Global Prevalence and Potential Influencing Factors of COVID-19 Vaccination Hesitancy: A Meta-Analysis

Countries worldwide have deployed mass COVID-19 vaccination drives, but there are people who are hesitant to receive the vaccine. Studies assessing the factors associated with COVID-19 vaccination hesitancy are inconclusive. This study aimed to assess the global prevalence of COVID-19 vaccination hesitancy and determine the potential factors associated with such hesitancy. We performed an organized search for relevant articles in PubMed, Scopus, and Web of Science. Extraction of the required information was performed for each study. A single-arm meta-analysis was performed to determine the global prevalence of COVID-19 vaccination hesitancy; the potential factors related to vaccine hesitancy were analyzed using a Z-test. A total of 56 articles were included in our analysis. We found that the global prevalence of COVID-19 vaccination hesitancy was 25%. Being a woman, being a 50-year-old or younger, being single, being unemployed, living in a household with five or more individuals, having an educational attainment lower than an undergraduate degree, having a non-healthcare-related job and considering COVID-19 vaccines to be unsafe were associated with a higher risk of vaccination hesitancy. In contrast, living with children at home, maintaining physical distancing norms, having ever tested for COVID-19, and having a history of influenza vaccination in the past few years were associated with a lower risk of hesitancy to COVID-19 vaccination. Our study provides valuable information on COVID-19 vaccination hesitancy, and we recommend special interventions in the sub-populations with increased risk to reduce COVID-19 vaccine hesitancy.


Introduction
Coronavirus disease 2019 (COVID- 19) vaccination has been progressing globally since the beginning of 2021. Several types of vaccines, including inactivated, vector-based, messenger ribonucleic acid (mRNA), and protein subunit vaccines, are being administered to recipients [1]. Since the vaccines became available, there have been expectations of the COVID-19 pandemic ending, considering that previous vaccination programs have been effective in managing several infectious diseases such as rubella, mumps, measles, and polio. These vaccination programs have been proven to improve global health and the economy [2,3]. However, the probability of failure of any vaccination program should be assessed. A study reported that the barriers to effective vaccination programs include inconvenient and limited clinic hours for immunization, inadequate access to healthcare, high vaccine administration fees, and vaccine hesitancy [4]. Of these factors, vaccine hesitancy is considered one of the most critical [5]. Individuals who are hesitant to be immunized have a tendency to spread incorrect information about vaccination, which may influence people close to them to reject vaccines as well [6].
Vaccine hesitancy is commonly observed in the case of new vaccines or vaccine candidates [7,8]. This phenomenon was reported in the case of malaria [9], dengue [10], and Ebola [11]. The factors contributing to vaccine hesitancy are complex and may include a lack of awareness regarding disease prevention and socioeconomic status [12,13]. This phenomenon poses a dilemma to vaccine coverage. Moreover, governments-as the highest regulatory authority of any nation-seemingly do not provide special interventions to reduce hesitancy toward vaccination programs. It is observed in the guidelines on COVID-19 vaccination, that the primary recommendation only focused on dose allocation, outreach, delivery, and monitoring; there was no information on how to reduce COVID-19 vaccination hesitancy [14].
Regarding COVID-19 vaccination, several studies have been conducted to assess the prevalence of COVID-19 vaccine hesitancy and its associated predictors [7,8,15]. However, the findings were inconclusive with variability regarding the correlation between COVID-19 vaccine acceptance and the following: sociodemographic factors, vaccine confidence and trust regarding vaccine safety, complacency towards the disease, conspiracy beliefs towards COVID-19 vaccination and willingness to pay for the vaccine [7,[16][17][18][19][20][21]. Therefore, a meta-analysis is necessary to determine the potential factors influencing COVID-19 vaccination hesitancy.

Study Design
During the period May-June 2022, we conducted a meta-analysis that followed the protocols of the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) [22,23]. In line with the purpose of our study, we first performed an organized search of PubMed, Scopus, and Web of Science and, subsequently, collected the required information to calculate the global prevalence of vaccine hesitancy and effect estimates of the potential influencing factors. The PRISMA checklist for this review is provided in (Supplementary Materials). Additionally, data used in this review are available in Figshare (https://doi.org/10.6084/m9.figshare.20055539.v3, accessed on 6 December 2022) [23].

Eligibility Criteria
We determined the eligibility criteria before conducting the organized search. An article was included in the analysis if the following inclusion criteria were met: (1) whether it assessed the prevalence of COVID-19 vaccination hesitancy or (2) identified potential factors influencing COVID-19 vaccination hesitancy. Reviews, commentaries, letters to the editor, grey literature, and double publications were excluded.

Search Strategy and Data Extraction
As of 25 May 2022, we performed an organized search of PubMed, Scopus, and Web of Science. Prior to the search for the main outcomes, the potential factors associated with COVID-19 vaccination hesitancy were determined. We used keywords from the following medical subject headings: "vaccine", "vaccination", or "immunization"; "COVID-19" or "coronavirus disease 2019"; "hesitancy" or "acceptance". We limited the organized search to the English language. If we found any duplication, we included the studies with larger sample sizes. Furthermore, we also conducted an organized search of the reference lists of the relevant articles to obtain additional papers. Thereupon, the following information was collected from the selected articles: (1) first author name, (2) year of publication, (3) study design, (4) study period, (5) Newcastle-Ottawa scale (NOS), (6) the prevalence of COVID-19 vaccination hesitancy, and (7) event rate of potential factors associated with COVID-19 vaccination hesitancy. Two independent teams, led by JKF and SAPK, conducted the article search and data extraction. Prior to the systematic search, the kappa statistic was used to measure the agreement between the two investigators. If the kappa statistic was greater than the p-value, agreement was established.

Assessment of the Methodological Quality
All potential articles for inclusion in the study were assessed for quality using NOS [24]. The quality was considered high, moderate, or low if the score was 7-9, 4-6, or 0-3, respectively. Low-quality articles were excluded from the analysis. Using a pilot form, the Vaccines 2022, 10, 1356 4 of 20 two independent teams, led by JKF and SAPK, conducted the NOS assessment, and any discrepancies were resolved through discussion.

Outcome Measures
The major outcomes were global prevalence and potential influencing factors of COVID-19 vaccination hesitancy. To identify the potential factors associated with vaccine hesitancy, we performed an initial organized search in PubMed, Scopus, and Web of Science. We identified the following potential factors: age group, gender, marital status, educational attainment, religion, employment status, healthcare-related job, socioeconomic status, urbanity, presence of children and elderly people at home, household size, and presence of family members with a medical background. Additionally, wearing masks, hand hygiene, compliance with physical distancing norms, smoking, history of chronic disease, personal history of COVID-19 diagnosis, COVID-19 diagnosis of a family member/friend, hospitalization due to COVID-19 among people in the same social circle, death owing to COVID-19 among people in the same social circle, safety conceptions about COVID-19 vaccines, and history of influenza vaccination in the past few years were also factors of interest.

Statistical Analysis
Before calculating the global prevalence of COVID-19 vaccination hesitancy and effect estimates of potential predictors of such hesitancy, we conducted an analysis of potential publication bias and heterogeneity among the studies. We analyzed the risk of publication bias using the Egger's test, with a p-value of <0.05 suggesting the existence of publication bias. Furthermore, we performed an analysis of heterogeneity among studies using the Q test, with a p-value of <0.10 indicating heterogeneity; thus, a random effects model was applied for data analysis-in cases where there was no heterogeneity, a fixed-effects model was used. A single-arm meta-analysis was performed using the dichotomous covariate method to calculate the event rate from each study to discern the global prevalence of COVID-19 vaccination hesitancy. The effect estimate was presented as the event rate. The analysis was performed using the R package (RStudio version 4.1.1, R Studio, Boston, MA, USA). The effect estimates of potential factors associated with COVID-19 vaccination hesitancy were outlined in a forest plot as a pooled odds ratio and 95% confidence interval (OR, 95% CI).

Selection of Studies
We retrieved 4299 potential papers from the databases mentioned and 18 from the reference lists of related articles. Of these, 23 papers were excluded owing to duplication and 4219 papers with irrelevant subjects. Thus, 75 articles were included in the fulltext review. Subsequently, six reviews and thirteen articles were excluded because of insufficient data. Eventually, a total of 56 articles were included in the final analysis to calculate the global prevalence and potential influencing factors in COVID-19 vaccination hesitancy [19,. The plotting of article selection in our study is presented in (Figure 1), and the characteristics of the included articles are listed in (Table 1). and 4219 papers with irrelevant subjects. Thus, 75 articles were included in the fu review. Subsequently, six reviews and thirteen articles were excluded because of in cient data. Eventually, a total of 56 articles were included in the final analysis to cal the global prevalence and potential influencing factors in COVID-19 vaccination hesi [19,. The plotting of article selection in our study is presented in (Figure 1), an characteristics of the included articles are listed in (Table 1).

Potential Factors Associated with COVID-19 Vaccination Hesitancy
The potential factors associated with COVID-19 vaccination hesitancy are summarized in (Table 2) and presented in (Figures 3-6). Our analysis revealed that out of 25 factors, 12 of them were associated with COVID-19 vaccine hesitancy. The following factors were associated with a higher risk of vaccination hesitancy: being a woman (compared to man) ( Figure 3A), being ≤ 50 years old (compared to those older than 50 years ( Figure 3B), being single (compared to married individuals) ( Figure 3C), being unemployed (compared to employed) ( Figure 4A), living in a household with five or more individuals (compared to living in smaller households) ( Figure 4B), having an educational attainment

Potential Factors Associated with COVID-19 Vaccination Hesitancy
The potential factors associated with COVID-19 vaccination hesitancy are summarized in (Table 2) and presented in (Figures 3-6). Our analysis revealed that out of 25 factors, 12 of them were associated with COVID-19 vaccine hesitancy. The following factors were associated with a higher risk of vaccination hesitancy: being a woman (compared to man) ( Figure 3A), being ≤50 years old (compared to those older than 50 years ( Figure 3B), being single (compared to married individuals) ( Figure 3C), being unemployed (compared to employed) ( Figure 4A), living in a household with five or more individuals (compared to living in smaller households) ( Figure 4B), having an educational attainment lower than an undergraduate degree (compared to those with an undergraduate degree or higher) ( Figure 5A), having a non-healthcare-related job (compared to having a healthcare-related  (Figure 5B), and considering COVID-19 vaccines to be unsafe (compared to those who consider COVID-19 vaccines to be safe) ( Figure 5C).

Source of Heterogeneity and Potential Publication Bias
Heterogeneity was not found for six variables (Christian religion, household size ≥ 5 individuals, family members with a medical background, smoking, hospitalization due to COVID-19 among people in the same social circle, and death owing to COVID-19 among people in the same social circle). Therefore, we used a fixed-effects model. Conversely, other variables were assessed using a random effects model. The Egger's test was used to assess potential bias among the studies. Our pooled analyses revealed a risk of publication bias for the following covariates: the Christian religion, family members with a medical background, hospitalization due to COVID-19 among people in the same social circle, and death owing to COVID-19 among people in the same social circle (Table 2). In contrast, living with children at home (compared to having no child at home) ( Figure 6A), maintaining physical distancing norms (compared to not following such norms) ( Figure 6B), having ever tested for COVID-19 (compared to having never tested for COVID-19) (Figure 6C), and having a history of influenza vaccination in the past few years (compared to not having been vaccinated for influenza in the past few years) ( Figure 6D) were associated with a lower risk of hesitancy to COVID-19 vaccination.

Source of Heterogeneity and Potential Publication Bias
Heterogeneity was not found for six variables (Christian religion, household size ≥ 5 individuals, family members with a medical background, smoking, hospitalization due to COVID-19 among people in the same social circle, and death owing to COVID-19 among people in the same social circle). Therefore, we used a fixed-effects model. Conversely, other variables were assessed using a random effects model. The Egger's test was used to assess potential bias among the studies. Our pooled analyses revealed a risk of publication bias for the following covariates: the Christian religion, family members with a medical background, hospitalization due to COVID-19 among people in the same social circle, and death owing to COVID-19 among people in the same social circle (Table 2).

Discussion
Our study estimated the global prevalence of COVID-19 vaccination hesitancy at 25%. The current findings are consistent with those of previous meta-analyses, which estimated the prevalence of vaccine hesitancy in the general population at 26-42% [15,[80][81][82][83]. In special populations, previous meta-analyses revealed that the hesitancy for COVID-19 vaccination was estimated at 24%, 27%, and 26% in multiple sclerosis patients, older people, and healthcare students, respectively [84][85][86]. Furthermore, hesitancy to receive a COVID-19 booster was reported at 21% in the general population [87]. Moreover, high rates of COVID-19 vaccine hesitancy were reported among the ethnic minorities in the UK [88]. Our estimate was in the range of extant literature. However, our study had a larger sample size, which might have provided a more accurate calculation. Moreover, in sub-group analysis, we also identified that the prevalence of COVID-19 vaccination hesitancy was 25%, 26%, and 25% in the general population, healthcare workers, and students, respectively. This study also identified the potential predictors of COVID-19 vaccine hesitancy, thereby providing more comprehensive evidence on this phenomenon.
The current study noted that the potential factors associated with COVID-19 vaccination hesitancy can be contextualized in terms of awareness, knowledge, and socioeconomic status. In the context of awareness of COVID-19 vaccination, we found that older people (>50 years), those living with children at home, individuals who have ever tested for COVID-19, and those with a history of influenza vaccination had a lower risk of COVID-19 vaccination hesitancy. In contrast, several factors, such as single marital status and unemployment, were associated with an increased risk of hesitancy toward COVID-19 vaccination.
The precise underlying factors contributing to COVID-19 vaccination hesitancy could not be defined clearly. However, some presumptions may explain these findings. Older individuals are more likely to suffer from one or more chronic diseases compared to younger people. In our previous investigation, we found that advanced age and comorbidity were associated with an increased risk of severity in COVID-19 patients [89]. Therefore, the possibility of an increased risk of severe COVID-19 might influence the awareness of such individuals and contribute to a lower risk of COVID-19 vaccination hesitancy in this group due to low levels of complacency [90].
Similarly, individuals living with children at home might be afraid of transmitting the virus to their children should they be infected with COVID-19. Therefore, it is reasonable that this group is less hesitant to receive COVID-19 vaccination. Interestingly, a similar impact was not observed in individuals living with elderly people at home. This ironic finding is supported by previous studies, which found that living with children was a crucial determinant of health-related behavior [91], whereas this was not the case for individuals living with elderly people [92].
Furthermore, individuals who have ever tested for COVID-19 and had a history of influenza vaccination might have had good practice in disease prevention. Disease screening and vaccination history have been shown to affect health behavior, which can possibly explain why this group is less averse to COVID-19 vaccination. Moreover, married individuals might engage in protective behavior toward their spouse; couples have mutual concern and might have a better life expectancy than single individuals. A previous study found that married individuals had better health behavior and a lower risk of mortality than single individuals [93]. Thus, single individuals might be more averse than married individuals to COVID-19 vaccination. This reason, in the context of poor health behavior, might also explain vaccine hesitancy in unemployed individuals.
Our study also found that individuals with lower educational levels and those who consider COVID-19 vaccines to be unsafe had a higher risk of COVID-19 vaccination hesitancy. In contrast, individuals with healthcare-related jobs had a lower risk of COVID-19 vaccination hesitancy. The association between higher educational levels, knowledge of disease prevention, and vaccine acceptance or hesitancy has been widely investigated [16,[94][95][96]. Individuals with higher educational levels and healthcare-related jobs might have adequate information on the global pandemic and consider vaccination to be a great step toward ending the pandemic, which can explain why this group had a lower risk of vaccine hesitancy. Our current findings are supported by previous studies in the context of dengue, Ebola, and monkeypox vaccines. Those studies also showed that knowledge of disease prevention and good health practices had a significant impact on the acceptance of vaccine candidates [97][98][99][100][101].
Although we could not elucidate the role of socioeconomic status in COVID-19 vaccination hesitancy in this study, we found that some factors related to socioeconomic status, such as unemployment and household size (≥5 individuals), were associated with COVID-19 vaccination hesitancy. Socioeconomic status has been proven to affect healthrelated behavior [102]. Individuals with a low socioeconomic status might lack knowledge of the pandemic and the role of vaccination in the pandemic. Moreover, individuals with a low socioeconomic status might also lack social interaction; therefore, they might lack adequate knowledge concerning disease prevention, which could contribute to COVID-19 vaccination hesitancy. Previous meta-analyses in this context did not assess the role of socioeconomic status in COVID-19 vaccination hesitancy. However, in other settings, such as in the case of dengue vaccines, socioeconomic status was found to affect vaccination acceptance [100].
To the best of our knowledge, this meta-analysis is the first comprehensive study to assess COVID-19 vaccination hesitancy worldwide. In sub-group analysis, our study identified similar prevalence rates of hesitancy to COVID-19 vaccination in the general population, healthcare workers, and students; suggesting that interventions to reduce the risk of COVID-19 vaccination hesitancy in those populations do not need to be differentiated. In addition to reporting the global prevalence, we also identified the potential factors associated with hesitancy to receive COVID-19 vaccination. Although the COVID-19 vaccination program targets the global population, some people have been hesitant to receive the vaccine. Our study identified the factors associated with such hesitancy, thereby shedding light on the populations that require special attention in order for the vaccination program to be successful. We recommend customized interventions and education for these special populations. A study suggested that customized effective, ethical, and evidence-based communication may be able to increase the acceptance of the COVID-19 vaccination [103]. This customized intervention was suggested to provide by community leaders and healthcare practitioners to establish the trust of COVID-19 vaccination [88]. Moreover, a recent study also reported that providing the population with reliable information regarding the COVID-19 pandemic and the COVID-19 vaccination was associated with the increased rate of vaccination acceptance among the Israeli parents [104]. On the other hand, while we have provided the valuable information on the factors associated with the risk COVID-19 vaccine hesitancy, it should be realized that the main factor driving individuals to be able to receive the vaccinations is the proven effectiveness and safety of vaccines in well-documented long-term studies. However, among those exhibiting COVID-19 vaccination hesitancy, there are people who outright refuse vaccination. Therefore, further studies should be performed with a focus on this group.
There are some potential limitations of our study. First, a meta-analysis is a methodological approach conducted by calculating the crude effect of the related factors to determine the evidence. However, the impact of potential confounding factors is difficult to evaluate. In the current study, potential confounding factors such as the types of COVID-19 vaccine, government regulations, source of information regarding COVID-19 vaccination, and environmental factors were not included; therefore, our findings should be interpreted carefully. We reported that considering the COVID-19 vaccine to be unsafe was one of the factors associated with increased risk of COVID-19 vaccination hesitancy. Considering that the different types of COVID-19 vaccine have different side effects; the factors of the type of vaccine might also govern the final findings. Moreover, the government regulations in several countries have implemented COVID-19 vaccination as a condition of administration, and the regulations in each country may have differences; thereby, this circumstance may also affect the final findings of this study. Therefore, we reiterate the importance of the basic tenet in studying the phenomenon of vaccine hesitancy, which is the time, context, place, and type specificity. All these peculiarities need to be considered in the efforts aiming to fathom the determinants of vaccine hesitancy [105]. Second, our study involved a multi-national population, and the knowledge of disease prevention among people with similar socioeconomic status and educational level might vary in each region. Third, all the papers included in our study had an observational research design. Therefore, further studies including randomized controlled trials are required in order to obtain better levels of evidence.

Conclusions
Our study estimated the global prevalence of COVID-19 vaccination hesitancy at 25%. It also recommended special interventions to minimize COVID-19 vaccination hesitancy among unmarried individuals, women, people with low educational levels, the unemployed, people living in households with five or more individuals, and those who believe COVID-19 vaccines to be unsafe.