This study primarily aims to determine the prevalence of COVID-19 vaccine hesitancy among healthcare providers in Nigeria. The secondary objective of this review was to identify the factors contributing to COVID-19 vaccine refusal among Nigerian healthcare workers.

The research question was; what is the prevalence of COVID-19 vaccine hesitancy among Nigerian healthcare workers?

This review included observational studies involving healthcare providers from various states in Nigeria who refused COVID-19 vaccination. These providers work in healthcare facilities and deliver care directly (e.g., doctors, nurses) or indirectly (e.g., pharmacists, laboratory scientists) [19]. Publications focusing on medical students, hospital administrative staff, non-observational studies on healthcare workers, animal studies, and studies on the general population were excluded. Healthcare workers represent a heterogeneous population; therefore, studies that were similar in terms of the participants were grouped for analysis.

The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [20] and is registered in the PROSPERO database (registration number CRD42022365489). The review protocol is available in Ibom Medical Journal at http://dx.doi.org/10.61386/imj.v16i2.305. A comprehensive literature search was conducted across four major databases: PubMed, Cochrane Library, Scopus (via Publish or Perish software), and the African Index Medicus [21]. The search terms were based on the condition under study (COVID-19 vaccine hesitancy): context (Nigeria) and population of interest (Healthcare workers) to retrieve relevant articles published from March 1st, 2021, to March 27, 2022, without language restriction. In response to the COVID-19 pandemic, Nigeria received its first shipment of vaccines from the COVID-19 Vaccines Global Access Facility (COVAX) on March 2, 2021 [7]. The country implemented a four-phase National Deployment and Vaccination Plan (NDVP), with the first phase prioritizing healthcare workers and other frontline personnel [22]. The authors anticipated that reports on vaccine non-acceptance would surface within a year of the vaccine’s introduction and applied a date filter, though the review extended beyond 1 year. However, all articles found during this period were written in English, with no available translations. The search string applied in PubMed is as follows: (COVID-19 OR COVID-19 OR Coronavirus OR COVID OR SARS-CoV-2 OR sars-cove-2) AND (Vaccination OR vaccine OR vaccine* OR immunisation OR immunisation) AND (“healthcare workers” OR “health personnel” OR physician OR nurse OR doctor OR residents OR pharmacist OR “laboratory scientist” OR “lab technician”) AND (Rejection OR hesitancy OR compliance OR attitude OR refusal OR non-acceptability) AND Nigeria. Records found were reviewed to determine whether they met the inclusion criteria. We also performed a hand search for grey literature, conference abstract proceedings, reference lists of the included publications, and citations in Google Scholar.

Three independent co-authors (CO, UA, and TA) screened the title and abstract, followed by a full-text screening of the articles identified based on predefined eligibility criteria. Studies were included if they assessed the prevalence of COVID-19 vaccine hesitancy among healthcare providers (medical doctors, public health officers, chemists, medical laboratory scientists, nurses, and others) [21, 23, 24] through observational designs in any of the 36 states of Nigeria plus the federal capital territory [22]. Of the 389 articles identified, 250 duplicate items were removed and 117 were eliminated after title and abstract screening. We excluded studies unrelated to COVID-19 vaccine hesitancy, studies on general populations or non-observational designs, and those conducted outside Nigeria, resulting in 22 abstracts for full-text review. Three articles were excluded during full-text screening (see Figure 1; PRISMA flow diagram). Author disagreements were resolved through discussion.

Quality assessment utilized the adapted Newcastle–Ottawa Scale for cross-sectional studies [25]. This scale is graded on a 10-point scale and consists of three domains. Domain 1 evaluates the methodological quality of each study (with a maximum of 5 stars), domain 2 assesses the comparability of the studies (with a maximum of 2 stars), and domain 3 evaluates the outcome measures and related statistical analyses (with a maximum of 3 stars.) [25]. Furthermore, the review categorised the overall quality of the studies into three groups: low risk of bias (scored 7–10), moderate risk of bias (scored 5–6), and high risk of bias (scored 0–4). This assessment process was conducted by three independent co-authors (EE, TA, and OI), and the final score for each study was determined by averaging their assessments. Any discrepancies that arose during this process were resolved through discussion. See Supplementary Material S1.

Data were independently extracted by three co-authors (OI, WO, and TA), capturing authorship, publication year, research location, study design, population, and sample size. The main outcome measures (proportion of vaccine hesitancy) and details related to the secondary objectives of the review (factors contributing to COVID-19 vaccine hesitancy) were also abstracted. For studies that reported the outcome of interest (vaccine hesitancy) in percentages along with sample sizes [26–31], we manually computed the absolute prevalence rate. In one study with missing data on the main outcome [12], we attempted to contact one of the study authors but did not receive a response. Despite this, the study was still included in the qualitative synthesis. Data extraction was performed using Microsoft Office Excel, which was prepared by the team. To ensure the appropriateness of the Excel sheet, we piloted it using four studies. In cases of disagreement, resolution was achieved through discussion.

A single-arm meta-analysis with a random-effects model was used to estimate the pooled prevalence of COVID-19 vaccine hesitancy among healthcare providers in Nigeria at a 95% confidence interval presented in a forest plot. The relative weight of each study and the prediction interval are also presented. Separate estimations were made for the proportions of COVID-19 vaccine hesitancy and factors influencing vaccine uptake. Factors contributing to COVID-19 vaccine hesitancy were subjected to meta-analysis if they had been assessed and data from at least two studies were available. To enhance variance stability, the proportions were transformed using the Freeman–Tukey Double Arcsine Transformation method [32, 33]. The assessment of heterogeneity was performed using the Cochrane Q statistic to determine data variability and its statistical significance from naught, Tau, and Tau² evaluated by analysing the variance in effect size measurement and the inconsistency index (I²) [32, 33]. The I² statistic was interpreted based on Higgins and Thompson’s classification, in which percentages of 25%, 50%, and 75% were considered indicative of low, moderate, and high heterogeneity, respectively [32]. Funnel plot asymmetry, Egger’s regression, and Begg’s rank correlation tests were used to assess publication bias. Subgroup analysis was performed on studies that recruited a cadre of HCWs (medical doctors). All statistical analysis were carried out in JAMOVI 2023 version 2.4.5 [34].

The articles included in this study were published in 2021 (n = 6) and in 2022 (n = 13). The studies involved 20,724 participants and were conducted across six geopolitical zones and states in Nigeria and published in English. All studies employed a cross-sectional design with physicians, nurses, medical laboratory scientists, pharmacists, physiotherapists, radiographers, optometrists, community health workers, and dental technicians as study participants. The majority of them (79%) were assessed to have low quality but were included in the review. An overview of the included studies is presented in Table 1.

Common determinants of vaccine hesitancy included fears of side effects, lack of trust, and safety concerns. Additional reasons cited for vaccine reluctance included beliefs that the vaccine contains harmful substances, concerns over effectiveness, adverse events following immunization, vaccine had not undergone sufficient clinical trial, government ulterior motive, social media influence, fear of biological chips, conspiracy theory, faulty storage, no exposure to COVID-19, health concerns, RNA component of the vaccine, difficulty in vaccination request, and religious belief. These were not amenable to quantitative analysis.

Several studies in the review [9, 11, 14, 26, 29, 30, 44] revealed that many HCWs were hesitant to accept the newly introduced COVID-19 vaccine due to uncertainties about its safety, which became a major barrier to vaccination uptake. Some of these studies did not report the proportion of participants whose reasons for refusal bordered on safety concerns, but it was common knowledge that the general population dreaded vaccination due to safety concerns.

An essential characteristic influencing vaccine acceptance is the side effect profile. Five studies [13, 30, 36, 37, 44] mentioned this outcome, but the actual proportion was not reported. Other studies have also documented that HCWs were deterred from accepting COVID-19 vaccination because of untoward effects [45, 46].

Negative social media reports and beliefs that vaccines were manufactured to wipe out Africans, were a premise for the non-acceptance of COVID-19 vaccination among HCWs [13, 14]. Conspiracy theories, including government ulterior motives, AstraZeneca not being genuine in Nigeria, fear of biological chips, and fears of vaccines containing dangerous substances are known determinants of vaccine hesitancy among the Nigerian health workforce [14, 39].

Vaccine hesitancy among Nigerian HCWs is linked to several other factors. These factors include lack of exposure to the COVID-19 virus, fear of unknown origin, lack of effectiveness, adverse events following immunisation, belief that the vaccine had not undergone sufficient clinical trials and the MRA component of AstraZeneca [14]. In a 2022 study by Emmanuel et al., health concerns, such as blood clots in women (21%), allergic reactions (25%), and innate immunity issues (28%), were identified as significant deterrents to COVID-19 vaccine acceptance among healthcare workers in Nigeria [13].

The forest plot depicting the prevalence of vaccine hesitancy due to safety concerns in the included studies yielded a p-value of <0.001, indicating statistical differences in the effects among the studies. Safety concerns were prevalent at 68% (CI: 0.047–0.89). The effect size variance measure was Tau = 0.213, with (SD: Tau² = 0.0452) when, compared with the overall effect size of 0.68 across the included studies. The variability in the meta-analysis, expressed as I², was 98.59%, primarily due to the variance in the observed events within the studies. See Supplementary Material S7 for the statistics and forest plots.

The GRADE framework was applied to assess evidence quality in this systematic review and meta-analysis [50, 51]. Observational studies begin with a low evidence rating, which may be adjusted based on specific criteria. As there was no dose-response relationship but rather observed variability due to confounding, the findings could only be rated down. Hence, we examined these domains to assess the robustness of the evidence: risk of bias, inconsistency, indirectness, imprecision, and publication bias.

Risk of Bias: The review relied mainly on cross-sectional studies, a design prone to bias. The Newcastle–Ottawa Scale used for the assessment showed that, 79% of the studies were of low quality, which may be due to non-representative sampling and lack of control of confounders. The overall risk of bias was rated further down in this domain.

Inconsistency: The pooled prevalence of COVID-19 vaccine hesitancy among healthcare workers showed substantial variability (I² = 99.69%) that remained unchanged after a subgroup analysis. This result may be attributable to differences in study populations, regions, and methods. Hence, the evidence was downgraded by one level.

Indirectness: All studies Included in the review recruited healthcare workers in Nigeria to evaluate COVID-19 vaccine hesitancy, so there were no concerns about indirectness. The evidence was not downgraded in this domain.

Imprecision: The evidence was downgraded in this domain as the confidence interval was wide: point estimate 75% (95% CI: 61%–88%). Additionally, the sample sizes showed wide variation especially in smaller studies, leading to concerns about imprecision.

Publication Bias: Funnel plot analysis and Egger’s test did not reveal significant publication bias. Hence no downgrading was applied in this domain.

The overarching certainty of evidence for the prevalence of COVID-19 vaccine hesitancy among healthcare workers in Nigeria was judged as low, as a result of serious concerns about bias, inconsistency, and imprecision. Although the studies were relevant, the high heterogeneity and methodological limitations led to low confidence in the level of evidence. Future high-quality, longitudinal studies are needed to improve the certainty of evidence in this area.

A thorough literature search was conducted to identify current evidence on COVID-19 vaccine hesitancy rates among HCWs in Nigeria. Findings indicate a high rate of vaccine hesitancy among healthcare providers, with an estimated prevalence of 75% (95% CI: 61%–88%, I² = 99.69%, P < 0.001). Primary reasons for hesitancy include concerns over side effects, distrust in vaccine safety, fear of unknown origins, and perceptions of insufficient clinical testing. These obstacles and myths preventing frontline workers from accepting COVID-19 vaccines need to be urgently addressed to improve vaccine uptake among the general population. Future observational studies should adopt written protocols to minimise variability and ensure comprehensive outcome reporting. A systematic review of longitudinal studies could enhance the evidence base on COVID-19 vaccine hesitancy among HCWs.

This is the first systematic review and meta-analysis of COVID-19 vaccine hesitancy among Nigerian HCWs. However, it has some limitations. First, the substantial heterogeneity observed in this meta-analysis (I² = 99.69%) highlights the diverse nature of the included studies. Differences in study settings, healthcare worker categories, data collection approaches and the timing of vaccine rollouts likely contributed to this variability. Despite subgroup analyses and sensitivity testing, significant residual heterogeneity persisted, suggesting unmeasured variations in study methodology.

This review primarily included cross-sectional studies, limiting causal inferences between healthcare providers’ characteristics and vaccine hesitancy [78]. Additionally, none of the included cross-sectional studies offered a protocol that would have helped to guarantee the validity of the findings.

Thirdly, many studies were rated as high risk of bias, potentially inflating the reported prevalence of hesitancy. Publication bias, as indicated by funnel plot asymmetry, may also have influenced our results. This could have skewed the pooled estimate of vaccine hesitancy.

Furthermore, the confidence intervals in several studies were wide, reflecting imprecision in the estimates of vaccine hesitancy. This imprecision was especially pronounced in studies with small sample sizes, contributing to uncertainty around the true prevalence of hesitancy.

This review is limited in its generalizability beyond the Nigerian healthcare context. Furthermore, the heterogeneity in healthcare worker roles included in the studies (e.g., doctors, nurses, pharmacists) suggests that the findings may not be universally applicable to all healthcare professionals.

Throughout the pandemic, people have expressed their opinions and concerns about the vaccine. Thus, eliciting patient perspectives from qualitative evidence may be worthwhile. Despite these limitations, the authors argue that heterogeneity is common in meta-analyses of prevalence studies and should not be used to judge the quality of evidence. Another area for improvement while conducting this systematic review is the need for more access to relevant databases. Although subscription-based databases were not searched, the authors believe that all relevant articles were discovered during the review period, as manual searches were also conducted.

This systematic review followed the PRISMA guidelines, but a few deviations were noted from the published protocol [24]. The Web of Science and Embase databases were not searched, and a different statistical package (JAMOVI software) was used for statistical analysis instead of Review Manager Version 5.4. The RevMan software was designed for intervention studies, but a meta-analysis of prevalence studies was conducted.

This systematic review and meta-analysis relied on secondary data from cross-sectional studies across Nigeria. To enhance quality, future prevalence studies should adhere to standardized protocols and methodological rigor. A heterogeneous healthcare population was recruited from various study centres of which all are not frontline workers. Given the high morbidity and mortality rates associated with the COVID-19 pandemic and the, crucial role of physicians and nurses in flattening the curve, knowing the proportion of physicians and nurses who are hesitant about vaccines would be useful in planning and implementing strategies to increase their uptake. A mixed-method systematic review and meta-analysis is required to address the various myths surrounding vaccine acceptance.