C subscales and education, income, and self-care ability, indicating that the influence of psychological factors varied across subpopulations. The psychological drivers of influenza vaccination among Chinese older adults are heterogeneous across sociodemographic groups. Precision public health strategies are essential: while high-SES individuals may benefit from data-driven information, low-income and low-literacy groups require simplified, emotionally resonant communication focused on personal protection rather than broad altruism. Furthermore, interventions should aim to reduce perceived constraints by streamlining vaccination processes and offering personalized risk-benefit assessments for older adults with chronic conditions.

Introduction

Seasonal influenza infection can lead to hospitalization or death, contributing to 290,000 to 650,000 respiratory deaths globally each year.¹ Older adults constitute a high-risk population for influenza, with increased susceptibility and a greater likelihood of developing severe complications, requiring intensive care, and experiencing higher mortality.^2–4 Influenza vaccination is an important strategy for preventing influenza infection, particularly among older adults.^5–8 However, vaccination coverage among this population varies substantially across countries. For example, in 2023, the vaccination rate among adults aged ≥65 y was only 9.3% in Poland, compared with 84.8% in Korea, according to available estimates.⁹ Recent evidence identifies both demographic variables and psychological dimensions as key predictors of vaccination decisions.^10–13 Previous studies have also found high heterogeneity in the determinants of vaccination, with substantial differences across subpopulations.^14,15

Vaccine hesitancy refers to delay in acceptance or refusal of safe vaccines despite the availability of vaccination services.¹⁶ In 2019, the World Health Organization (WHO) included vaccine hesitancy on its list of the top ten global health threats for the first time.¹⁷ The 5C scale provides a novel tool to monitor the psychological antecedents of vaccination.^18–20 It consists of five components: Confidence, Collective Responsibility, Complacency, Constraints, and Calculation. Confidence is defined as trust in the effectiveness and safety of vaccines.¹⁸ Collective Responsibility refers to the willingness to vaccinate oneself to protect others through herd immunity.^18,21 Individuals who lack Confidence and Collective Responsibility tend to hold negative attitudes toward vaccination. Complacency is defined as the perception of low personal risk from infectious diseases, leading individuals to view vaccination as unnecessary.¹⁸ Constraints refers to the structural or perceived psychological barriers individuals encounter when seeking vaccination.¹⁸ Calculation refers to individuals’ engagement in extensive information searching and deliberation.^18,22,23 The 5C scale thus facilitates the analysis of vaccination behavior from a psychological perspective.

While previous studies have typically treated the 5C scale as a set of independent predictors of vaccination behavior, few have examined their interaction with demographic or socioeconomic factors. It remains unclear whether psychological drivers operate uniformly across all older adults, or if their influence varies by socioeconomic status and health condition. Identifying high-risk subpopulations through the 5C framework – and tailoring interventions accordingly – is a critical step toward improving influenza vaccination coverage among older adults. Furthermore, exploring the determinants of vaccination among elderly populations with different characteristics is essential for improving influenza vaccination rates.

The primary aim of this study is to determine whether and how the effect of each 5C subscales on influenza vaccination differ across subgroups defined by sociodemographic factors. Specifically, the study aims to: (1) assess the main effects of each 5C subscale on influenza vaccination differ across subgroups; (2) test for interaction effects between 5C subscales and demographic characteristics using likelihood ratio tests (LRT) and visualize effect heterogeneity through marginal effects plots; (3) further explore these relationships through stratified modeling; and (4) provide actionable insights for tailoring public health interventions for older adults.

Method

Study design and data collection

Study participants

This cross-sectional study was executed between December 2024 and January 2025 across six major Chinese cities: Beijing, Hangzhou, Qingdao, Shenzhen, Chongqing, and Chengdu. Within each city, five to eight Community Health Service Centers (CHSCs) were selected via simple random sampling from the local CHSCs. A minimum recruitment quota of 300 participants was established for each selected center. The study population comprised community-dwelling residents aged 60 y or older who visited these facilities for routine medical care or vaccination services.

Participant recruitment was conducted consecutively in waiting areas by trained investigators. Data were collected through face-to-face interviews utilizing an electronic questionnaire platform. To guarantee data completeness, the system was configured to prohibit submission unless all items were answered, thereby eliminating missing values. Of the 13,754 individuals approached, 13,363 valid responses were retained following a rigorous screening process. The effective response rate was 97.2%.

Sample size Calculation

Using the standard cross-sectional study formula, we calculated a required sample size of 8850 based on a 4.16% vaccination rate among Chinese older adult (60 y and above) reported by the China CDC in 2023 (α = 0.05; δ = 0.00416). We inflated this number by 35% to address potential non-response, resulting in a final recruitement goal of 11,948 participants.

$$\mathit{n}=\frac{Z_{1-\alpha /2}^2\ast \mathit{p}(1-\mathit{p})}{{\delta }^2}$$

Measures

Influenza vaccination

Influenza vaccination status was measured using a single question: “Have you ever received the influenza vaccine in the past one year?” with response options of “Yes” or “No.”

Sociodemographic characteristics

Sociodemographic data were collected from each participant, including gender (male, female), age (60–69, 70–79 and ≥80 y), educational level (primary and below, middle school, high school, college and above), monthly income (≤2500 CNY, 2501–5000 CNY, 5001–7500 CNY, and >7500 CNY), chronic illness (with chronic illness or no chronic illness), and self-care ability (fully independent or not fully independent).

Vaccine hesitancy (5C scale)

Influenza vaccine hesitancy was measured using the 5C scale, which assesses five subscales: Confidence, Collective Responsibility, Complacency, Constraints, and Calculation.

Confidence was measured using the following items: (1) “I think the influenza vaccine is safe.” (2) “I trust physicians’ recommendation for influenza vaccination to benefit my health.” (3) “I trust the influenza vaccination provided by the government to protect older adults.”

Collective Responsibility was measured using: (1) “I trust the influenza vaccination provided by the government to protect older adults.” (2) “I think getting vaccinated against influenza can reduce the spread of diseases among the population.” (3) “I am willing to protect populations with low immunity by getting influenza vaccinated.”

Complacency was assessed by: (1) “The probability of getting diseases is low, so I do not need to be vaccinated.” (2) “Even if I get infected, I can resist it, so I don’t need the influenza vaccine.” (3) “I believe influenza vaccines are not the only effective way to prevent diseases.”

Constraints were measured using: (1) “It was easy to obtain information about influenza vaccines.” (2) “It is easy and quick for me to get vaccinated.” (3) “I know the vaccination process at CHSCs.” Additional items included the exact vaccination site, service hours, and available appointment channels. It is important to note that this subscale specifically operationalizes constraints as perceived barriers regarding the accessibility and convenience of the vaccination process, rather than purely objective structural limitations.

Calculation was measured using: (1) “I carefully considered the efficacy and risks of the influenza vaccines.” (2) “I seriously sought information about influenza vaccines before vaccination.” (3) “I proactively consulted a physician for vaccination advice.”

All items were rated on 5-point Likert scales, and every subscale score ranged from 3 to 15. Reliability and validity analyses were as follows: Confidence (Cronbach’s α = 0.868); Collective Responsibility (α = 0.959); Complacency (α = 0.626); Constraints (α = 0.715); and Calculation (α = 0.678). The KMO value of the 5C scale was 0.886, and Bartlett’s test of sphericity yielded χ² = 55681.6 (p < .001), indicating good construct validity.

Statistical analysis

All analyses were performed using R version 4.4.3. Initially, univariate analyses and logistic regression were used to assess the independent associations of 5C subscales and demographic characteristics (gender, age, educational level, monthly income, chronic illness, and self-care ability) with influenza vaccination behavior.

A series of logistic regression models were then fitted, including both main effects and interaction terms between each 5C subscale and the demographic variables. The statistical significance of interaction effects was assessed using the Likelihood Ratio Test (LRT). Significant interactions (p < .05) were visualized using interaction effect plots and further explored through stratified modeling, where separate logistic regressions were estimated within each subgroup defined by the modifying variable. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) for each 5C subscale were computed to quantify and visualize differential associations across subpopulations.

Ethics approval and consent

This study was performed in strict accordance with the ethical principles outlined in the Declaration of Helsinki. The research protocol was reviewed and approved by the Institutional Review Board of the Chinese Academy of Medical Sciences & Peking Union Medical College, the official ethics oversight body for human subjects’ research (approval number: CAMS&PUMC-IEC-2025-002).

Prior to participation, all individuals were provided with comprehensive information about the scientific nature of the study. Strict confidentiality protocols were implemented throughout the research process, and all data were anonymized to ensure participant privacy. The collected data were used exclusively for research purposes in accordance with the approved protocol.

Result

Demographics characteristics

The participants’ demographic characteristics are presented in Table 1. Among the 13,363 valid surveys, 45.26% were male and 54.74% were female. More than half of participants were aged 60–69 y (57.46%). Overall, about one-third of older adults did not have chronic illness (33.36%), while two-thirds reported having at least one chronic illness (66.64%). Additionally, most older adults were fully independent in self-care (84.70%).

Table 1.

Demographics characteristics and vaccination rate.

Variable		N (%)	Vaccination N (%)
Gender	Male	6048 (45.26)	2110 (34.89)
	Female	7315 (54.74)	2440 (33.36)
Age	60–69	7678 (57.46)	2640 (34.38)
	70–79	4302 (32.19)	2456 (33.84)
	≥80	1383 (10.35)	454 (32.83)
Education Level	Primary and Below	3720 (27.84)	1043 (28.04)
	Middle School	3992 (29.87)	1383 (34.64)
	High School	3461 (25.90)	1256 (36.29)
	College and Above	2190 (16.39)	868 (39.63)
Monthly Income (CNY)*	≤2500	3850 (28.81)	1001 (26.00)
	2501–5000	5829 (43.62)	2153 (36.94)
	5001–7500	2457 (18.39)	937 (38.14)
	>7500	1227 (9.18)	459 (37.41)
Chronic Illness	No Chronic Illness	4458 (33.36)	1470 (32.97)
	With Chronic Illness	8905 (66.64)	3080 (34.59)
Self-Care Ability	Not Fully Independent	2045 (15.30)	739 (36.14)
	Fully Independent	11318 (84.70)	3811 (33.67)

*CNY: Chinese yuan; US dollars (USD) have an average exchange rate of CNY 7.19 per USD in December 2024.²⁴.

The overall influenza vaccination rate was 34.05%. Vaccination rates were highest among college-educated older adults (39.63%) and those with monthly incomes >7500 CNY (37.41%), and lowest among those with primary education (28.04%) or monthly incomes ≤2500 CNY (26.00%). Older adults with chronic illnesses (34.59%) or limited self-care ability (36.14%) were more likely to be vaccinated than their healthier or more independent counterparts.

5C scores among all participant and by characteristics

The mean (SD) scores for each 5C subscale were: Confidence (11.38 ± 2.28), Collective Responsibility (11.61 ± 2.61), Calculation (7.04 ± 2.66), Complacency (8.55 ± 2.32), and Constraint (6.71 ± 2.50). Confidence and Collective Responsibility scores were higher than those of other subscales across all subgroups. Women scored higher than men on most dimensions. Having a chronic illness was associated with higher Confidence and Collective responsibility, but lower Complacency. See Figure 1 for details.

Figure 1.

Mean 5C scale scores (±95% CI) across sociodemographic subgroups.

Multivariable logistic analysis of sociodemographic variables and 5C scale on influenza vaccination

The results of the multivariable logistic regression are presented in Table 2. Regarding the 5C scale, Calculation (OR = 1.075, 95%CI: 1.057–1.093), Complacency (OR = 0.658, 95%CI: 0.643–0.673), and Constraint (OR = 0.849, 95%CI: 0.831–0.867) showed the strongest associations with influenza vaccination.

Table 2.

Multivariable logistic regression of vaccination.

Variable		Estimate	Std	z-value	p-value	OR	95%CI
5C Scale	Confidence	−0.002	0.014	−0.135	.892	0.998	(0.971, 1.026)
	Collective Responsibility	0.004	0.012	0.367	.713	1.004	(0.981, 1.029)
	Calculation	0.072	0.009	8.287	<.001	1.075	(1.057, 1.093)
	Complacency	−0.418	0.012	−35.886	<.001	0.658	(0.643, 0.673)
	Constraint	−0.164	0.011	−15.111	<.001	0.849	(0.831, 0.867)
Gender	Male	Ref.
	Female	−0.000	0.042	−0.011	.991	1.000	(0.920, 1.086)
Age	60–69	Ref.
	70–79	0.075	0.047	1.604	.109	1.078	(0.984, 1.181)
	≥80	−0.032	0.075	−0.432	.666	0.968	(0.836, 1.121)
Education Level	Primary and Below	Ref.
	Middle School	0.240	0.061	3.960	<.001	1.271	(1.129, 1.431)
	High School	0.413	0.067	6.157	<.001	1.512	(1.325, 1.724)
	College and Above	0.512	0.083	6.139	<.001	1.669	(1.417, 1.965)
Monthly Income (CNY)	≤2500	Ref.
	2501–5000	0.286	0.057	5.020	<.001	1.331	(1.191, 1.489)
	5001–7500	0.394	0.075	5.262	<.001	1.483	(1.281, 1.718)
	>7500	0.347	0.097	3.568	<.001	1.415	(1.169, 1.711)
Chronic Illness	No Chronic Illness	Ref.
	With Chronic Illness	0.071	0.045	1.557	.120	1.073	(0.982, 1.173)
Self-Care Ability	Not Fully Independent	Ref.
	Fully Independent	−0.216	0.060	−3.586	<.001	0.806	(0.716, 0.907)

Older adults with higher Calculation scores were more likely to have been vaccinated against influenza, which contrast with general theoretical assumptions about Calculation. Those with higher Complacency and Constraint were less likely to have been vaccinated. Among demographic characteristic, education level, monthly income, and self-care ability were significantly associated with vaccination status.

Interaction effects between 5C scale and demographic characteristics

Likelihood ratio tests revealed significant interactions between Confidence and educational level (p < .001), monthly income (p < .001), self-care ability (p = .002). Significant interactions were also found between Collective Responsibility and age group (p = .002), educational level (p = .002), monthly income (p = .003), and self-care ability (p = .008). Furthermore, Calculation interacted significantly with monthly income (p < .001), self-care ability (p = .027), while Complacency showed a significant interaction with monthly income (p < .001). No significant interactions were observed for Constraint subscale. Detailed results are provided in Supplementary A Tables S1–S5.

Figure 2 present the marginal effects plots for the significant interactions. For Confidence subscale, the association with vaccination became positive only among those with higher education (high school and above) and middle-income levels, whereas it was negative for lower-income levels, whereas it was negative for lower-income and lower-education groups. Similarly, higher education and income levels amplified the positive association of Collective Responsibility with vaccination. For Calculation, higher income levels strengthened its positive association with vaccination.

Figure 2.

(A) Effect of Confidence by monthly income and education level. Note: All models adjusted. (B) Effect of Collective Responsibility by age group, monthly income, education level and self-care ability. Note: All models adjusted. (C) Effect of Complacency on monthly income; (D) Effect of Calculation by monthly income and self-care ability. Note: All models adjusted.

Figure 2.

(Continued).

Logistic regressions for influenza vaccination by subgroup

Stratified analysis revealed significant heterogeneity in the association between 5C dimensions and influenza vaccination across sociodemographic subgroups. While Constraint and Complacency consistently showed negative associations with vaccination across all subgroups, the effects of Confidence, Collective Responsibility, and Calculation varied substantially. The comprehensive subgroup analysis results are available in Supplementary Tables S6–S10 and visualized in Figures 3–6.

Figure 3.

Logistic regressions for influenza vaccination by age group.

Note: All models adjusted for gender, educational level, monthly income, chronic illness status, and self-care ability.

Figure 4.

Logistic regressions for influenza vaccination by education level.

Note: All models adjusted for gender, age, monthly income, chronic illness status, and self-care ability.

Figure 5.

Logistic regressions for influenza vaccination by monthly income.

Note: All models adjusted for gender, age, education, chronic illness status, and self-care ability.

Figure 6.

(A) Logistic regressions for influenza vaccination by chronic illness status; (B) logistic regressions by self-care ability.

Notes: (A) Models adjusted for gender, age, education, monthly income, and self-care ability.

(B) Models adjusted for gender, age, education, monthly income, and chronic illness status, self-care ability.

By age group, Calculation showed a positive association with vaccination. Collective Responsibility had a significant negative effect among participants aged 60–69 y, and positive effect among those aged 70–79 y. By education level, Calculation had a significant positive effect among those with middle school, high school, and college and above. Confidence had a significant negative effect among those with primary or below education and middle school education, but a significant positive effect among those with high school and college or above education. By monthly income, Confidence showed a significant positive effect in 5001–7500 CNY group, and Collective Responsibility had a significant positive effect among those with income >7500. Calculation also showed positive associations in the 2501–5000 CNY, 5001–7500 CNY, and >7500 CNY groups. By self-care ability, Confidence had a significant negative association among those with limited independence.

Discussion

A cross-sectional survey was conducted among 13,363 adults aged 60 y and older across six Chinese cities to examine the psychological antecedents of influenza vaccination. Crucially, our results demonstrate that the effects of the 5C components are not uniform across subpopulations – the strength and direction of associations vary meaningfully by age, education level, income, chronic illness status, and self-care ability. These findings underscore the importance of tailoring interventions to specific psychological profiles within distinct demographic strata.

This study found that Calculation, Complacency, and Constraints significantly influenced influenza vaccination. The original 5C model often conceptualizes Calculation as a marker of hesitancy, implying that extensive information-seeking reflects risk aversion and “paralysis by analysis.”¹⁸ However, a notable finding that contrasts with previous research is the positive association between Calculation and vaccination. This divergence may be attributed to the specific demographic composition of our sample, which consisted of urban older adults with relatively higher socioeconomic status. Participants may have been more capable of engaging in information-seeking behaviors and rational decision-making. In this context, extensive information-seeking likely reflects “pro-vaccination deliberation” rather than indecision.^25,26 Unlike individuals who search for information to confirm biases against vaccines, high-SES older adults with greater health literacy may engage in active information processing to confirm safety and efficacy, thereby transforming their “calculation” into a reinforcing mechanism for vaccination intent.²⁷ Furthermore, our subgroup analysis showed that this positive association was strongest among those with higher incomes. This supports the notion that individuals with greater resources possess the capacity to effectively filter and interpret medical information, enabling them to make informed, rational decisions that favor health protection rather than being overwhelmed by conflicting data.^28,29

Regarding Constraints, our results showed a universal negative association with vaccination. The significant negative association suggests that older adults who perceive the vaccination process as convenient and transparent are significantly more likely to be vaccinated. This implies that even in the absence of severe physical barriers, the psychological perception of the process as cumbersome, opaque, or time-consuming acts as a major deterrent.^30,31 Therefore, increasing vaccination rates requires not only establishing more clinics but also simplifying the “user experience” – such as by providing clear, jargon-free guidelines and streamlining the appointment process to reduce these perceived hurdles.^32–35

Complacency, on the other hand, was universally associated with lower odds of vaccination in all subgroups, suggesting a consistently negative influence. Older adults with high Complacency perceive a low threat from influenza and are less likely to receive the vaccine.^18,36

Significant heterogeneity was also observed regarding Confidence, particularly across economic and educational levels. We found that Confidence had a positive effect on influenza vaccination only among older adults with higher education (high school or above) and high income (5001–7500 CNY per month). This likely reflects greater vaccine literacy in high-SES groups, enabling them to interpret scientific information accurately and base their confidence on evidence on evidence.³¹ However, among low-income and low-education groups, vaccine hesitancy may arise because these individuals interpret vaccine safety as “I don’t need it,” rather than recognizing its benefits. They may perceive that the potential risks of vaccination outweigh its health benefits.^13,37

While Collective Responsibility did not show a significant correlation in the overall analysis of the total population, it exerted a notable influence among specific subgroups. We found that Collective Responsibility had a positive effect among younger seniors, those with higher education and income, and those with limited independence, but a negative effect among older, low-education, low-income, and fully independent adults. While Collective Responsibility typically correlates with altruism,^18,38 its negative impact in these specific groups warrants nuanced interpretation. We speculate that for older adults facing resource scarcity (low income) or age-related frailty, the abstract concept of “herd immunity” may feel distant compared to immediate survival needs. For these individuals, appeals to civic duty might inadvertently feel like an additional burden. Research indicates that lower-SES individuals often perceive higher environmental threats and may prioritize immediate self-preservation strategies over medical interventions that they perceive as carrying risks to their own health, effectively prioritizing “survival of the self” over the “protection of the community.”³⁹

Furthermore, given that approximately two-thirds of our respondents reported at least one chronic condition, the role of health status is profound. While not all chronic conditions result in severe immune compromise, the high prevalence of multimorbidity implies a degree of immunosenescence. For these older adults, the decision often hinges on a subjective assessment of risks and benefits, where the fear of disrupting their stable health status outweighs the fear of influenza.¹³

The observed heterogeneity highlights the need for precision public health strategies. First, high-income and highly educated individuals tend to evaluate risks rationally. Therefore, they may benefit from messaging that emphasizes personal health benefits and cost-effectiveness. Second, for low-income and low-education groups, public health education should use emotionally resonant communication. Messages focusing on personal safety are likely more useful.⁴⁰ Finally, addressing the concerns of the chronically ill requires a personalized approach. Physicians should explicitly explain the balance of risks and benefits, clarifying that vaccination serves as a protective shield for their specific condition rather than a health burden. For those who remain hesitant, culturally appropriate strategies should be adopted. This includes integrating advice from trusted Traditional Chinese Medicine practitioners to build trust.⁴¹

Limitation

As a cross-sectional study, this research cannot infer causality. Additionally, because the study was conducted in urban settings, rural older adults were not included. Vaccination status was self-reported, and although the survey was conducted face-to-face, there remains the potential for information bias.

Conclusion

This study provides large-scale empirical evidence on the psychological determinants of influenza vaccination among Chinese older adults using the 5C model. The findings demonstrate that psychological factors – particularly Calculation, Complacency, and Constraints – play crucial and heterogeneous roles in shaping vaccination behavior. Calculation was positively associated with vaccine uptake, suggesting that informed evaluation and rational decision-making facilitate vaccination. In contrast, higher Complacency and greater perceived Constraints were consistently linked to lower vaccination rates, highlighting the need to address low perceived susceptibility and logistical barriers. Moreover, the effects of 5C dimensions varied significantly across subgroups defined by education, income, and self-care ability, emphasizing the importance of precision public health approaches.

Targeted interventions that enhance vaccine literacy, reduce complacency, and simplify vaccination procedures are essential to increase coverage among older adults. Future studies should employ longitudinal designs to clarify causal pathways and evaluate tailored behavioral strategies that promote vaccine acceptance within diverse aging populations.

Biography

Lili You, PhD, is an Associate Professor a of Peking Union Medical College, specializing in health education and promotion, health management, essential public health services, adolescent and child health literacy, and health city evaluation. Her research focuses on chronic disease management, digital interventions for vaccine hesitancy, and the development of health literacy assessment tools. She currently serves as a member of the National Expert Group on Essential Public Health Services under the National Health Commission and holds editorial positions.

Funding Statement

This study was supported by the Beijing Natural Science Foundation-Haidian Original Innovation Joint Fund Project [grant number L242146].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data underlying this article will be shared upon reasonable request to the corresponding author.

Supplementary Information

Supplemental data for this article can be accessed online at https://doi.org/10.1080/21645515.2026.2616903