Mapping global acceptance and uptake of COVID-19 vaccination: A systematic review and meta-analysis.

Background: The COVID-19 pandemic exit strategies depend on widespread acceptance of COVID-19 vaccines. We aim to estimate the global acceptance and uptake of COVID-19 vaccination, and their variations across populations, countries, time, and sociodemographic subgroups.
Methods: We searched four peer-reviewed databases (PubMed, EMBASE, Web of Science, and EBSCO) for papers published in English from December 1, 2019 to February 27, 2022. This review included original survey studies which investigated acceptance or uptake of COVID-19 vaccination, and study quality was assessed using the Appraisal tool for Cross-Sectional Studies. We reported the pooled acceptance or uptake rates and 95% confidence interval (CI) using meta-analysis with a random-effects model.
Results: Among 15690 identified studies, 519 articles with 7,990,117 participants are eligible for meta-analysis. The global acceptance and uptake rate of COVID-19 vaccination are 67.8% (95% CI: 67.1-68.6) and 42.3% (95% CI: 38.2-46.5), respectively. Among all population groups, pregnant/breastfeeding women have the lowest acceptance (54.0%, 46.3-61.7) and uptake rates (7.3%, 1.7-12.8). The acceptance rate varies across countries, ranging from 35.9% (34.3-37.5) to 86.9% (81.4-92.5) for adults, and the lowest acceptance is found in Russia, Ghana, Jordan, Lebanon, and Syria (below 50%). The acceptance rate declines globally in 2020, then recovers from December 2020 to June 2021, and further drops in late 2021. Females, those aged < 60 years old, Black individuals, those with lower education or income have the lower acceptance than their counterparts. There are large gaps (around 20%) between acceptance and uptake rates for populations with low education or income.
Conclusion: COVID-19 vaccine acceptance needs to be improved globally. Continuous vaccine acceptance monitoring is necessary to inform public health decision making.

Introduction

The COVID-19 pandemic has become the most threatening global health issue[1]. The cataclysmic impact of the COVID-19 pandemic has contributed to an unprecedented pace in COVID-19 vaccine development[2]. Effective vaccine development usually takes almost 10 years, but COVID-19 vaccines have been developed and issued for use within a one-year timeframe[3]. Considering the delta variant, around 85% of the population should get immunity through natural infection or vaccination[4]. Given the powerful capability of the omicron variant to escape neutralizing antibodies elicited by current vaccines, more than 85% of the population need to get immunity[5].

Public confidence and acceptance of COVID-19 vaccines need to be ensured to achieve high vaccination uptake and herd immunity[6,7]. However, the accelerated development and issue process of COVID-19 vaccines may exacerbate public concerns regarding their safety and effectiveness[3]. The novelty of the COVID-19 disease, the anti-vaccine movement, and politicization of the COVID-19 vaccine may also negatively influence vaccine acceptance[8]. Previous studies have investigated public acceptance of COVID-19 vaccines, with substantial heterogeneity across the world[9-13]. Vaccine acceptance is defined as the individual or group decision to accept or refuse, when presented with an opportunity to vaccinate[14]. It is a complex and context specific issue that varies across time, place, and vaccines[15]. With the evolution of the pandemic and widespread dissemination of COVID-19 related misinformation[16], public acceptance may change over time. Although a growing body of literature has investigated public acceptance of COVID-19 vaccination, few studies have systematically reviewed and synthesized the current evidence[3,17-19].

We conducted a systematic review and meta-analysis to estimate the global acceptance and uptake of COVID-19 vaccination, including 1) global acceptance and uptake of COVID-19 vaccination in each population group, 2) cross-country comparison and time trends of vaccination acceptance, and 3) variations in vaccination acceptance and uptake across subgroups according to sociodemographic characteristics. We find that the global acceptance rate of COVID-19 vaccines is lower than 70%, with large variations between countries. The lowest acceptance rates are found in Russia, Ghana, Jordan, Lebanon, and Syria. The acceptance rates decline globally in 2020, then recover in the first half of 2021, and further drop in late 2021.Vulnerable populations with low acceptance rates include pregnant or breastfeeding women, Black people, and those with low socioeconomic status. Our findings highlight differences in vaccine acceptance between different populations, and suggest the need to carefully monitor and improve vaccine acceptance rates.

Method

Search strategy and selection criteria

This review was developed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines[20]. We employed the following search terms on four peer-reviewed databases (PubMed, EMBASE, Web of Science, and EBSCO): coronavirus terms (“coronavirus disease” OR coronavirus OR coronaviruses OR 2019-nCoV OR 2019ncov OR COVID-19 OR “severe acute respiratory syndrome coronavirus 2” OR SARS-2 OR SARS-COV-2) AND vaccine terms (vaccin* OR immunis* OR immuniz*) AND survey terms (survey OR questionnaire OR poll). All papers published in English from December 1, 2019 to February 27, 2022 were collected with the above search terms in all fields of studies. The detailed search strategy for each database is included in Supplementary Data 1.

Articles were included in this review if they investigated acceptance, willingness, intention, or uptake of COVID-19 vaccination, and if they were original survey studies. We excluded studies that investigated (1) non-COVID-19, clinical-trial, emergency or boosting vaccination acceptance, (2) studies which assessed willingness-to-pay or conditional acceptance, (3) studies without outcomes of interest, or (4) studies that applied continuous variables to evaluate vaccination acceptance. Studies using continuous variables usually adopted different response ranges with no consistent meanings for response options across studies, and there were also no clear cut-off points for vaccination acceptance or refusal in continuous variables. Therefore, studies with continuous variables were excluded in our review for conducting the meta-analysis of vaccination acceptance rate. The following study designs were also excluded: editorials, letters, commentaries, correspondences, study protocols, reviews, qualitative studies, intervention studies, and non-survey studies such as crawling information from social media. Two independent researchers (SH, QW) first screened titles and abstracts, and then scrutinized the full texts to estimate their eligibility. When they had disagreements on study selection, the third researcher (FD) was consulted. The review protocol is available on International prospective register of systematic reviews (PROSPERO) (ID: CRD42021261022). This study was exempt from ethical review due to use of publicly available data.

Data abstraction and quality assessment

Two researchers (SH, QW) independently performed article extraction and assessed the quality of included studies. When inconsistency arose, they were asked to discuss and revisit the article until reaching a consensus. We extracted the following information from the included articles: title, first author, publication date, journal, study design, sampling method, sample size, survey period, survey location, target population, and measurement questions about COVID-19 vaccination acceptance. To achieve the study objectives, we also extracted four outcomes: (1) overall acceptance of COVID-19 vaccination (total sample); (2) subgroups’ acceptance of COVID-19 vaccination (by gender, age, race, education, and income); (3) overall uptake of COVID-19 vaccination (total sample); and (4) subgroups’ uptake of COVID-19 vaccination (by gender, age, race, education, and income). For each included study, we described its characteristics, study design, and primary outcomes in Sheet 1 in Supplementary Data 2.

The Appraisal tool for Cross-Sectional Studies (AXIS tool), a novel critical appraisal tool that addresses study design and reporting quality as well as the risk of bias in cross-sectional studies, was used to assess the quality of the included studies[21]. This tool, shown in Supplementary Data 3, includes three domains and twenty items with a total possible score of 20: quality of reporting (7 items), study design quality (7 items), and the possible introduction of biases (6 items). Given 12 of 20 scores (60%) are considered pass, studies > 12 scores were considered with the high-quality, and data from those high-quality studies were extracted for the review and meta-analysis. The quality assessment scores of each included study are shown in Supplementary Data 4.

Statistical analysis

Data organization and meta-analysis were carried out using Microsoft Excel and STATA 15.1 software respectively. Figures were done with R (version 4.1). Acceptance, willingness, or intention of COVID-19 vaccination were categorized into three groups: (1) Yes/ Definitely/ Probably; (2) Unsure/ Neutral/ I don’t know; and (3) No/ Definitely not / Probably not (Sheet 1 in Supplementary Data 2). The first one was labelled as “accept”, and the latter two were labelled as “vaccine hesitant”. For studies covering the vaccinated individuals, we took vaccinated individuals as the “accept” group when calculating vaccine acceptance rates. The acceptance rate of COVID-19 vaccination was defined as the proportion of participants willing to / accept / intend to / will get vaccination against COVID-19 in total surveyed participants. Uptake status of COVID-19 vaccination were categorized into two groups: (1) Yes and (2) No.

We reported the pooled acceptance or uptake rates and 95% confidence interval (CI). We employed a DerSimonian and Laird random-effects models[22] to conservatively estimate the pooled acceptance or uptake of COVID-19 vaccination, in case of significant heterogeneity (I² > 50%) between studies. Variability between studies was determined by the heterogeneity tests with Higgins’ I² statistic. Stratified subgroup analyses were conducted to explore the sources of heterogeneity. A value of P < 0.05 was considered statistically significant.

The pooled acceptance or uptake rates of COVID-19 vaccination were estimated by different populations, countries, survey times, and participants’ characteristics. We categorized all study participants into seven groups: (1) adults, (2) healthcare workers, (3) patients with chronic diseases, (4) pregnant or breastfeeding women, (5) university students, (6) children and adolescents, and (7) other populations (Sheet 1 in Supplementary Data 2). Other populations were defined as study populations that cannot be categorized into the first six study populations, such as the homeless, those in a particular occupation, and elderly persons with Medicare. We first estimated the pooled acceptance or uptake of COVID-19 vaccination for each population group, and within each population group, we then estimated the acceptance or uptake of COVID-19 vaccination in individual countries by synthesizing all studies from the same country. To compare the trends of COVID-19 vaccination acceptance over time, we reported the pooled acceptance rate of all studies from the same survey month, and developed graphs to illustrate the time trends. For acceptance or uptake estimates from adults, we also conducted subgroup analyses based on their sociodemographic characteristics such as gender, age group, race, education, and income. Additionally, we did a sensitivity analysis of the pooled acceptance rate of COVID-19 vaccination with studies whose sample size was more than 300 (Supplementary Table 1 in Supplementary Information).

Table 1

Estimated acceptance and uptake of COVID-19 vaccination by study populations

	Acceptance of COVID-19 vaccines										Uptake of COVID-19 vaccines
	Willing to get a COVID-19 vaccine				Unsure			Unwilling to get a COVID-19 vaccine
Population groups	No. of studies		No. of participants	Estimated acceptance (%, 95% CI)	No. of studies	No. of participants	Estimated unsure rate (%, 95% CI)	No. of studies	No. of participants	Estimated unwillingness (%, 95% CI)	No. of studies	No. of participants	Estimated uptake (%, 95% CI)
Overall	476		7967690	67.8 (67.1–68.6)	221	1392225	20.3 (19.3–21.3)	404	7748451	20.4 (19.6–21.3)	139	6179852	42.3 (38.2–46.5)
Adults	202		7068345	69.1 (68.2–70.1)	88	972239	19.4 (18.1–20.8)	177	6948770	19.8 (18.5–21.1)	49	5877618	39.7 (32.4–47.1)
Healthcare workers	98		129265	67.5 (64.4–70.6)	47	58906	21.7 (18.5–24.9)	83	92347	19.8 (17.7–21.9)	28	57789	54.1 (46.5–61.7)
Patients with chronic diseases	52		166278	67.4 (63.9–70.9)	29	152918	19.6 (16.9–22.2)	40	160584	16.9 (13.9–19.9)	19	132502	39.3 (31.9–46.7)
Pregnant/breastfeeding women	13		25102	54.0 (46.3–61.7)	3	2949	24.2 (19.4–28.9)	12	9252	41.9 (33.0–50.8)	3	2955	7.3 (1.7–12.8)
University students	45		50471	67.7 (62.7–72.8)	20	30469	19.6 (14.8–24.4)	38	45910	20.7 (17.2–24.2)	16	18937	43.7 (31.2–56.1)
Children and adolescents	46		363825	70.7 (67.6–73.9)	16	70757	25.2 (19.0–31.4)	33	338814	19.8 (17.9–21.7)	6	3969	37.9 (22.5–53.4)
Others	58		164404	65.9 (60.8–71.0)	35	103987	19.6 (16.4–22.8)	52	152774	21.7 (18.4–25.1)	23	86082	39.4 (26.1–52.8)

For each pooled estimate, heterogeneity tests between studies reached Higgins’ I² statistic > 92%, P < 0.001.