The refusal of COVID-19 vaccination and its associated factors: a systematic review

Introduction

At the beginning of 2021, the coronavirus disease 2019 (COVID-19) vaccination program involving several designs including protein subunit, vector, inactivated, and mRNA, was started.¹ Currently, referring to data presented on Our World in Data, this vaccination program has included 84.6% of the global population, and the reason the rest of the population (15.4%) did not receive vaccination is still unknown.² The high number of vaccinated country populations is the result of the hard work of various parties, and this may be associated with factors such as the seriousness of governments in promoting vaccination programs, equitable distributions of vaccines, hard work of healthcare workers, good public awareness about the importance of vaccination, and good promotion of vaccines to populations who have the power to hesitate about vaccines.³ Contrarily, in the unvaccinated population, several factors may contribute to hesitation, including fear of harmful ingredients in vaccines, distrust of pharmaceutical companies, lack of knowledge about COVID-19, belief that a healthy lifestyle and a good diet are enough to fight against COVID-19, preference for natural immunity, lack of seriousness from the government in promoting vaccination programs, religious rules suggesting not to vaccinate, and limited information regarding the safety of vaccination. These factors have been reported to trigger hesitancy and refusal of the COVID-19 vaccination.^4–7 In our previous study, we had explored the global prevalence of COVID-19 vaccination hesitancy and its potential associated factors.⁸ However, because the hesitancy population consists of hesitancy and refusal populations, and the refusal population can influence people in their circle to become hesitant or refuse the COVID-19 vaccine, the prevalence of the COVID-19 vaccine refusal should also be investigated.

It is widely known that new vaccines or vaccine candidates are commonly hesitated against or rejected by the public. Before the COVID-19 pandemic, this phenomenon has been widely reported in several cases, such as: dengue,⁹ malaria,¹⁰ Ebola,¹¹ chikungunya,¹² and monkeypox.¹³ This might be caused by poor public knowledge regarding the vaccine, including the poor knowledge of vaccine efficacy and side effects. In the case of COVID-19, this phenomenon might be affected by multiple factors, and theoretically, the factors had been contextualized into three major categories, including poor knowledge of vaccination programs, socioeconomic status, and social interaction.¹⁴ Moreover, recently, influencers in their podcasts discussed the rejection of the COVID-19 vaccine, which is a dilemma that can govern people in society to reject COVID-19 vaccinations, thereby threatening the success of the COVID-19 vaccination program.¹⁵ However, to date, there are no precise data on the prevalence of COVID-19 vaccination refusal and its potential associated factors. Several previous studies have investigated the refusal of COVID-19 vaccines; however, the results of these studies have been inconclusive. In the present study, we seek to explore the global prevalence of COVID-19 vaccination refusal and identify the associated factors using a meta-analysis approach.

Methods

Results

Selection of studies

A total of 3,422 papers and 4 papers were assessed from the databases and reference lists of related articles, respectively. In the initial evaluation, we excluded 33 papers because of duplication, and 3,318 papers because of irrelevant topics. Subsequently, 75 articles were included in further review. We further excluded 17 articles as they were reviews and 34 articles because of insufficient data. Finally, the data retrieved from 24 articles were analyzed to estimate the cumulative prevalence and factors associated with COVID-19 vaccination refusal.^17–40 The flow diagram of article selection in our study is outlined in Figure 1, and the characteristics of the articles included in our study are listed in Table 1.

Figure 1. A flowchart of article selection in our study.

Table 1. Baseline characteristics of articles included in our study.

Author & year	Country	Study design	Sample size	Study period	Population	NOS
Al- Sanaf et al. 2021	Kuwait	Cross-sectional	1019	March 2021	HCW	5
Aurilio et al. 2021	Italy	Cross-sectional	531	December 2020	HCW	6
Baniak et al. 2021	USA	Cross-sectional	275	February 2021	HCW	6
Fakonti et al. 2021	Cyprus	Cross-sectional	435	December 2020	HCW	6
Fares et al. 2021	Egypt	Cross-sectional	385	December 2020–January 2021	HCW	7
Fisher et al. 2020	USA	Cross-sectional	991	April 2020	GP	6
Grochowska et al. 2021	Poland	Cross-sectional	419	September–November 2020	HCW	5
Handam et al. 2021	Lebanon	Cross-sectional	758	May–June 2021	Students	6
İkiışık et al. 2021	Turkey	Cross-sectional	384	December 2020	GP	5
Janssen et al. 2021	France	Cross-sectional	4349	December 2020–March 2021	HCW	5
Kose et al. 2020	Turkey	Cross-sectional	1138	September 2020	HCW	7
Kozak et al. 2021	Germany	Cross-sectional	3368	March–April 2021	HCW	6
Manning et al. 2021	USA	Cross-sectional	1205	August 2020–September 2020	Students	7
Mena et al. 2021	Spain	Cross-sectional	865	December 2020–January 2021	HCW	7
Ousseine et al. 2021	France	Cross-sectional	15426	February–April 2021	GP	6
Paris et al. 2021	France	Cross-sectional	1965	February–February 2021	HCW	7
Pataka et al. 2021	Greece	Cross-sectional	656	December 2020	HCW	6
Rodriguez-Blanco et al. 2021	Spain	Cross-sectional	2494	November–December 2020	GP	7
Saied et al. 2021	Egypt	Cross-sectional	2133	January 2021	Students	7
Schwarzinger et al. 2021	France	Cross-sectional	1942	July 2020	GP	6
Shaw et al. 2021	USA	Cross-sectional	5287	November–December 2020	HCW	6
Spinewine et al. 2021	Belgium	Cross-sectional	1132	January 2021	HCW	6
Vignier et al. 2021	French Guiana	Cross-sectional	579	January–March 2021	HCW	8
Wang et al. 2020	Hong Kong	Cross-sectional	806	February–March 2020	HCW	6

The cumulative prevalence of the refusal to COVID-19 vaccination

Our analysis identified that the cumulative prevalence of the refusal to COVID-19 vaccination was 12% (event rate: 0.12; 95%CI: 0.10, 0.15; p Egger: 0.5290; p Heterogeneity<0.0001; p<0.0001) (Figure 2A). Subsequently, sub – group analysis found that the prevalence of the refusal to COVID-19 in general population was 20% (Figure 2B), healthcare workers 10% (Figure 2C), and students 11% (Figure 2D).

Figure 2. The prevalence of COVID-19 vaccines refusal.

A). All prevalence of COVID-19 refusal (Event rate: 0.12; 95%CI: 0.10, 0.15; p Egger: 0.5290; p Heterogeneity<0.0001; p<0.0001).

B). The prevalence in general population subgroup (Event rate: 0.20; 95%CI: 0.15, 0.26; p Egger: 0.4100; p Heterogeneity<0.0001; p<0.0001).

C). The prevalence in healthcare workers subgroup (Event rate: 0.10; 95%CI: 0.07, 0.14; p Egger: 0.6760; p Heterogeneity<0.0001; p<0.0001).

D). The prevalence in student subgroup (Event rate: 0.11; 95%CI: 0.06, 0.20; p Egger: 0.5830; p Heterogeneity<0.0001; p<0.0001).

Factors associated with COVID-19 vaccination refusal

Table 2 and Figures 3–5 summarize the factors associated with refusal of COVID-19 vaccination. Our calculation revealed that six of the 15 factors had a significant impact on COVID-19 vaccine refusal. We found that an increased risk of COVID-19 vaccine refusal was observed in the following covariates: female (Figure 3A), educational attainment lower than an undergraduate degree (Figure 4A) and living in rural areas (Figure 5B).

Table 2. Factors associated with refusal to COVID-19 vaccination.

Covariates	Refusal/Total (n [%])	NS	p Egger	p Het	OR	95% CI	p
Age group (years)
<30	1827/9397 (19.44%)	9	0.2570	<0.0001	1.20	0.96-1.50	0.1090
30-50	1575/9579 (16.44%)	9	0.1600	0.0200	0.96	0.82-1.13	0.6210
>50	639/5344 (11.96%)	9	0.5240	<0.0001	0.91	0.60-1.36	0.6300
Sex
Male	1194/8499 (14.05%)	22	0.2550	<0.0001	0.71	0.61-0.82	<0.0001
Female	3659/19842 (18.44%)	22	0.2550	<0.0001	1.42	1.22-1.65	<0.0001
Marital status
Married	1702/9960 (17.09%)	6	<0.0001	<0.0001	0.77	0.71-0.82	<0.0001
Single	1954/9693 (20.16%)	6	<0.0001	0.9620	1.31	1.21-1.41	<0.0001
Educational attainment
<BSc	2892/12199 (23.71%)	10	0.3620	<0.0001	1.74	1.34-2.26	<0.0001
≥BSc	2397/16931 (14.16%)	10	0.3620	<0.0001	0.58	0.44-0.75	<0.0001
Employment
Not Working	1751/8770 (19.97%)	7	0.1240	0.0550	1.11	0.96-1.28	0.1670
Working	2666/14218 (18.75%)	7	0.1450	0.0240	0.93	0.79-1.09	0.3850
Socioeconomic status
Low	712/2947 (24.16%)	6	0.6470	<0.0001	1.05	0.59-1.84	0.8770
Medium	1947/8561 (22.74%)	6	1.0000	<0.0001	1.86	0.81-4.28	0.1450
High	1238/10286 (12.04%)	6	0.9900	<0.0001	0.42	0.18-0.99	0.0480
Having children at home	421/1775 (23.72%)	4	0.5940	<0.0001	0.65	0.35-1.21	0.1730
Hospitalization due to COVID-19 among people in the same social circle	16/668 (2.40%)	2	0.5890	0.0250	0.62	0.05-7.13	0.7030
Health literacy about COVID-19 vaccine	1795/11153 (16.09%)	4	0.5630	0.0010	0.60	0.31-1.17	0.1330
History of previous vaccination	2456/18868 (13.02%)	14	0.9710	<0.0001	0.28	0.17-0.48	<0.0001
History of chronic disease(s)	275/2317 (11.87%)	9	0.5130	<0.0001	0.94	0.64-1.38	0.7510
Personal history of COVID-19 diagnosis	494/3498 (14.12%)	8	0.1970	0.0280	0.94	0.77-1.15	0.5560
Family member/friend ever diagnosed with COVID-19	932/6739 (13.83%)	8	0.2800	<0.0001	0.94	0.72-1.21	0.6120
Ever tested for COVID-19	26/582 (4.47%)	2	<0.0001	0.6930	0.80	0.48-1.34	0.3950
Medical background	200/3642 (5.49%)	11	0.9170	<0.0001	0.32	0.17-0.58	<0.0001
Residential
Urban	2629/16172 (16.26%)	4	0.3030	<0.0001	0.62	0.44-0.88	0.007
Rural	859/3324 (25.84%)	4	0.3030	<0.0001	1.61	1.14-2.28	0.007

Figure 3. Females had higher risk of refusal to COVID-19 vaccination than males (A), and individual history of previous vaccination had lower risk of refusal to COVID-19 vaccination (B).

Figure 4. Individual with educational attainment <BSc had higher risk of refusal to COVID-19 vaccination than ≥BSc (A), and individual with medical background had lower risk of refusal to COVID-19 vaccination compared to general population (B).

Figure 5. Individual with high SES had lower risk of refusal to COVID-19 vaccination than low SES (A), and rural residential living was associated with increased risk of refusal to COVID-19 vaccination than urban population (B).

In contrast, the decreased risk of refusal of COVID-19 vaccination was affected by the following factors: high socioeconomic status (Figure 5A), history of previous vaccination (Figure 3B), and individuals with medical background (Figure 4B).

Discussion

Our meta-analysis revealed that the prevalence of refusal to undergo the COVID-19 vaccination was 12%. Our findings were lower than those reported by Cenat et al. and Robinson et al.^41,42 In our study, we had a larger sample size than those reported by in these studies. Moreover, studies by Cenat et al. and Robinson et al. also involved articles that reported COVID-19 vaccination hesitancy.^41,42 It is well known that the terminologies of refusal and hesitancy to vaccinate are different, and not everyone is hesitant to vaccinate. Thus, it can be assumed that the context of previous studies has a gap in the definition of vaccine refusal. Therefore, our study may provide better data on the prevalence rate of COVID-19 vaccination refusal. Moreover, we also reported the prevalence of COVID-19 vaccination refusal in some subgroup populations: general population, healthcare workers, and students. We found that healthcare workers had the lowest prevalence of COVID-19 vaccination refusal, followed by students, and the general population. Our current findings indicate that vaccination knowledge might affect our findings. We assumed that healthcare workers and students may have a better knowledge of vaccination programs than the general population. This assumption is supported by the results of previous studies, which found that healthcare workers and students had better knowledge of COVID-19 vaccination than the general population,^43,44 and this factor was also shown to contribute to the acceptance of vaccination programs.⁴⁵

Our study found that the increased risk of COVID-19 vaccination refusal was higher in females and individuals with educational levels below an undergraduate degree (BSc). In contrast, lower risk of COVID-19 vaccination refusal was found in individuals with a history of previous vaccination and medical background. Our current findings suggest that the factors related to knowledge of COVID-19 vaccination had the potential to affect the refusal to vaccinate against COVID-19. As previously reported, a study revealed that females lacked literacy regarding COVID-19 vaccination than males.⁴⁶ This may be due to the fact that the majority of females are housewives, and therefore, may have less social interaction than males, as they are based at home rather than going out to work.⁴⁷ This possibility might contribute to the lack of knowledge on COVID-19 vaccination in the female population. Furthermore, one study found that the majority of the side effects of COVID-19 vaccination were found among female individuals⁴⁸ Taken together, those factors may affect the decision to accept or refuse the vaccines. Moreover, individuals with educational level below the undergraduate (BSc) degree might have an inadequate source of literature regarding COVID-19 vaccination compared to those with an educational level higher than an undergraduate (BSc) degree. In the context of vaccination knowledge, a study found that educational attainment was one of the predictors of vaccination knowledge, where lower educational attainment was associated with poorer knowledge of vaccination.⁴⁹ Therefore, the population with an educational level below the undergraduate (BSc) degree might have insufficient consideration for COVID-19 vaccination compared to those with an educational level higher than the undergraduate (BSc) degree. Further, individuals with a history of previous vaccination and a medical background may have adequate information regarding the importance of COVID-19 vaccination, therefore, may have sufficient awareness regarding COVID-19 vaccination. Previous studies found that individuals with a medical background had better knowledge of COVID-19 vaccination than the general population.^43,44 Likewise, a study revealed that individuals with a history of previous annual vaccination had good awareness and knowledge of the importance of vaccination programs.⁵⁰ Prior to the COVID-19 pandemic, studies have widely reported that knowledge on disease prevention and the good practice of health behavior had a significant impact on the acceptance rate of vaccination in the case of Monkeypox, Ebola, and Dengue.^51–53 Thus, this might imply that this population (individuals with a history of previous vaccination and medical background) has a low rate of refusal to vaccinate against COVID-19, as reported in our meta-analysis.

Our study also found that a higher risk of COVID-19 vaccination refusal was observed in rural than urban populations, and a lower risk of COVID-19 vaccination refusal was observed in individuals with high SES than in those with low SES. Currently, the precise explanations regarding the reasons underlying our findings might be difficult to explain and might have differences between each region. However, we can propose the following reasons: social privileges, administrative requirements, and social circles. First, in the social privilege aspect, individuals with high SES might have certain pride if they have been vaccinated, while this pride might not belong to the rural population. Studies found that COVID-19 vaccination was considered a socioeconomic privilege and political ideology,⁵⁴ while the rural population did not consider that the COVID-19 vaccine was a privilege and tended to have poor perception toward vaccine safety.⁵⁵ The second reason is administrative requirements. Individuals with high SES might need COVID-19 vaccination as an administrative requirement for their activities; for example, business, travelling, and carrier requirements as the World Health Organization (WHO) has implemented a COVID-19 vaccine certificate as an administrative requirement for travel or business.⁵⁶ However, these administrative requirements were not necessary for rural individuals, as the majority of rural individual jobs are in private and traditional sectors, such as farmers, fishermen, and manual laborers.^57,58 The third factor is social circle. Individuals with high SES might have social circles that interact with high intellectual content, whereas in rural populations, their social circle might be limited to their neighbors with similar intellectual contents. This factor might also indirectly contribute to the understanding of COVID-19 vaccinations, and consequently, affect their decision to accept or refuse the COVID-19 vaccine. This is supported by previous studies that revealed that SES was associated with the level of knowledge of vaccination programs and physical health status.^59,60 Moreover, our previous study on dengue also revealed that SES was one of the predictive indicators for the acceptance of vaccination.⁵³

Our meta-analysis is one of the first to report the prevalence of COVID-19 vaccination refusal and the potential factors associated with refusal of COVID-19 vaccination. Our study also had a larger sample size compared to previous meta-analyses in a similar context.^41,42 The findings of our study might serve as the initial step to prevent the failure of COVID-19 vaccination programs. By identifying the potential factors associated with refusal to vaccinate against COVID-19, we expect that governments may provide advanced intervention to those populations. As previously reported, the main concern to obtain public trust regarding the COVID-19 vaccines was the lack of adequate evidence from long-term and large-scale studies on the effectiveness and safety of COVID-19 vaccination.⁶¹ However, several studies have provided suggestions on how to provide interventions to the refusal population: providing reliable information regarding the COVID-19 pandemic and the COVID-19 vaccination, providing customized, effective, ethical, and evidence-based communication; and the information should be provided by community leaders and healthcare practitioners.^62–64

Our meta-analysis has several limitations. First, several potential confounding factors, such as the level of knowledge about COVID-19 vaccination, attitude toward COVID-19 prevention, government regulation, types of vaccine, environmental factors, and the source of literature regarding COVID-19 vaccination, were not included in the analysis due to the lack of available data. Second, the sample size in our present study was limited; therefore, further studies involving larger sample sizes are needed. Third, our meta-analysis could not reflect the prevalence of the global numbers because the proportion of sample sizes in each region was unequal. Fourth, all articles included in our analysis were cross-sectional studies. Further meta-analyses involving only randomized controlled trials are required.

Conclusion

In conclusion, we revealed that the cumulative prevalence of the refusal to COVID-19 vaccination was 12%, the highest in the general population, and the lowest in the healthcare worker subgroup. The individuals with the following characteristics are found increased risk to refuse COVID-19 vaccination: female, having an educational attainment lower than an undergraduate degree, and living in a rural area, while the following factors are associated with reduced risk to refuse COVID-19 vaccination: high socioeconomic status, history of previous vaccination, and individuals with a medical background.

Author contribution

Conceptualization: FT, JKF, GS; Data Curation: FT, JKF, GS, APK, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP; Formal Analysis: JKF, APK, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP; Investigation: FT, JKF, GS, LW, APK, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP; Project Administration: APK, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP; Resources: APK, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP; Methodology: FT, JKF, GS, LW, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP; Software: FT, JKF, APK, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP; Visualization: APK, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP, CC, KD, HH; Supervision: FT, JKF, GS, LW, CC, KD, HH; Validation: FT, JKF, LW, CC, KD, HH; Writing – Original Draft Preparation: FT, JKF, LW, APK, EAP, RPP, MR, TA, MN, SW, GP, AP, QA, MVPHM, RY, YSP; Writing – Review & Editing: FT, JKF, GS, LW, CC, KD, HH. All authors have critically reviewed and approved the final draft and are responsible for the content and similarity index of the manuscript.