The emergence of the Omicron variant in late 2021 and early 2022 has reinforced the effects of COVID-19 vaccine inequity on the pandemic. As many Western countries such as the United Kingdom and United States responded to the identification of the highly transmissible Omicron variant by closing borders to citizens of southern African countries [1,2], scientists have questioned the effectiveness of border restrictions, highlighted the knock-on effect on economies and households, and rather emphasized the need to reduce the huge disparity in vaccine access in low- and middle-income countries (LMICs) as the sustainable solution [3,4].
Vaccination rates in LMICs, particularly in Africa, continue to fall short of global goals [5,6]. Although it is clear that access to vaccines is a key determinant of vaccine coverage, it is also affected by vaccine hesitancy and implementation challenges. However, better data are needed to determine the barriers to vaccination to assist with tailoring interventions that will improve vaccine uptake in the African region. Research in South Africa has shown that there has been a general shift towards the acceptance of vaccines since the start of the COVID-19 pandemic; however, approximately 29% of adults reported being unwilling to be vaccinated in May/April of 2021 . Some primary concerns that have fueled COVID-19 vaccine hesitancy in South Africa and the African region include concerns about safety, which in part is due to the rapid pace of vaccine development [7,8].
Growing anti-vaccine sentiment in Africa, especially in South Africa, has recently garnered global attention . Even before the COVID-19 pandemic, vaccine hesitancy in South Africa was implicated in several vaccine-preventable disease outbreaks, including a measles outbreak involving 18 311 cases between 2009 and 2010 . Expectedly, COVID-19 vaccination efforts in the country have again brought this issue to the forefront. For instance, in their evaluation of the COVID-19 vaccine rollout for health care workers in South Africa in early 2021, Wiysonge et al. revealed high levels of vaccine hesitancy at 41% among health care workers in Cape Town, those who would be charged with administering vaccines . In their evaluation of vaccine hesitancy surveys conducted in South Africa prior to and during the start of COVID-19 vaccine rollout, Cooper et al. acknowledge the fluctuating COVID-19 vaccine acceptance , reiterating the urgent need for studies that would deepen the understanding of vaccine hesitancy in South Africa and inform interventions for increasing uptake.
Aiming to cover this information gap, we conducted a population-based, cross-sectional telephonic survey among individuals ages 18 and older from August 25 to October 29, 2021, in the Agincourt Health and Socio-Demographic Surveillance System (AHDSS) area located in Mpumalanga province, northeast South Africa, adjacent to southern Mozambique. The rural setting is characterized by high levels of poverty, unemployment, and circular labour migration . From May 2021, COVID-19 vaccines became available to adults aged 60 and older, and from September 1, 2021, South Africans 18 and older were eligible to receive the COVID-19 vaccine (Pfizer and Jansen are available). 1664 adults completed the telephonic interview, and 1662 (99.9%) completed the vaccine module and were included in this analysis. This survey was part of an ongoing panel, comprising a cohort of individuals measuring the impact of COVID-19 on economic, behavioural, and mental health outcomes in the AHDSS. The original sample was selected in August 2020 and included 2300 households (one individual per household was sampled). The sample was selected to be representative of the population and to ensure a sufficient sample size to generate precision around estimates by gender and age groups.
We measured factors that have been shown to impact vaccine uptake. Based on the theoretical model of increasing vaccination put forward by Brewer et al. and adapted by the World Health Organization (WHO) working group, we measured items in the constructs that shape vaccination uptake [13,14]: 1) what people think and feel – confidence in the vaccine safety and efficacy; 2) social processes – family and social support for vaccination; 3) motivation – intention to get vaccinated and 4) practical issues – knowledge of where the vaccine is available, ease of access, etc. The survey asked individuals about demographics, vaccine uptake (whether they had received the COVID-19 vaccine, which vaccine, when they received it, how many doses), social processes (disclosure of vaccine status to others and encouraging others to be vaccinated); the proportion of family and friends that are vaccinated or intend to be vaccinated; trusted sources of vaccine information; support from local and national leaders around vaccination; what people think and feel (confidence in vaccine safety and efficacy, reasons for vaccination or non-vaccination); practical issues (ease of access to the vaccine and knowledge of where to get the vaccine). Data were weighted by age and gender to be representative of the AHDSS from 2020 (total AHDSS eligible population = 34 582). Specifically, inverse probability of selection weights were applied to each stratum of gender (female/male) and age (age groups: 18-29, 30-39, 40-49, 50-59, 60-69, 70-79). Weighted descriptive results are presented below (counts, percentages, and 95% confidence intervals (CIs)) and χ2 tests were conducted to preliminarily assess differences between groups. All statistical testing was two-sided with an alpha of 0.05; analyses were conducted using Stata version 16.1 (StataCorp LLC, College City Texas, USA). Verbal informed consent was obtained from all participants and ethical approval was obtained from the Human Research Ethics Committee at the University of the Witwatersrand and the Institutional Review Board of the University of North Carolina.
Half of the participants reported receiving a COVID-19 vaccine (50.4%, 95% CI = 47.9-52.9) with 41.1% (95% CI = 38.7-43.5) being fully vaccinated and 9.3% (95% CI = 8.0-10.7) partially vaccinated, while 49.6% (95% CI = 47.1-52.1) were not vaccinated. More women than men were vaccinated (55.5% of women vs 42.8% of men, P < 0.001), and older age groups were more likely to be vaccinated than younger age groups (P < 0.001) (Table 1). Of those vaccinated, 68.6% (95% CI = 65.4-71.7) reported having been vaccinated at a community venue such as a church or school, and 96.4% (95% CI = 94.8-97.4) reported that the main reason they accepted vaccination was to protect their health and that of their families, friends, and communities. When asked if they had disclosed their vaccination status to anyone, 95.2% (95% CI = 93.6-96.5) reported that they had. Among vaccinated individuals, 87.3% (95% CI = 85.0-89.4) anticipated that most of their family and friends would get a COVID-19 vaccine compared to only 70.5% (95% CI = 67.0-73.7) of unvaccinated individuals (P < 0.001). 96.5% (95% CI = 95.1-97.5) of vaccinated participants and 85.9% (95% CI = 83.1-88.3) of unvaccinated participants personally knew someone who had been vaccinated (P < 0.001).
Table 1. Demographic characteristics among adults 18 and older in the Agincourt Health and Socio-Demographic Surveillance System (AHDSS) area, Mpumalanga, South Africa. *†
CI – Confidence interval
*August 25 to October 29, 2021, stratified by COVID-19 vaccination status and weighted to be representative of the AHDSS by age and gender.
† All counts (n), percentages (%), and 95% CIs are weighted values, including those for missing data.
Of the 49.6% who had not received any dose of a COVID-19 vaccine, when asked if they would get vaccinated, 69.0% (95% CI = 65.5-72.3) said they would get vaccinated as soon as they could, 8.0% (95% CI = 6.2-10.2) said they would wait a while, 7.4% (95% CI = 5.7-9.6) reported being unsure, and 14.7% (95% CI = 12.3-17.5) said they would definitely not get vaccinated. When unvaccinated participants were asked why they had not been vaccinated, the most common reasons were lack of information (36.6%, 95% CI = 33.1-40.2), such as not knowing if they were eligible and not knowing where to go (Table 2). This was followed by COVID-19 vaccine hesitancy (19.7%, 95% CI = 16.9-22.8), such as concerns about side effects, and health care access issues (14.3%, 95% CI = 11.9-17.1), such as vaccination sites being open at inconvenient hours, long wait times, or distance from the site. The hesitancy toward vaccines in general (8.4%, 95% CI = 6.6-10.7) and possible contraindications to the vaccine (6.2%, 95% CI = 4.7-8.0) were other relevant factors contributing to not receiving the vaccine. Among the 14.7% who said they would definitely not get the COVID-19 vaccine, there was a clear age gradient from youngest to oldest, 50.9% were 18-29-year-olds, 21.3% were 30-39-year-olds, 12.5% were 40-49-year-olds, 8.8% were 50-59-year-olds, 4.1% were 60-69-year-olds, and 2.5% were 70+ year-olds; there was no difference by gender. When examining the perceived risk of COVID-19 among the 14.7% who said they had no plans to be vaccinated compared to those who said they did plan to get vaccinated, 27.9% (95% CI = 20.0-37.3) of those who did not plan to get vaccinated thought it was moderately/extremely likely they would get seriously ill if they got COVID-19 compared to 49.2% (95% CI = 45.3-53.2), (P < 0.001) of those who planned to get vaccinated.
Table 2. Primary reason for not getting a COVID-19 vaccine among adults in Mpumalanga, South Africa, stratified by gender*
CI – Confidence interval
*All counts (n), percentages (%), and 95% CIs are weighted values, including those for missing data.
As expected, confidence in COVID-19 vaccine safety was higher among those vaccinated with 75.3% (95% CI = 72.3-78.0) being somewhat or very confident the COVID-19 vaccine is safe vs only 51.2% (95% CI = 47.6-54.9) of unvaccinated individuals (P < 0.001). Similarly, 75.8% (95% CI = 72.9-78.5) of the vaccinated were somewhat or very confident that the COVID-19 vaccine is effective vs 51.0% (95% CI = 47.3-54.6) of the unvaccinated (P < 0.001). The individuals most trusted for providing accurate information on COVID-19 vaccines were reported to be doctors (58.0%, 95% CI = 55.5-60.4) and nurses (49.1%, 95% CI = 46.6-51.5) for both vaccinated and unvaccinated individuals (Table 3). Government officials were endorsed by 15.9% (95% CI = 14.2-17.8), and family and friends were trusted by only 6.1% (95% CI = 5.0-7.5). Most individuals obtained information on COVID-19 vaccines from TV (59.2%, 95% CI = 56.7-61.6) and radio (46.1%, 95% CI = 43.7-48.6) followed by social media (24.8%, 95% CI = 22.6-27.1). Far fewer obtained information from friends, family, church, school, or websites (all less than or equal to 5.0% each); importantly, for the unvaccinated among the 18-29 age group, 42.4% (95% CI = 37.1-47.9) reported getting information on vaccines from social media.
Table 3. Sources of COVID-19 vaccination information, support, and trust among adults in Mpumalanga, South Africa, stratified by vaccination status*
CI – Confidence interval
*All counts (n), percentages (%), and 95% confidence intervals (CIs) are weighted values, including those for missing data.
While there has been much debate on the reasons for low vaccine coverage in Africa, clearly the causes are multi-faceted and involve both structural factors, such as vaccine supply and distribution , and on-the-ground demand [8,9]. Across Africa, access to vaccines remains an issue, as the majority of countries have less than 10% of adults fully vaccinated . Wealthier nations must continue to provide vaccines to LMICs and to support vaccine distribution, especially in rural, inner-city, and remote locations. While South Africa is an outlier in the region, now having sufficient availability of vaccines, ensuring easy access to those who want them, and addressing concerns around vaccine safety and efficacy are still major challenges. In particular, efforts should prioritize those for whom access is lower and most tenuous, including men, younger populations, labour migrants, undocumented migrants, and the unhoused, to name a few.
Our findings support the Increasing Vaccination Model put forward by the WHO working group on behavioural and social drivers of vaccination . We observed significant differences between vaccinated and unvaccinated individuals in their thoughts and feelings about vaccines, particularly concerning confidence in vaccine safety and efficacy, and also in their risk perception about the severity of illness if they were to become infected with COVID-19. This relationship is also supported by Kollamparabil et al., who highlight the moderating effect risk perception has on COVID-19 vaccine confidence and thus behaviour in the South African context . Social processes seemed less important in this setting, and while unvaccinated individuals knew fewer vaccinated individuals and anticipated fewer friends and family would be vaccinated than vaccinated individuals, both groups reported very high levels of vaccination in their peer groups. In this study population, access was a major contributor to non-vaccination with confusion about where to get vaccinated, eligibility for vaccination, and also factors associated with the ability to access vaccination sites. Interestingly, interventions that directly impact behaviour change such as vaccine mandates, incentives, default appointments, and on-site vaccinations have shown the greatest likelihood of impact in increasing vaccination uptake . In addition, health care provider recommendations for vaccination have been found to be one of the best interventions for increasing vaccine uptake . In our study population, doctors and nurses were the most trusted source for accurate information on vaccines among vaccinated and unvaccinated individuals. In the analysis in early 2021 looking at COVID-19 vaccine hesitancy in South Africa, trust in health care workers did not predict lower vaccine hesitancy thus the authors posit that health care workers have not been used effectively to address misinformation . How trust in health care workers can be best translated into interventions in LMIC settings to address misinformation and increase vaccination coverage is key. Therefore, along with the aforementioned interventions, we recommend a concerted effort by doctors and nurses to address vaccine concerns. We also endorse the continued use of television and radio to inform the public about vaccinations. However, a greater emphasis needs to be put on using these outlets to mitigate confusion about eligibility and vaccination site locations.
The limitations of this research were in the telephonic conducting of interviews, because of which verification of vaccination status was not possible. Further, the survey was done during the roll-out of the COVID-19 vaccine in South Africa, thus vaccine uptake will currently likely be much higher in this setting, given the interest in vaccination among those reporting to be unvaccinated at the time of the survey. Nevertheless, at the time of writing this article, only 40% of adults 18 and older were vaccinated in Mpumalanga . The study site is located in a rural part of South Africa, so the results are mostly generalizable to other rural populations in South Africa and the southern African region. However, vaccine access is much higher in South Africa than in its neighbour countries.
Increasing vaccination in South Africa beyond current levels will require a concerted effort to address concerns around vaccine safety and increase confidence in vaccine efficacy. Understanding the local context and barriers to uptake as vaccines are being rolled out is imperative for effective messaging that addresses real concerns about vaccines, engages with hesitancy sensitively and effectively, and makes it clear where and when to obtain them. Further, we recommend continuing to provide vaccines at convenient times (e.g., outside of working hours), through mobile vaccination clinics, in community venues, and in workplaces, which will help to reach those in need (e.g., younger populations, men, and vulnerable populations), particularly as COVID-19 variants of concern continue to sweep through the region.
The authors would like to acknowledge the Human Research Ethics Committee at the University of the Witwatersrand and the Institutional Review Board of the University of North Carolina.