Socioeconomic disparities in childhood vaccine hesitancy among parents in China: The mediating role of social support and health literacy

ABSTRACT

Parental vaccine hesitancy is a major obstacle to childhood vaccination. We examined parental socioeconomic status (SES) disparities in vaccine hesitancy, and the potential mediating roles of perceived social support and health literacy. A questionnaire survey was given to parents with children aged below 6 years from six provinces in China. SES was examined by educational attainment, annual household income, and a subjective measure of SES (using a scale of 1–10). Linear regression was applied to assess the association between SES and vaccine hesitancy. Bootstrapping mediation analysis was performed with 5,000 samples bootstrapped. A total of 1,638 parents were included. Using annual household income > 200,000 Chinese yuan (CNY) as a reference, parents with lower household income (CNY 100,001–150,000) experienced higher vaccine hesitancy. Educational attainment was not associated with vaccine hesitancy. Subjective SES had a U-shaped relationship with vaccine hesitancy. Perceived social support and health literacy independently and sequentially mediated the effects of subjective SES (indirect effect: −0.240) and annual household income (indirect effect: 1.250 for ≤ CNY 100,000 and 0.759 for CNY 100,001–150,000) on vaccine hesitancy. Socioeconomic disparities influenced parental vaccine hesitancy in China, which were mediated by perceptions of social support and health literacy.

Introduction

Vaccine hesitancy is one of the top ten threats to global health and is prominent in 90% of countries worldwide.^1,2 Vaccine hesitancy contributes to inadequate vaccination and undermines the achievements of immunization programs, which may result in resurgences of vaccine-preventable diseases. Parental vaccine hesitancy is an obstacle to improve vaccination rates of children, especially for those with children under six years old because they are recommended to complete most vaccination in China. Parental vaccine hesitancy is influenced by their socioeconomic status (SES).^3,4 Parental SES shapes childhood vaccination attitudes and choices owing to fewer resources such as knowledge, money, and beneficial social relationships. Assessing parental SES disparities in vaccine hesitancy is important because it is an essential aspect of promoting health equalities among children around the world.⁵

SES can be measured by objective and subjective ways.⁶ Objective SES is mainly measured by educational attainment and income, while subjective SES reflects an individual’s perceived social status relative to others and more accurately captures subtle aspects of social status.^7,8 Studies have found a significant linear relationship between objective SES and vaccine hesitancy, but the direction of the relationship (i.e., higher levels of vaccine hesitancy in advantaged SES groups, or higher levels of vaccine hesitancy in disadvantaged SES groups) is inconsistent and varies by country.^9,10 The effect of subjective SES on vaccine hesitancy has rarely been studied, but a US study revealed a negative association between subjective SES and vaccine hesitancy.¹¹ In China, evidence of SES disparities in vaccine hesitancy is scarce. However, one study on disparities in human papillomavirus vaccination found that parents with lower income and educational attainment were more likely to vaccinate their daughters.¹²

The pathway by which SES influences vaccine hesitancy has not been clarified. Few studies have explored the relevant mechanisms. However, it has been reported that commitment to good health-related decisions, trust in the health system, perceived healthcare quality, and vaccination knowledge and beliefs may mediate the association between SES and vaccine hesitancy.^11,13,14 The World Health Organization Commission on Social Determinants of Health and the MacArthur Network (a research network supported by the MacArthur Foundation aiming to reduce health disparities) on SES and Health have each developed a conceptual framework to show how SES may influence health outcomes through intermediary determinants.^15,16 The two frameworks indicate that SES influences vaccine hesitancy through environmental factors (physical, social, and financial resources) and individual factors (psychology, cognition, and ability).

Social support and health literacy are important predictors for vaccine hesitancy.^17,18 Social support has been defined as a cognitive appraisal of being connected to others, and knowing that support is there if needed.¹⁹ There is robust evidence that advantaged SES is beneficial to health literacy and social support.^20,21 Based on these findings, it can be inferred that social support and health literacy may be intermediate factors by which SES influences vaccine hesitancy. Previous studies have found that social support and health literacy mediate the association of SES with health outcomes such as mental health and quality of life.^22–24

Although previous studies reported mixed directions on the association between SES and vaccine hesitancy, the majority indicated significant findings. Since the association was influenced by multiple factors, such as vaccination environment and policies, it varied among countries. Chinese parents of children have high levels of vaccine hesitancy and uneven pattern of SES.²⁵ Thus, similar studies were warranted to perform in China. Potential vaccination strategies could be identified to improve uptake rates and reduce inequality of childhood vaccination in China. Following recommendations to use multiple measures of SES and consider their interaction,²⁶ in the present study, parents with children under six years old were investigated by using objective and subjective measures of SES to 1) assess the associations between parental SES measures and vaccine hesitancy; 2) explore the influence of interactions between different SES indicators on parental vaccine hesitancy; and 3) analyze the mediating role of social support and health literacy in the association between parental SES and vaccine hesitancy.

Methods

Theory framework

Previous studies have found that social support and health literacy are associated with SES and vaccine hesitancy.^17,18,20,21 Existing theoretical frameworks suggest that environmental and individual factors might explain the impact of SES on health outcomes.¹⁶ Social support and health literacy are environmental and individual factors, respectively. Strong social support is beneficial to individual access to information and improves health literacy through interpersonal interactions.²⁷ We therefore hypothesized that social support and health literacy play a mediating role in the relationship between SES and vaccine hesitancy. Based on previous studies on vaccine hesitancy and theoretical frameworks on how SES influences health outcomes, we established a theoretical framework to examine the pathway between SES and vaccine hesitancy in the present study, as shown in Figure 1.

Figure 1.

Theoretical framework of the effect of socioeconomic status on vaccine hesitancy via the mediation of perceived social support and health literacy.

Measures

A questionnaire survey was conducted for children’s parents to assess their hesitancy on childhood vaccination. The questionnaire was consisted of participants’ individual socio-demographic information (sex, age, residence, medical insurance, whether working in healthcare settings, number of children, and age and sex of the youngest child), vaccine hesitancy, social support, health literacy, and SES indicators (participants’ education, income, and subjective SES).

Vaccine hesitancy was measured by our vaccine hesitancy scale, which consists of 17 items to measure parents’ confidence, complacency, convenience, and calculation in relation to childhood vaccination.²⁸ In this scale, vaccine hesitancy is defined as a motivational state of being conflicted about or opposed to childhood vaccination. Each item was rated on a 5-point Likert scale from “strongly agree” to “strongly disagree,” with higher scores indicating a higher level of vaccine hesitancy. The scale showed good reliability and validity among parents with children below 6 years old in China.

Social support was assessed using the perceived social support scale.²⁹ The scale is a 3-item instrument measuring perceived support from family, friends, and important others, which was rated on a 7-point Likert scale ranging from “strongly disagree” to “strongly agree.” Higher scores indicated greater perceived social support. The Chinese version of the scale has been validated and has shown good reliability.²⁹

Health literacy was measured using a simplified version of the health literacy scale.³⁰ The simplified scale included four items and was validated as a reliable and effective tool for assessing health literacy levels in Chinese populations. Each item was rated on a 4-point Likert scale from “very difficult” to “very easy.” The higher the score, the greater the participant’s level of health literacy.

SES was assessed by both objective and subjective measures. Parents’ educational attainment and income were used as indicators of their objective SES. Educational attainment consisted of the following categories: high school or below, junior college, college, and master’s degree or higher. Income was assessed by yearly household income and categorized as follows according to the average household disposable income (HDI) in China for 2023: ≤ Chinese yuan (CNY) 100,000 (1-fold HDI), CNY 100,001–150,000 (1- to 1.5-fold HDI), CNY 150,001–200,000 (1.5- to 2-fold HDI), and > CNY 200,000 (2-fold HDI). For the assessment of subjective SES, we referred to the MacArthur scale, requiring respondents to assess their social position on a ladder with 10 rungs.³¹ Participants were asked to rate themselves on a scale of 1 to 10 prompted by the statement: “What level do you think your family’s current socioeconomic status generally is?”

Sampling and data collection

In China, children are suggested to complete most vaccination by age of six. The vaccine hesitancy scale was also applicable to parents with children ≤6 years old, which was consistent with the selection criteria of participants in our study. Parents with children beyond the age were therefore excluded. Based on geographical location and Gross Domestic Product per capita,³² we selected six Chinese provinces: Guangdong and Jiangsu (eastern regions), Henan and Jiangxi (central regions), and Sichuan and Shaanxi (western regions). The six provinces were at different layers based on China’s Division of Central and Local Financial Expenditure Responsibilities in the Healthcare Sector, which stratified all provinces into five layers.³³ Data were collected from November 2023 to January 2024 using Questionnaire Star, the most widely used online survey platform in China. The sample size was calculated using the formula $\mathit{n}=\mathit{deff}\times \frac{Z_{\left(1-\alpha /2\right)}^2\times \mathit{p}\times \left(1-\mathit{p}\right)}{d^2}$,³⁴ in which α error was 0.05, permissible error (d) was 0.04, and design effect (deff) was 2. The prevalence of vaccine hesitancy was assumed to be 50% in order to obtain the maximum sample size. The required sample size was estimated to be 1,201.

Statistical analysis

Vaccine hesitancy, health literacy, perceived social support, and subjective SES were presented by means and standard deviation (SD), while objective SES including educational attainment and household income, and demographic variables (except for age of the youngest child) were presented by percentages. Bivariate analyses were performed to study the associations between vaccine hesitancy and demographic variables.

To analyze SES disparities in vaccine hesitancy, we first used analysis of variance and the Pearson correlation coefficient to assess the associations between educational attainment, household income, subjective SES, and vaccine hesitancy. Next, we performed multiple linear regression for each SES indicator and adjusted for demographic variables found to be significantly associated with vaccine hesitancy in bivariate analyses. Finally, we constructed multiple linear regression and added interaction terms between educational attainment, household income, and subjective SES to determine the effects of these interactions on vaccine hesitancy. To examine whether there is a U-shaped relationship between subjective SES and parental vaccine hesitancy, we plotted the scatter plot and added the square of subjective SES as a new independent variable to the regression model. If subjective SES and its square both showed significant associations with vaccine hesitancy and the signs for their regression coefficients were opposite, there was a U-shaped relationship. We calculated the turning point of the fitted U-shaped curve based on the regression results and estimated the marginal effects (slope of the curve) to further confirm the relationship.

Bootstrapping methods were employed to test the mediating role of perceived social support and health literacy. We conducted bivariate analyses to evaluate whether perceived social support and health literacy were associated with SES indicators and vaccine hesitancy. We next performed bootstrapping mediation analyses with biases corrected by using 5,000 bootstrapped resamples and controlling for the influence of demographic variables.³⁵ Indirect effect was considered significant when the 95% confidence interval (CI) did not include zero. Direct effect refers to the impact of SES on vaccine hesitancy after controlling for mediators. Total effect refers to the impact of SES on vaccine hesitancy without controlling for mediators, which is the sum of direct and indirect effects. Mediation analyses were performed in Stata version 15.0 (StataCorp., College Station, TX, USA) and all other analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). The level of significance used was p < .05.

Ethical consideration

The study was approved by the Ethics Committee for Medical Research of Xi’an Jiaotong University. The survey was conducted upon the informed consent of all participants.

Results

Characteristics of participants

A total of 1,715 questionnaires were collected, of which 54 questionnaires were excluded for residency in non-target provinces or having children of ineligible age, while 23 questionnaires were excluded for recorded response times of less than 2 minutes (the shortest time we tested) or identical responses to every question on a single page. A total of 1,638 valid questionnaires were included in statistical analyses. The socio-demographic characteristics of all participants are shown in Table 1. Of all participants, 64.5% were mothers, 60.6% resided in urban areas, 83.2% were under the age of 35, and 11.9% worked in healthcare settings. In addition, 81.9% of parents had a bachelor’s degree or above and 47.0% had an annual household income below CNY 150,000. The mean subjective SES score was 5.82 (SD = 1.45).

Table 1.

Sociodemographic characteristics of participants and associations with vaccine hesitancy.

	N (%)/ Mean (SD)	Vaccine hesitancy, mean (SD)	P for difference in vaccine hesitancy
Demographic Characteristics
Gender			.195
Male	582 (35.5)	37.15 (7.34)
Female	1,056 (64.5)	37.67 (7.93)
Age (years)			.327
≤30	635 (38.8)	37.24 (6.79)
31 − 35	727 (44.4)	37.49 (8.27)
>35	276 (16.9)	38.05 (8.24)
Place of residence			.247
Urban	992 (60.6)	37.67 (7.66)
Rural	646 (39.4)	37.21 (7.83)
Medical insurance			<.001
Urban employee medical insurance	969 (59.2)	37.50 (7.84)
Urban and rural resident medical insurance	646 (39.4)	37.25 (7.48)
Other/No insurance	23 (1.4)	43.65 (7.40)
Working in healthcare settings			<.001
Yes	195 (11.9)	39.53 (8.54)
No	1,443 (88.1)	37.21 (7.57)
Number of children			.601
1	1,266 (77.3)	37.43 (7.79)
≥2	372 (22.7)	37.67 (7.51)
Gender of child			.060
Boy	838 (51.2)	37.14 (7.33)
Girl	800 (48.8)	37.86 (8.10)
Age of child (years)	2.15 (1.56)	37.49 (7.73)	.025
Regions			.820
Eastern	749 (45.7)	37.56 (8.09)
Central	480 (29.3)	37.11 (6.80)
Western	409 (24.8)	37.79 (8.04)
Socioeconomic status
Education level			.661
High school or below	69 (4.2)	38.07 (9.56)
Junior college	228 (13.9)	37.58 (7.85)
College	1,243 (75.9)	37.38 (7.69)
Master’s degree or higher	98 (6.0)	38.23 (6.39)
Annual household income (CNY)			.018
≤100,000	283 (17.3)	37.63 (6.59)
100,001 − 150,000	487 (29.7)	38.31 (8.15)
150,001 − 200,000	510 (31.1)	36.79 (7.49)
>200,000	358 (21.9)	37.25 (8.20)
Subjective SES	5.82 (1.45)	37.49 (7.73)	.413

SES: socioeconomic status; CNY: Chinese yuan; SD: standard deviation.

Association between SES and vaccine hesitancy

Bivariate analysis and linear regression showed a significant association between household income and vaccine hesitancy, while there was no evidence that education was associated with vaccine hesitancy (Table 1 , Table 2). Using annual household income above CNY 200,000 as a reference, parents with household incomes of CNY 100,001–150,000 had higher levels of vaccine hesitancy. In addition, working in healthcare settings, type of medical insurance, and age of child were significantly associated with vaccine hesitancy.

Table 2.

Linear regression between SES and vaccine hesitancy.

	Model 1	Model 2	Model 3	Model 4
Subjective SES	−0.123	−2.374***	−3.214	−7.023**
Square of Subjective SES		0.195***	0.215	0.526**
Education level (reference: Master’ degree or higher)
High school or below	−0.333	−0.390	6.012	−0.450
Junior college	−0.685	−0.726	−6.425	−0.679
College	−0.757	−0.685	−6.080	−0.746
Annual household income (reference: > CNY 200,000)
≤100,000	0.540	0.350	0.408	−21.731**
100,001 − 150,000	1.277*	1.234*	1.297*	−9.623
150,001 − 200,000	−0.240	−0.240	−0.181	−16.423
Age of child (years)	0.331**	0.326**	0.321**	0.326**
Working in healthcare settings (reference: No)
Yes	2.492***	2.328***	2.326***	2.134
Medical insurance (reference: Urban employer medical insurance)
Urban and rural resident medical insurance	−0.514	−0.543	−0.545	−0.555
Other/No insurance	6.057***	5.643***	5.573**	5.518**
Interaction
High school or below*Subjective SES			−2.809
High school or below*Square of subjective SES			0.251
Junior college*Subjective SES			1.671
Junior college*Square of subjective SES			−0.122
College*Subjective SES			0.973
College*Square of subjective SES			−0.016
≤100,000*Subjective SES				6.891**
≤100,000*Square of subjective SES				−0.506*
100,001 − 150,000*Subjective SES				2.714
100,001 − 150,000*Square of subjective SES				−0.149
150,001 − 200,000*Subjective SES				4.997
150,001 − 200,000*Square of subjective SES				−0.371
R2	0.031	0.037	0.041	0.045

Model 1: Linear regression between all SES indicators and vaccine hesitancy, adjusted for age of child, working in healthcare settings, and medical insurance type.

Model 2: Adding the quadratic term of subjective SES in Model 1.

Model 3: Adding the interaction terms between educational attainment and subjective SES in Model 2.

Model 4: Adding the interaction terms between annual household income and subjective SES in Model 2.

SES: socioeconomic status; CNY: Chinese yuan.

***p < .001, **p < .01, *p < .05.

Regression analyses found no linear relationship between subjective SES and vaccine hesitancy (Table 2). However, through the scatter plot, we found that a U-shaped association between these two variables might exist (Figure 2a). Linear regression found that the effect of subjective SES was significantly negative (−2.374), while the effect of its quadratic term was significantly positive (0.195), indicating a U-shaped relationship (Table 2). The fitted curve illustrated that increasing subjective SES first decreased vaccine hesitancy, and then increased it after crossing the turning point (6.09; 95% CI: 5.36 to 7.02; p < .001). The marginal effects significantly increased from − 1.98 (95% CI: −3.12 to − 0.85) to 1.52 (95% CI: 0.51 to 2.54) as subjective SES score increased from 1 to 10 (Figure 2b).

Figure 2.

The scatter plot (a) and marginal effects (b) between subjective SES and vaccine hesitancy.

We also analyzed the association between SES and four dimensions of vaccine hesitancy (confidence, complacency, convenience, and calculation) (Table S1, Figure S1). Linear regression indicated that parents with annual household incomes of CNY 100,001–150,000 had lower levels of confidence and higher levels of complacency compared with the reference category of household incomes above CNY 200,000. Subjective SES was positively associated with confidence, and had a U-shaped effect on complacency.

Interaction between objective and subjective SES on vaccine hesitancy

There was no evidence of interaction between educational attainment and subjective SES, while a significant interaction between annual household income (≤ CNY 100,000) and subjective SES was identified (Table 2). The estimate (−0.506, p = .016) for the interaction between household income below CNY 100,000 and the squared term of subjective SES was negative, indicating that the U-shaped curve for the relationship between subjective SES and vaccine hesitancy among parents with household income below CNY 100,000 was flatter compared with the reference category of household income above CNY 200,000 (Figure S2). Therefore, each unit of increase in subjective SES resulted in smaller changes in vaccine hesitancy among parents with lower household incomes.

Mediating effect of perceived social support and health literacy

Bivariate analyses found that perceived social support and health literacy were positively associated with annual household income and subjective SES and negatively correlated with vaccine hesitancy, but not correlated with educational attainment (Table S2).

The mediating effects of perceived social support and health literacy on the association between SES and vaccine hesitancy are shown in Table 3, while the mediating pathways are shown in Figure S3. Using annual household income above CNY 200,000 as the reference group, there was no evidence of the direct effect and total effects of household income below CNY 100,000 on vaccine hesitancy. However, the indirect paths of household income below CNY 100,000 influencing vaccine hesitancy through perceived social support, health literacy, and the chain of perceived social support and health literacy were all statistically significant. The estimated effect sizes for the three indirect paths were 0.752 (95% CI: 0.456 to 1.123), 0.383 (95% CI: 0.181 to 0.658), and 0.115 (95% CI: 0.062 to 0.195), respectively. Among the three indirect paths, perceived social support as a single mediator explained most of the indirect effects (60.2%). For annual household incomes CNY 100,001–150,000, the total effect (1.249, 95% CI: 0.111 to 2.389) was significant, while there was no evidence of direct effect. The effect sizes for the three indirect paths through perceived social support, health literacy, and the chain of perceived social support and health literacy were 0.458 (95% CI: 0.237 to 0.725), 0.231 (95% CI: 0.083 to 0.435), and 0.070 (95% CI: 0.034 to 0.126), respectively.

Table 3.

Mediating effects of perceived social support and health literacy on the association between SES and vaccine hesitancy.

Paths	Estimate	SE	95% CI
Annual household income (reference: > CNY 200,000)
≤100,000
Indirect effect
≤100,000 → Perceived social support → Vaccine hesitancy	0.752	0.169	0.456, 1.123
≤100,000 → Health literacy → Vaccine hesitancy	0.383	0.120	0.181, 0.658
≤100,000 → Perceived social support → Health literacy → Vaccine hesitancy	0.115	0.033	0.062, 0.195
Direct effect	−0.834	0.658	−2.115, 0.463
Total effect	0.416	0.678	−0.910, 1.748
100,001 − 150,000
Indirect effect
100,001 − 150,000 → Perceived social support → Vaccine hesitancy	0.458	0.124	0.237, 0.725
100,001 − 150,000→ Health literacy → Vaccine hesitancy	0.231	0.088	0.083, 0.435
100,001 − 150,000 → Perceived social support → Health literacy → Vaccine hesitancy	0.070	0.023	0.034, 0.126
Direct effect	0.489	0.565	−0.638, 1.597
Total effect	1.249	0.580	0.111, 2.389
150,001 − 200,000
Indirect effect
150,001 − 200,000 → Perceived social support → Vaccine hesitancy	0.181	0.105	−0.024, 0.387
150,001 − 200,000 → Health literacy → Vaccine hesitancy	−0.016	0.072	−0.165, 0.127
150,001 − 200,000 → Perceived social support → Health literacy → Vaccine hesitancy	0.028	0.017	−0.001, 0.066
Direct effect	−0.432	0.523	−1.512, 0.585
Total effect	−0.240	0.553	−1.349, 0.815
Subjective SES
Indirect effect
Subjective SES → Perceived social support → Vaccine hesitancy	−0.105	0.040	−0.191, −0.036
Subjective SES → Health literacy → Vaccine hesitancy	−0.119	0.031	−0.188, −0.067
Subjective SES → Perceived social support → Health literacy → Vaccine hesitancy	−0.016	0.007	−0.033, −0.006
Direct effect	−1.356	0.750	−2.871, 0.056
Total effect	−1.596	0.756	−3.115, −0.150

SES: socioeconomic status; CNY: Chinese yuan; SE: standard error; CI: confidence interval.

The total effect of subjective SES on vaccine hesitancy was significant (−1.596, 95% CI: −3.115 to − 0.150), but there was no evidence of direct effect. The indirect paths through perceived social support, health literacy, and the chain of perceived social support and health literacy were all statistically significant. The effect sizes for the three indirect paths were − 0.105 (95% CI: −0.191 to − 0.036), −0.119 (95% CI: −0.188 to − 0.067), and − 0.016 (95% CI: −0.033 to − 0.006), respectively. The effects of the indirect path through health literacy explained 49.6% of overall indirect effects.

Discussion

Our study found that parents with relatively low household incomes were more hesitant about vaccines, and subjective SES had a U-shaped effect on vaccine hesitancy. Educational attainment was not associated with vaccine hesitancy. Subjective SES and vaccine hesitancy exhibited a flatter U-shaped relationship among parents with relatively low household incomes. Social support and health literacy may explain the association between SES and parental vaccine hesitancy.

Most previous studies have found linear relationships between educational attainment or income with vaccine hesitancy, which were either positive or negative.^9,10,26 Our findings that parents with relatively low household incomes had high levels of vaccine hesitancy were consistent with previous studies.^12,36 However, we found that educational attainment was not correlated with vaccine hesitancy. There may be two possible reasons to explain this. First, most participants in our study were highly educated. Second, previous studies mainly used vaccination behavior, vaccination refusal, and vaccination delay as proxy variables for vaccine hesitancy,^13,37 while in the present study, vaccine hesitancy was measured as a set of beliefs and motivational states of being conflicted about or opposed to vaccination. The U-shaped relationship between subjective SES and vaccine hesitancy in our study was inconsistent with a US study, which described a negative association between subjective SES and COVID-19 vaccine hesitancy.¹¹ Differences between the instruments used to measure vaccine hesitancy and subjective SES may explain the inconsistent results. In the US survey,¹¹ vaccine hesitancy was measured as participants’ willingness to get a COVID-19 vaccination, while subjective SES was determined by summing survey items reflecting financial satisfaction, perceived financial stability, and the MacArthur ladder. In the present study, healthcare workers were found to have greater vaccine hesitancy, which was consistent with the result reported in a previous study from another country.³⁸ Perception of less vulnerability, concerns about vaccine safety, and mistrust in health authorities were the main reasons for possibly high levels of vaccine hesitancy among healthcare workers.³⁹ Social support and health literacy explained the effects of income on parental vaccine hesitancy. With increasing household incomes, parents have more favorable social relationships and better access to medical resources and health-related information, thereby gaining stronger social support and higher levels of health literacy, which may reduce vaccine hesitancy.⁴⁰ Results of mediation analyses support previous findings that health literacy mediates the relationship between income and vaccine confidence.²⁴ In particular, we found that the mediating effect of social support explained most of the indirect effects, suggesting that increased social support for parents with low household incomes has a prominent effect on reducing vaccine hesitancy. The results support renewed policy focus on parents with low household incomes, such as the adoption of public health strategies tailored to reduce vaccine hesitancy in this group.⁴¹ For example, the government and health systems could provide more health services to support parents’ decisions on childhood vaccination, such as family doctor services or community healthcare center outreach. Education and communication are also vital to enhance the ability of parents to access and understand information about vaccination and vaccine-preventable diseases.⁴¹

Our study found that subjective SES enhanced social support and health literacy, while a mediating effect of social support and health literacy was found prior to the turning point of the U-shaped relationship between subjective SES and vaccine hesitancy. However, the phenomenon that subjective SES increased vaccine hesitancy after this turning point could not be well explained. We found that the increase of complacency levels may help to explain this. Complacency refers to parents believing that diseases preventable by vaccine are low-risk, and that their children are healthy and do not require vaccination.²⁸ High levels of complacency among parents with high subjective SES correspond with a socio-cultural phenomenon in the middle classes: healthism.⁴² Healthism is characterized by high health awareness, expectations and information-seeking, emphasizing that individuals are responsible for their own health-related behaviors and decisions.^42,43 Individuals thus actively seek information through various sources such as the internet and conduct their own research on vaccines. They are more likely to receive misinformation and underestimate the risk and threat of vaccine-preventable diseases,⁴⁴ and thus more likely to spread false information and to refuse effective preventive measures.⁴⁵ Some studies have found that healthism is an underlying reason for the high levels of vaccine hesitancy among advantaged SES groups.^44,46,47

Although the present study found that income and subjective SES were associated with vaccine hesitancy, it proved to be weak. Two major reasons might explain this finding. Participants in our study had high levels of education, and variations of vaccine hesitancy in highly educated parents might be minimal, which weakened the relationship between SES and vaccine hesitancy. Additionally, unmeasured confounding variables could also contribute to the heterogeneity of vaccine hesitancy among parents, and SES was merely able to explain part of it. Although the association between SES and vaccine hesitancy was weak, it was pivotal to identify specific pathways that SES influencing vaccine hesitancy. Mediation analysis was therefore performed in the present study to explore the mechanism behind the association. The mediating variables were sound evidence to support policy making on guiding the implementation of interventions to reduce vaccine hesitancy, particularly in low SES parents.

This study has some limitations. First, parents were randomly selected by the Questionnaire Star platform, and they participated in the present study voluntarily without any incentives. Self-selection bias might occur if participants had stronger interests or were more willing to vaccinate children, which possibly underestimated the levels of vaccine hesitancy. Second, all data were self-reported by parents, which might cause recall and response bias. Third, majority of participants in our study were highly educated, which shared similar limitation as other online studies. The proportion of participants with a bachelor degree in our study was largely greater than the actual data (8.7%) in China,³² which might underestimate parents’ vaccine hesitancy levels and limit the generalizability of our findings. Forth, subjective SES is on the basis of objective SES. The U-shaped relationship between subjective SES and vaccine hesitancy led us to speculate that annual household income may also have a U-shaped effect on vaccine hesitancy. We measured annual household income as a categorical variable with four categories, which limited further exploration of its relationship with vaccine hesitancy. Future research can use income as a continuous variable to further explore this relationship.

Conclusions

SES disparities influence parental vaccine hesitancy in China. Parents with low household income experience high vaccine hesitancy. A U-shaped relationship between subjective SES and vaccine hesitancy may complicate vaccine policy formation. Social support and health literacy mediate the influence of household income and subjective SES on vaccine hesitancy.

Biography

Minghuan Jiang received his Ph.D. in the Chinese University of Hong Kong. Prof. Jiang is an associate Professor at School of Pharmacy in Xi’an Jiaotong University and serves as an adjunct researcher at Institute for Global Health and Development in Peking University. He was the principal investigator or participated in over 20 scientific research projects, including the National Natural Science Foundation, International Postdoctoral Exchange Fellowship Program (Talent-Introduction Program), and Shaanxi Provincial Natural Science Foundation, etc. He participated in the compilation of China’s first version of Guidelines on Vaccine Economics Evaluation. He has over 70 research papers published in peer-reviewed journals, which were mainly published in Lancet Global Health, American Journal of Preventive Medicine, PharmacoEconomics, Vaccine, etc. He currently serves as the vice president of ISPOR Northwest China Chapter, a member of Pharmacoeconomics Professional Committee of Chinese Pharmaceutical Association, and a member of Vaccine Economics Professional Committee of China Association for Vaccines, etc.

Funding Statement

This work was supported by the Natural Science Foundation of Shaanxi Province [grant number 2023-YBSF-374], Shaanxi Province Postdoctoral Science Foundation [grant number 2023], and International (Regional) Exchange and Cooperation Project of the National Natural Science Foundation of China [grant number 2023YFVA1004]. The funders of this study had no role in the management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

Disclosure statement

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

Author contributions

Yao X: Writing – original draft, Conceptualization, Visualization, Software, Methodology, Formal analysis, Data curation. Fu M: Investigation; Methodology, Project administration, Visualization, Software. Peng J: Investigation, Methodology, Project administration. Feng D: Methodology, Visualization, Writing – review & editing; Ma Y: Investigation, Methodology, Project administration. Wu Y: Investigation, Methodology, Project administration. Feng L: Conceptualization, Supervision. Fang Y: Writing – review & editing, Conceptualization, Resources, Supervision. Jiang M: Writing – review & editing, Conceptualization, Supervision, Validation. All authors had approved the final version of the manuscript for submission.

Data availability statement

All data and analysis code were available on request.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2024.2444008