Is vaccine hesitancy related to mental health after the adjustment of the zero-COVID-19 strategy in the elderly? A mediation analysis in China

ABSTRACT

With the global Omicron pandemic and the adjustment of the zero-coronavirus disease 2019 (zero-COVID-19) strategy in China, there is a critical need to improve vaccination rates among older adults while addressing the mental health issues associated with vaccination. This study investigated levels of COVID-19-related anxiety, depression, benefit finding, and fear in older adults and explored the relationship between vaccine hesitancy, sociodemographic factors, and mental health. Participants aged 60 and older (n = 658) were recruited from several cities in the eastern, central, and western China regions. Of these, 347 exhibited vaccine hesitancy. The effects of residence, education, health status, and COVID-19 vaccination on anxiety/depression/benefit-finding were found to be mediated/suppressed by vaccine hesitancy. Additionally, in investigating psychological antecedents, older people without vaccine hesitancy showed higher confidence, lower complacency, fewer constraints, and a greater sense of collective responsibility. This study advances our understanding of mental health differences in anxiety, depression, and benefit-finding across sociodemographic characteristics. It is essential to improve population confidence related to vaccines, accessibility to vaccination services, and responsibility to mitigate vaccine hesitancy while paying close attention to the mental health associated with vaccination in older adults.

KEYWORDS:

• COVID-19
• elderly
• vaccine hesitancy
• mental health
• zero-COVID-19 strategy
• mediation analysis

Introduction

China has revised its coronavirus disease 2019 (COVID-19) response strategy to balance sustainability and cost-effectiveness in implementing a dynamic zero-COVID-19 strategy. This resulted in the management of COVID-19 being downgraded from Class A to Class B under the country’s Law on the Prevention and Treatment of Infectious Diseases.^1,2 Consequently, centralized isolation, close contact tracing, or mass nucleic acid testing were no longer available. China’s response to COVID-19 was entering a new phase.³ During this phase, China confronted unprecedented challenges and opportunities to improve its public health system. However, the sudden policy change has an immediate impact on the Chinese public, who are experiencing a sudden outbreak of COVID-19 before life can return to normal.⁴

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) poses a high risk to the general population, with individuals aged 60 and above being particularly vulnerable to severe and potentially fatal infections.⁵ This heightened susceptibility is due to lower immune function, underlying chronic diseases, and malnutrition.^6,7 Moreover, the virus’s high mutation rate is of ongoing concern as it continues to generate new and potentially more infectious and atypical variants, including the Omicron subvariant BQ.1.1, BQ.1, and XBB.1.5 are currently prevalent worldwide.⁸ As China experiences a surge in cases due to the relaxation of strict infection control policies and a decline in vaccine-induced antibody titers among the elderly population,⁹ older adults have been significantly impacted.¹⁰ In Hong Kong’s fifth wave of COVID-19 in early 2022, the unvaccinated or incompletely vaccinated elderly population exhibited an extremely high mortality rate.¹¹ Notably, unvaccinated individuals had a case-fatality rate of 3.04%, compared to 0.17% and 0.04% for those who received two or three doses of vaccination, respectively, with the vaccine’s safety well-established.^12,13 Previous research has demonstrated a correlation between non-COVID-19 vaccination and lower quality of life among older individuals.¹⁴ Mathematical methods also demonstrated that vaccinating the oldest individuals first is the most effective strategy to reduce the age-related mortality risk from COVID-19 while maximizing the remaining life expectancy.¹⁵

Medical scientists and policymakers increasingly recognize that vaccination offers broader societal benefits beyond preventing infectious diseases and conserving medical resources. These benefits include improvements in health, economics, and social outcomes.^16,17 Therefore, the Chinese government is encouraging vaccination of older people, particularly booster shots.¹⁸ In April 2021, China opened COVID-19 vaccination to older adults after completing the first phase of emergency vaccination for high-risk and critically ill populations and the second phase of mass vaccination for adults aged 18–59.¹⁹ Regardless of residence, the government provides the COVID-19 vaccines free of charge. Despite the government’s efforts to promote vaccination among older adults, COVID-19 vaccination coverage increased gradually from January to June 2022, particularly for booster vaccination.¹⁹ The China CDC examined the vaccination status of 12,900 participants between July and August 2022, with an average age of 65.2 years.²⁰ The weighted vaccination rate (at least one dose) of the elderly aged 60 and above was 92.3%, and the full course and booster dose rates in this age group were 88.6% and 72.4%, respectively. The rates of vaccination exhibited a reduction as individuals advanced in age. In the oldest age group (80 years and older), 80.5% received a single dose; however, the proportions of individuals who completed the full course and received booster doses were 71.9% and 46.7%, respectively. These rates were lower than in the United States (92.1% for full vaccination, 70.7% for booster vaccination), Germany (91.2% for full vaccination, 85.9% for booster vaccination), and Japan (92.4% for full vaccination, 90.3% for booster vaccination).

The fact that a vaccine against a disease does not necessarily indicate that the entire population accepts it is cause for concern. The global dissemination of hesitancy regarding the COVID-19 vaccine among the elderly, which has extended to China, has been an ongoing phenomenon.^20–22 According to a national survey conducted in China in 2021, older adults express greater apprehension regarding the safety and effectiveness of the COVID-19 vaccine compared to other age groups.²³ Vaccine hesitancy is a multifaceted and context-dependent phenomenon that can vary over time, location, and type of vaccine.^24,25 Understanding vaccine hesitancy requires considering various psychological factors typically modeled by the 5C framework: confidence, complacency, constraint, calculation, and collective responsibility.^26,27

Epidemics have significant psychosocial impacts on individuals, often resulting in mental health issues such as health anxiety, depression, panic, and adjustment disorders.^28,29 Ambiguity and uncertainty surrounding vaccination can exacerbate vaccine hesitancy and create stress and anxiety. Previous literature has suggested a correlation between the delay and psychological disorders, including depression, anxiety, and psychotic symptoms.^30,31 Research has found a link between vaccine hesitancy and mental health psychopathology,³² which emphasizes the critical importance of mental health in public health, particularly for older adults who are especially susceptible to misinformation and uncertainty.³³ However, despite the unique challenges faced by older adults, including be weakened immune system and complex psychological stress during this stage of life,³⁴ few studies in China have explored the effect of vaccine hesitancy on mental health in this population.^32,35 To address this knowledge gap, more attention must be given to the mental health of older adults.³⁶ Mediated analyses effectively examine the relationships among multiple factors within a population, including sociodemographic characteristics and mental disorders.³⁷ Consequently, conducting a mediation analysis to investigate the link between vaccine hesitancy and psychological health in older adults would facilitate the identification of factors contributing to vaccine hesitancy, ultimately leading to improved vaccination rates and better overall mental health for this vulnerable population.

The low vaccination rates among older adults in China necessitate urgent action to accelerate the vaccination process and mitigate the impact of the adjustment of the dynamic zero-COVID-19 strategy. To this end, this study aims to investigate the relationship between sociodemographic factors, vaccine hesitancy, and mental health outcomes, such as depression, anxiety, benefit-finding, and fear related to the COVID-19 epidemic in older Chinese adults. Additionally, this study seeks to explore the psychological antecedents of vaccine hesitancy, aiming to develop informed educational programs and interventions to enhance vaccination rates.

Materials and methods

Study design and data collection

The study was conducted in several cities in eastern, central, and western China. Participants were recruited using a convenience sampling method from multiple community health centers between late December 2022 and February 2023. The eligibility criteria required participants to be older adults aged 60 years and above, except those who could not complete the questionnaire due to severe hearing, vision, or mobility impairments.

Study subjects who consented to participate were allowed to fill out a paper questionnaire or scan a QR code through their cell phones to complete the questionnaire. Additionally, each participant was permitted to fill out the questionnaire only once. Participants who completed the questionnaire were offered incentives, such as masks and disinfectant wipes, as compensation for their time. Seven hundred and twenty-six participants completed the questionnaire, but 68 were excluded from the study due to incomplete or unreliable responses (e.g., identical questionnaires or responses completed in less than 120 seconds). Thus, the final sample size for the study was 658 participants.

Questionnaire development and measures

The survey consisted of five distinct parts: (1) informed consent; (2) sociodemographic characteristics; (3) vaccine hesitancy status following the adjustment of the zero-COVID-19 strategy; (4) mental health status, which includes assessments of depressive and anxiety symptoms, benefit-finding from the epidemic, and fear of COVID-19; and (5) psychological antecedents of COVID-19 vaccination. See Supplemental Material_Measurement_1 for details.

The first part of the survey was the informed consent section, which aimed to ensure that participants had sufficient knowledge of the study’s goals and procedures before agreeing to participate. The second section gathered demographic and general information about the participants. The third section assessed vaccine hesitancy status after the adjustment of strategy. The fourth section focused on participants’ mental health and included evaluations of depressive and anxiety symptoms, benefit-finding from the epidemic, and fear of COVID-19. Finally, the fifth section aimed to investigate the psychological precursors of COVID-19 vaccination.

Participants were requested to provide sociodemographic data, including gender, age, area of residence, marital status, religious beliefs, education level, medical background, health status, SARS-CoV-2 infection, COVID-19 vaccination status, and frequency of influenza vaccination in the past three years, using closed-ended questions with provided response options.

Vaccine hesitancy was evaluated based on the WHO’s definition and relevant literature,^24,38 and a question asking participants to choose their genuine willingness to vaccinate based on their actual situation. Participants who selected any of the first five items (complete refusal, inclined to refuse but unsure, wait and see/hold back/delay, partially inclined to be willing, or willing to be keen but unsure) were considered vaccine hesitancy.

Anxiety symptoms were evaluated using the Chinese version of the Generalized Anxiety Disorder Assessment (GAD-7), which includes a total of seven items. Participants were asked to report the frequency of each symptom in the past two weeks using a four-point scale ranging from “not at all” to “almost every day.” The GAD-7 total score for the seven items ranges from 0 to 21. Scores of 5, 10, and 15 represent cut points for mild, moderate, and severe anxiety, respectively.³⁹ The Chinese version of the GAD-7 has been validated and showed good reliability and validity.⁴⁰ In this study, Cronbach’s α of GAD-7 was 0.940.

The assessment of depressive symptoms was conducted using the Patient Health Questionnaire (PHQ-9), a nine-item instrument for depression self-assessment.⁴¹ Participants were asked to report the frequency of nine problems in the past two weeks on a four-point scale ranging from “not at all” to “almost every day.” Scores ranged from 0 to 27, with severity ratings of 5–9 for mild, 10–14 for moderate, 15–19 for moderately severe, and 20–27 for severe.⁴² The Chinese version of the PHQ-9 has good reliability.⁴³ In this study, Cronbach’s α of PHQ-9 was 0.945.

The benefit was assessed using the Benefit-Finding from Epidemic Scale (BFES), developed by Chinese scholars to evaluate the sense of honor of the general Chinese population in response to an epidemic.⁴⁴ The scale consists of 10 items, including health consciousness, family value, meaning in life, and social connectedness, assessed using a 7-point Likert scale. Scores ranged from 10 to 70, with higher scores indicating higher benefit levels. The reliability and validity of the scale have been validated,⁴⁵ and in this study, Cronbach’s α of BFES was 0.87.

Fear of COVID-19 was assessed using the Chinese version of the Fear of Coronavirus Disease 2019 Scale (FCV-19-C).⁴⁶ The FCV-19-C questionnaire has been used and validated in different countries⁴⁷ and among different groups of people, such as teachers, students,⁴⁸ and patients with mental illness.⁴⁹ At the same time, the scale has been developed into a Chinese version, showing good construct validity and high internal consistency. The scale consists of seven items, and scores range from 7 to 35, with higher scores indicating greater fear of COVID-19.⁵⁰ In this study, the Cronbach’s α of the FCV-19-C was 0.927.

The psychological antecedents of the COVID-19 vaccination were assessed using the 5C scale, which had been previously authorized in China and demonstrated good reliability in previous research conducted by our team.⁵¹ The scale comprised 14 items and measured five constructs: self-confidence, complacency, constraint, calculation, and collective responsibility. Participants rated each item on a 7-point Likert scale, with higher scores indicating more excellent agreement with the corresponding construct. In our sample, Cronbach’s alpha values for the 5C scale ranged from 0.678 to 0.893, indicating acceptable internal consistency. Although the original scale was developed to gauge vaccination behavior across various contexts, we adapted it to evaluate COVID-19 vaccine uptake by incorporating additional prompts to precede the relevant section of the questionnaire. See Supplemental Material_Measurement_2 for details.

Statistical analysis

Before conducting the analysis, the data were cleaned to ensure their quality. Descriptive statistics were employed to summarize the sociodemographic characteristics of the participants, as well as their levels of vaccine hesitancy and mental health status. Spearman’s rank-order correlation coefficient was employed to conduct correlation analyses, exploring the associations among sociodemographic variables, vaccine hesitancy, anxiety, depression, benefit-finding, and fear. The subsequent pathway analyses included only those demographic factors significantly associated with vaccine hesitancy and mental health status.

To explore the model fit, pathway coefficients, the relationship between vaccine hesitancy and mental health factors, and the mediating role of vaccine hesitancy, four structural equation models (SEMs) were developed. Each model accounted for several independent variables, including gender, area of residence, religious beliefs, education level, health status, and vaccination status for COVID-19 and influenza. To evaluate the initial model fit, maximum likelihood estimation was utilized, and three indices were employed: root mean square error of approximation (RMSEA), comparative appropriate index (CFI), and standardized root mean residual (SRMR). An acceptable model fit was determined using cutoffs of RMSEA < 0.08, CFI > 0.90, and SRMR < 0.06, according to source.⁵² Path coefficients were analyzed to explore sociodemographic variables’ direct and indirect effects on anxiety, depression, benefit-finding, and fear levels. The Sobel test was employed to examine the mediating role of vaccine hesitancy.⁵³ Additionally, the RIT (Indirect effect/Total effect) and RID (Indirect effect/Direct effect) methods were used to estimate the proportion of mediating results explained by vaccine hesitancy.

The normality of the data obtained from the 5C scale was assessed and found that the scores were slightly skewed. Therefore, the median and interquartile range (IQR) were used to describe these data. The Wilcoxon rank-sum test was used to compare the differences in the 5C dimensions between different groups. All data cleaning and analysis were performed using SPSS (version 26.0) and Stata (version 17.0), and P values < .05 (two-tailed) were considered statistically significant.

Ethics statement

This study was executed strictly according to the ethical principles of the Declaration of Helsinki. The study protocol received approval from the Public Health and Nursing Research Ethics Committee of Shanghai Jiao Tong University School of Medicine, with the reference number SJUPN-202018.

Results

Characteristics of the study sample

Table 1 presents data on the study sample, which comprised 658 individuals aged 60 and older who met the inclusion criteria. The sample was evenly split by gender, with 50.0% male and 50.0% female participants (n = 329). Most participants were between 60 and 74 years old (83.7%), 38.91% were from Eastern, and 36.0% were from central regions of China. Most participants were married (87.8%), had no religious beliefs (87.5%), and had less than a high school education (62.2%). Additionally, 74.2% of the participants reported no personal or family medical background. A little over half of the participants (51.1%) reported no significant change in their health status. 18.7% of the participants or their family members (71.4%) had contracted SARS-CoV-2; less than half (42.1%) had received essential COVID-19 vaccines but not the booster shot. Additionally, over one-third (36.9%) did not take an influenza vaccine in the past three years, and more than half (52.7%) of the participants showed vaccine hesitancy toward COVID-19 vaccination.

Table 1. Sample characteristics of participants (n = 658).

Characteristics		Number (%)	Mean (SD)
Gender
	male	329 (50.0)
	female	329 (50.0)
Age(continuous)			68.36 (6.29)
Age(years)
	60–74	551 (83.7)
	>75	107 (16.3)
Area of residence
	Eastern	256 (38.9)
	Midlands	237 (36.0)
	Western	165 (25.1)
Marital status
	Unmarried	10 (1.5)
	Married	578 (87.8)
	Divorced/widowed	70 (10.6)
Religious beliefs
	No	576 (87.5)
	Yes	82 (12.5)
Education level
	Less than high school education	409 (62.2)
	High school diploma	146 (22.2)
	College degree or more	103 (15.7)
Medical background
	I had medical background	26 (4.0)
	My family had medical background	144 (21.9)
	none	488 (74.2)
Health status
	Better	99 (15.1)
	Merely changed	336 (51.1)
	Worse	223 (33.9)
SARS-CoV-2 infection
	I got infected	123 (18.7)
	At least one of my family got infected	470 (71.4)
	Nobody got infected in my family	65 (9.9)
COVID-19 vaccination
	Unvaccinated	71 (10.8)
	Basic doses	277 (42.1)
	Booster doses	310 (47.1)
Frequency of influenza vaccination in the past 3 years
	Every year	179 (27.2)
	Once/twice vaccinated	236 (35.9)
	Never	243 (36.9)
COVID-19 vaccine hesitancy
	Yes	347 (52.7)
	No	311 (47.3)
Anxiety(continuous)			3.52 (4.64)
Anxiety
	None	483 (73.4)
	Mild	105 (16.0)
	Moderate	45 (6.8)
	Severe	25 (3.8)
Depression(continuous)			3.82 (5.43)
Depression
	None	476 (72.3)
	Mild	109 (16.6)
	Moderate	31 (4.7)
	Moderately severe	23 (3.5)
	Severe	19 (2.9)
Benefit-finding(continuous)			5.70 (1.13)
Health consciousness			6.15 (1.26)
Family value			6.02 (1.25)
Meaning in life			4.78 (2.08)
Social connectedness			6.12 (1.25)
Fear(continuous)			17.77 (8.85)
5C scale
Confidence			5.51 (1.50)
Complacency			3.21 (1.47)
Constraint			2.69 (1.61)
Calculation			4.85 (2.09)
Collective responsibility			5.30 (1.58)

Standard Deviation (SD).

The mean scores of the GAD-7 and PHQ-9 were 3.52 (SD 4.64) and 3.82 (SD 5.43), respectively. A total of 3.8% of the participants exhibited severe anxiety, while 16.0% reported mild anxiety and 6.8% moderate anxiety. Additionally, there were 109 cases (16.6%) of mild depression, 31 cases (4.7%) of moderate depression, 23 cases (3.6%) of moderate to severe depression, and 19 cases (2.9%) of severe depression. The mean score of the BFES was 5.70 (SD 1.13), with the four dimensions being health consciousness (mean 6.15; SD 1.26), family value (mean 6.02; SD 1.25), meaning in life (mean 4.78; SD 2.08), and social connectedness (mean 6.12; SD 1.25). The fear of COVID-19 found an overall low score of 17.77 (SD 8.85). The five dimensions of the antecedent psychological description of vaccination were confidence (mean 5.51; SD 1.50), complacency (mean 3.21; SD 1.47), constraint (mean 2.69; SD 1.61), calculation (mean 4.85; SD 2.09), and collective responsibility (mean 5.30; SD 1.58).

Correlations among demographics, vaccine hesitancy, and mental health

Correlation analysis

Table 2 presents the results of Spearman’s correlation analysis, which revealed a positive association (P < .05) between vaccine hesitancy and almost all socio-demographic variables (e.g., gender, area of residence, education level, health status, and vaccination status) as well as psychological variables (e.g., anxiety, depression, benefit-findings, and fear).

Table 2. Correlations between variables of interest (n = 658).

	Gender	Area of residence	Religious beliefs	Education level	Health status	COVID-19 vaccination	Influenza vaccination	COVID-19 vaccine hesitancy	Anxiety	Depression	Benefit-finding	Fear
Gender	1
Area of residence	−.143**	1
Religious beliefs	.092*	−.124**	1
Education level	−.023	−.458**	.084*	1
Health status	.040	−.069	−.006	.036	1
COVID-19 vaccination	−.123**	.103**	−.117**	−.009	−.008	1
Influenza vaccination	.091*	−.476**	.030	.243**	.065	−.322**	1
COVID-19 vaccine hesitancy	−.033	.240**	−.090*	−.196**	−.117**	.221**	−.150**	1
Anxiety	−.004	−.224**	−.104**	.095*	−.066	.043	.273**	.194**	1
Depression	.115**	−.476**	.083*	.213**	.082*	−.134**	.386**	−.126**	.811**	1
Benefit-finding	.103**	−.312**	.145**	.201**	.135**	−.084*	.201**	−.168**	−.017	−0.07	1
Fear	.104**	−.351**	.151**	.219**	.157**	−.070	.217**	−.226**	.613**	.577**	.209**	1

Table shows the Spearman’s correlation coefficient, *P < .05 **P < .01.

Factors of anxiety, depression, benefit-finding, fear, and the pathway

Figure 1 illustrates the proposed model for the mediation effect of vaccine hesitancy between demographics and anxiety symptoms. The model fit was excellent (RMSEA = 0.000; CFI = 1.000; SRMR = 0.000). After controlling for demographic factors, the area of residence (with “Eastern” as the reference) (β = 0.171, P < .001), education level (with “Less than high school” as the reference) (β = −0.103, P < .05), health status (with “Better” as the reference) (β = −0.099, P < .01), and COVID-19 vaccination status (with “Unvaccinated” as the reference) (β = 0.222, P < .001) were significantly associated with vaccine hesitancy. Moreover, vaccine hesitancy was mainly related to anxiety symptoms (β = −0.086, P < .05).

Figure 1. Mediation effect of COVID-19 vaccine hesitancy between demographics and anxiety symptoms.

Note. RMSEA (Root mean square error of approximation) is an index used to evaluate model misfit. CFI (Comparative Fit Index) is used to measure the consistency between the model and the original data. SRMR (standardized root mean square residual) value measures the residual mean between observed variables and predictor variables in the model. RMSEA ≤ 0.05, CFI ≥ 0.95 and SRMR ≤ 0.08 represent an excellent fitting model.

Figure 2 depicts the model for the mediation effect of vaccine hesitancy between demographics and depression symptoms. The model fit was excellent (RMSEA = 0.000; CFI = 1.000; SRMR = 0.000). After adjusting for demographic factors, the area of residence (with “Eastern” as the reference) (β = 0.171, P < .001), education level (with “Less than high school” as the reference) (β = −0.103, P < .05), health status (with “Better” as the reference) (β = −0.099, P < .01), and COVID-19 vaccination status (with “Unvaccinated” as the reference) (β = 0.222, P < .001) were significantly associated with vaccine hesitancy. Vaccine hesitancy was positively linked to depression symptoms (β = −0.134, P < .001).

Figure 2. Mediation effect of COVID-19 vaccine hesitancy between demographics and depression symptoms.

Figure 3 illustrates the proposed model for the mediation effect of vaccine hesitancy between demographics and benefit-finding. The model fit was good (RMSEA = 0.000; CFI = 1.000; SRMR = 0.000). After adjusting for demographic factors, including the area of residence (with “Eastern” as the reference) (β = 0.171, P < .001), education level (with “Less than high school” as the reference) (β = −0.103, P < .05), health status (with “Better” as the reference) (β = −0.099, P < .01), and COVID-19 vaccination status (with “Unvaccinated” as the reference) (β = 0.222, P < .001), vaccine hesitancy was significantly associated with these demographics. Specifically, vaccine hesitancy was positively associated with benefit-finding (β = 0.239, P < .001).

Figure 3. Mediation effect of COVID-19 vaccine hesitancy between demographics and benefit-finding.

Figure 4 depicts the proposed model for the mediation effect of vaccine hesitancy between demographics and fear. The model fit was excellent (RMSEA = 0.000; CFI = 1.000; SRMR = 0.000). After adjusting for demographic factors, with “Eastern” as the reference for area of residence (β = 0.171, P < .001), “Less than high school” as the reference for education level (β = −0.103, P < .05), “Better” as the reference for health status (β = −0.099, P < .01), and “Unvaccinated” as the reference for COVID-19 vaccination (β = 0.222, P < .001), significant associations with vaccine hesitancy were observed. Nevertheless, it’s worth noting that the interaction effect of vaccine hesitancy concerning demographics and fear did not reach statistical significance.

Figure 4. Mediation effect of COVID-19 vaccine hesitancy between demographics and fear.

Mediating mechanisms for anxiety, depression, and benefit-finding factors

In the simulation study, the Sobel method’s test power was higher than the sequential regression coefficient method.⁵⁴ Therefore, we performed the Sobel Test to confirm the mediation effects. Table 3 presents the mechanisms underlying the correlation between the area of residence, education level, health status, COVID-19 vaccination, anxiety, depression, and benefit-finding.

Table 3. Mediation effect of COVID-19 vaccine hesitancy between demographics and mental health.

		X→M	M→Y	X→Y	Sobel Test	RIT	RID
Anxiety	Area of residence	B = 0.171***	B = −0.086*	B = 0.196***	B = −0.127*	0.091	0.100
	Education Level	B = −0.103^*	B = −0.086*	B = 0.012	B = 0.148**	0.176	0.214
	Health Status	B = −0.100^**	B = −0.086*	B = 0.092*	B = 0.118*	0.141	0.164
	COVID-19 Vaccination	B = 0.223^***	B = −0.086*	B = 0.023	B = −0.249**	0.695	2.276
Depression	Area of residence	B = 0.171***	B = −0.135***	B = 0.215***	B = −0.023*	0.118	0.134
	Education Level	B = −0.103^*	B = −0.135***	B = 0.004	B = 0.014*	0.218	0.278
	Health Status	B = −0.100^**	B = −0.135***	B = 0.092*	B = 0.013*	0.179	0.218
	COVID-19 Vaccination	B = 0.223^***	B = −0.135***	B = 0.047	B = −0.030**	1.016	60.846
Benefit-Finding	Area of residence	B = 0.171***	B = 0.239^***	B = −0.169^***	B = 0.813^***	0.317	0.241
	Education Level	B = −0.103^*	B = 0.239^***	B = −0.009^***	B = −0.577^***	0.548	0.354
	Health Status	B = −0.100^**	B = 0.239^*	B = −0.042	B = −0.396^***	0.584	1.406
	COVID-19 Vaccination	B = 0.223^***	B = 0.239^***	B = 0.089^*	B = 0.794^***	0.983	58.29

RIT = Ratio of indirect effect to total effect; RID = Ratio of indirect effect to direct effect. X refers to the independent variables included in this mediation model, including geography, education level, health status, and COVID-19 vaccination; M represents the mediating variable COVID-19 vaccine hesitancy; and Y is the dependent variable in the mediation model, including anxiety, depression, and benefit-finding. The effect of X on Y in the mediation model is divided into two parts: one is the direct effect, the direct effect of X on Y, and the other is the indirect effect, the effect of X on Y through M.

The RIT results in Table 3 demonstrated that vaccine hesitancy suppressed 9.1% of the effect of region of residence on anxiety (B = −0.127, P < .05). Additionally, vaccine hesitancy played a mediating in role about 17.6% and 14.1% of the effects of education level (B = 0.148, P < .01) and health status (B = 0.118, P < .05) on anxiety, respectively. Vaccine hesitancy completely mediated the effect of the COVID-19 vaccination on anxiety (B = −0.249, P < .01), which means that an individual’s anxiety level is not directly affected by whether or not they get the COVID-19 vaccination; rather, vaccine hesitancy does.

Specifically, 11.8% of the total effect of area of residence on depression was suppressed by vaccine hesitancy (B = −0.023, P < .05). Additionally, about 21.8% and 17.9% of the effects of education level (B = 0.014, P < .05) and health status (B = 0.013, P < .05) on depression, respectively, were mediated by vaccine hesitancy. Finally, vaccine hesitancy showed a complete mediation role in the effects of COVID-19 vaccination on depression (B = −0.030, P < .01).

Vaccine hesitancy was also found to moderate the effects of the factors on benefit-finding. The findings indicate that vaccine hesitancy mediated 31.7% of the total effect of area of residence (B = 0.813, P < .001) and 98.3% of the impact of COVID-19 vaccination on benefit-finding (B = 0.794, P < .001). About 54.8% of the effect of education level on benefit-finding was depressed by vaccine hesitancy (B = −0.577, P < .001). Specifically, vaccine hesitancy showed a full mediation effect between health status and benefit-finding (B = −0.396, P < .01), as the direct effect of health status on benefit-finding became non-significant when the mediator (vaccine hesitancy) was added.

Psychological antecedents of the COVID-19 vaccine hesitancy

Table 4 describes the group differences in the psychological antecedents of different COVID-19 vaccine hesitancy states. The study found that older adults without vaccine hesitancy showed significantly higher levels of confidence (Z = 14.69, P < .001), lower levels of complacency (Z = 7.03, P < .001), lower levels of constraint (Z = 7.77, P < .001), and higher levels of collective responsibility (Z = 8.85, P < .001) than those with vaccine hesitancy. However, no statistically significant difference was found in the calculation dimension (P = .644).

Table 4. Total score of 5C and score of each dimension in participants (n = 658).

Dimensions	Median (IQR)	Vaccine Hesitancy		Wilcoxon test (Mann-Whitney)
		Yes (n = 347)	No (n = 311)	Z-value	P-value
Confidence	5.83 (2.33)	5.00 (2.00)	7.00 (1.00)	14.69	<.001
Complacency	3.25 (1.75)	3.50 (1.50)	2.50 (2.00)	7.03	<.001
Constraint	2.67 (2.67)	3.00 (2.33)	1.67 (2.00)	7.77	<.001
Calculation	5.33 (3.67)	5.00 (2.67)	5.33 (4.00)	0.47	.644
Collective responsibility	5.50 (3.00)	5.00 (2.00)	6.50 (2.00)	8.85	<.001

IQR (Inter Quartile Range) = High quartile(Q3) – low quartile(Q1).

Discussion

The purpose of this study was to explore the relationship between sociodemographic factors, vaccine hesitancy, and mental health outcomes in the elderly population. Through a survey of 658 respondents after adjustment of the dynamic zero-COVID-19 strategy in eastern, central, and western China, this study found that vaccine hesitancy served as a mediator/suppressor of the association between sociodemographic variables and symptoms of anxiety, depression, and benefit-finding.

Evidence has shown that vaccine hesitancy is positively related to anxiety and depression symptoms.^32,53 Addressing the mediation effect, the study revealed that the positive contribution of education level to anxiety and depression in older populations is not fully realized and is entirely dependent on the mediating effect of vaccine hesitancy. Therefore, higher education levels are associated with higher anxiety and depression because they are less willing to get vaccinated. Vaccine hesitancy is a partial mediator of health status, anxiety, and depression (14.1% and 17.9% of the total effect, respectively). This shows that the elderly with poorer self-rated health status had more anxiety and depression, primarily because they were less willing to get vaccinated. Between residency and anxiety/depression, vaccine hesitancy was only partially suppressed (9.1% and 11.8% of the total impact, respectively). This implies that older adults in the eastern region were less likely to get immunized, which increased anxiety and depression. Our study found that the generation of negative inhibition of anxiety and depression by vaccination status is not fully realized, and it is entirely dependent on the mediating effect of vaccine hesitancy. Thus, the elderly who had not yet received the COVID-19 vaccine were less motivated to do so, but reported less anxiety and depression following vaccination.

Several studies have additionally demonstrated a negative correlation between vaccine hesitancy and the benefit-finding of the COVID-19 epidemic.^44,55 Regarding the mediation effect, it was discovered that vaccine hesitancy partially moderated the relationship between residency and benefit-finding (31.8% of the total impact). This shows that older adults in the East were more hesitant to vaccinate, and this was associated with lower benefit-finding when an epidemic stuck. Additionally, vaccine hesitancy partially moderated the effect of COVID-19 vaccination status on the benefit-finding (98.3% of the total effect), as older adults who were not vaccinated were more hesitant to vaccinate but felt more epidemic-specific benefit after vaccination. Concerning the suppression effect, vaccine hesitancy partially suppressed the effect of education level on benefit-finding (54.8% of the total effect), indicating that higher-educated people exhibited less benefit-finding due to their increased hesitation toward vaccination. Also, the effect of health status on anxiety and depression was completely suppressed by vaccine hesitancy, suggesting that the suppression of benefit-finding in the elderly by health status is not fully realized and depends entirely on vaccine hesitancy. That is, older adults with poorer self-rated health status had lower benefit-finding, primarily due to their stronger hesitancy to vaccinate.

In this study, as in previous ones,^56,57 vaccine hesitancy among the elderly was negatively associated with fear of COVID-19, although it did not mediate fear of SARS-CoV-2. This lack of mediation may be due to the Chinese people’s continued confidence in their government. The public trusts the government’s ability to control the epidemic. It has implemented numerous active and effective measures to contain it, including city closures, extensive testing, and rapid vaccine development.

In summary, vaccine hesitancy may partially or fully mediate/suppress the effects of sociodemographic factors on mental health. This finding highlights the crucial role of vaccine hesitancy in shaping older people’s mental health following vaccination. Additionally, the study found that older adults residing in the eastern region of China, those with lower levels of education and poorer health status, and those who have not received the COVID-19 vaccine were more likely to exhibit vaccine hesitancy. This observation is broadly consistent with previous research conducted during different stages of the COVID-19 pandemic.^20,58

The study found that vaccine hesitancy was more prevalent among individuals living in developed eastern regions and those with higher levels of education. This finding is consistent with a previous national cross-sectional study in China, which reported that vaccine hesitancy was more prevalent among well-educated and urban-dwelling older adults.⁵⁹ Similar trends have been observed in other countries, with higher socioeconomic status associated with lower influenza vaccination rates.^60,61 Individuals’ higher levels of education and wealth are associated with exposure to various sources of information, which may contribute to information overload and susceptibility to misinformation, leading to vaccine hesitancy.⁶² However, it is undeniable that there is no conclusive evidence about the effect of education level on vaccine hesitancy. According to the study conducted by Kricorian et al.,⁶³ individuals with lower levels of educational have a lower propensity to vaccinate, which may be due to their concerns about the presentation of complicated data or extensive textual content within health-related information. A study conducted in China found a similar results.⁶⁴ Nevertheless, we believe that education can be a critical factor in facilitating the acquisition of accurate knowledge about vaccines among individuals who lack information and formal education. In addition, highly educated populations now have greater access to health information and vaccine-related knowledge through the Internet and other media. However, the spread of misinformation, which is increasingly disseminated through social media platforms and other channels, poses a significant threat to public health. Therefore, these new media should be subject to stricter rules so that they can serve as a link between immunization service providers and the public, disseminate official information about vaccines, and dispel misunderstandings and doubts about vaccination in society.

Furthermore, the study found vaccine hesitancy was more prevalent among older adults with self-reported poor health status and those who had never taken the COVID-19 vaccine. This is supported by findings from a study in Shanghai, China, where overall booster vaccination rates were lower among 39,498 patients aged 60 years and older with diabetes than among those aged 60 years and younger.⁶⁵ Immunization of the elderly in China with chronic diseases has not progressed well. Some older adults choose not to be vaccinated due to their health concerns or the possible adverse effects of vaccines on chronic diseases.⁶⁶ Moreover, although China has been vaccinating older adults with chronic conditions since the end of 2021, some vaccination facilities still refuse to vaccinate older adults with potentially life-threatening diseases, undermining trust in vaccines.⁶⁷

The results of our study highlight the urgent need for immediate vaccination efforts and outreach targeting older adults, particularly those living in economically developed areas with higher levels of education, poorer health status, and those who have never received the COVID-19 vaccine for various reasons. With infection rates already high in China’s large cities and the epidemic spreading to smaller towns and vast rural areas, time is of the essence. The National Health Commission in China has outlined specific strategies to increase vaccination rates among the elderly, including launching a second booster vaccination for those aged 60 years and older on December 14, 2022.⁶⁸ To effectively implement these strategies, China must work to rebuild confidence in the COVID-19 vaccine and combat misinformation through transparency. The results of our study hold significant ramifications for health promotion initiatives, as the widespread prevalence of vaccine hesitancy points to a considerable shortfall in health education efforts aimed at mitigating vaccine hesitancy. Furthermore, our research reveals a positive correlation between vaccine hesitancy and symptoms of anxiety/depression, along with a negative association with benefit-finding, thereby shedding light on the potential mental health implications for older adults who exhibit vaccine hesitancy. As a result, public health authorities must remain aware of the mental health ramifications of vaccine hesitancy and prioritize mental health support for individuals in the aftermath of large-scale vaccination campaigns.

Our study identified lower confidence, collective responsibility, and higher levels of complacency and constraint among older adults who exhibited vaccine hesitancy. As a result, the 5C model can be used as a framework for health promotion in practice. For instance, communities should prioritize promoting and educating older adults on COVID-19 vaccination by considering their physical and mental characteristics and preferred information access, providing sufficient vaccine information to make informed decisions, enhancing their vaccine confidence, reducing complacency, and addressing other potential barriers. Moreover, special vaccination channels for older people can be created, and the vaccination process can be optimized with peer-to-peer vaccination education that considers the collective responsibility characteristics of this group, ultimately helping to increase the COVID-19 vaccination rate and reduce vaccine hesitancy. Understanding these psychological factors is crucial in developing effective strategies to promote vaccine acceptance and increase vaccination rates. By comprehending the underlying reasons for vaccine hesitancy in older adults, healthcare providers and public health officials can tailor their communication and outreach efforts to address these factors directly, leading to a more successful vaccination campaign.

Nevertheless, it is critical to acknowledge several limitations of this study. First, health status and other variables were calculated based on self-report rather than precise laboratory tests, and the results may be susceptible to recall bias. Second, the infection rate of 18.7% for SARS-CoV-2 among the elderly was lower than that of subway passengers in Beijing (75.7–92.3%). This may be because the elderly tend to have fewer social activities and are more likely to be protected by family members. Moreover, convenience sampling methods may introduce sample selection bias, which could compromise the representativeness of the findings. While we used a sizable sample size and calculated the sample size according to the incidence rate to ensure that the questionnaire accurately represented the subject of the study, the possibility that the sample size could have influenced the results could not be ruled out entirely. In addition, the research encompassed an elderly cohort characterized by varying degrees of education, potentially resulting in challenges pertaining to the comprehension of certain responses to the questionnaire. Diverse strategies were implemented to alleviate this issue, encompassing the provision of comprehensive explanations, supplementary resources, and personalized support. But variations in educational attainment might have contributed to disparities in the interpretation of specific inquiries, potentially introducing a degree of bias into the data gathering process. An additional limitation is that the cross-sectional study design could not determine the causal relationship between sociodemographic characteristics, vaccine hesitancy, and symptoms of anxiety, depression, and benefit-finding.⁶⁹ Moreover, the present investigation evaluated vaccine hesitancy and mental health status only at a single point, thereby failing to account for potential fluctuations in these variables over time. Longitudinal research designs should be used in future studies to find out what causes these factors and how vaccine hesitancy and mental health status change over time.

Conclusions

In conclusion, this study was conducted within two months of the adjustment of the zero-COVID-19 strategy in China and aimed to explore the association between sociodemographic characteristics and mental health among older adults, as well as the mediating/suppressing effects of vaccine hesitancy. Specifically, older adults living in the East with higher education levels, poorer health statuses, and unvaccinated were more likely to experience anxiety, depression, and lower benefit-finding in response to the epidemic, possibly due to vaccine hesitancy. This study highlights potential barriers to vaccine uptake among older adults and provides evidence-based recommendations to improve vaccine confidence and mental health outcomes. With a deeper understanding of the psychological factors driving vaccine hesitancy, public health professionals can develop effective interventions and communication strategies tailored to the unique needs of older adults, ultimately improving the overall health outcomes of this vulnerable population.

Disclosure statement

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

Data availability statement

Data will be made available on request.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2023.2288726.

Additional information

Funding

This study was supported by the 3-year action plan for the construction of Shanghai's public health system (2020-2022), academic leaders cultivating project [grant number GWV-10.2-XD33], an Innovative research team of high-level local universities in Shanghai [grant number SHSMU-ZDCX20212801].