Seroprevalence and risk factors of COVID-19 in healthcare workers from 11 African countries: a scoping review and appraisal of existing evidence

Abstract A better understanding of serological data and risk factors for coronavirus disease 2019 (COVID-19) infection in healthcare workers (HCWs) is especially important in African countries where human resources and health services are more constrained. We reviewed and appraised the evidence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) seroprevalence and its risk factors in HCWs in Africa to inform response and preparedness strategies during the SARS-CoV-2 pandemic. We followed the Preferred Reporting Items for systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines in this scoping review. Databases including PubMed, Embase and preprint servers were searched accordingly from the start of the COVID-19 pandemic to 19 April 2021. Our search yielded 12 peer-reviewed and four pre-print articles comprising data on 9223 HCWs from 11 countries in Africa. Seroprevalence varied widely and ranged from 0% to 45.1%. Seropositivity was associated with older age, lower education, working as a nurse/non-clinical HCW or in gynaecology, emergency, outpatient or surgery departments. Asymptomatic rates were high and half of the studies recommended routine testing of HCWs. This scoping review found a varying but often high SARS-CoV-2 seroprevalence in HCWs in 11 African countries and identified certain risk factors. COVID-19 public health strategies for policy and planning should consider these risk factors and the potential for high seroprevalence among HCWs when prioritizing infection prevention and control measures and vaccine deployment.


Introduction
The coronavirus disease 2019 (COVID-19) pandemic has put health systems worldwide to an unprecedented test. As frontline workers, healthcare workers (HCWs) are a critical part of the health system's pandemic response and are themselves at high risk of infection. Vaccine roll-out, therefor, prioritizes HCWs worldwide (WHO, 2020b), and at the time of writing, 35 African countries had begun vaccine campaigns (Our World in Data, 2021).
Serological studies are investigations using serology testing to look for antibodies in blood (Centers for Disease Control and Prevention, 2021). These studies give insight into the true burden of infection as they have the ability to account for asymptomatic and unreported cases. These studies can provide data on infection trends, effects of interventions, vaccine deployment prioritization guidelines (Maeda and Nkengasong, 2021) and geographical distribution mapping for identification of populations at particularly high risk (Peeling et al., 2020).
A previous systematic review on seroprevalence in HCWs only included three African studies (Galanis et al., 2020). Nevertheless, data on seroprevalence and risk factors are specifically important in the African continent, where countries have predominantly lower access or quality of healthcare (Our-World-in-Data, 2015). This region furthermore experiences constrained human resources in healthcare. Compared to Europe, where there are 80 nurses per 10 000 population, across Africa there are only 10 (WHO, 2021a), emphasizing the high need to protect this scare workforce. Evidence on the implementation of measures that can help detect and mitigate exposures such as triage systems, dedicated COVID-19 wards, as well as surveillance testing and infection prevention and control (IPC) availability is necessary to better understand the risk of becoming overburdened.
This need is further emphasized in a region that does not have ownership of vaccine production and must depend on limited doses from other areas. Serological testing can also provide guidance on vaccine decision-making.
This scoping review aimed to better understand and identify gaps in knowledge on seroprevalence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) antibodies in HCWs in the African continent and their risk factors for COVID-19 infection, as well as how these seroprevalence rates compare with the respective general population. The added public health value of this review is that it provides the most up to-date synthesis of all available research and hence can inform public health strategies for policy and planning.

Materials and methods
Given the relative infancy and fast-moving pace of the field, we selected a scoping review method without prior registration in a public domain.

Data sources and search strategy
This scoping review was performed according to the Preferred Reporting Items for systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) Checklist (Tricco et al., 2018) and frameworks on scoping reviews (Levac et al., 2010;Arksey and O'malley, 2005;Peters Mdj et al., 2020) including the stages of (1) identifying a research question, (2) identifying relevant studies, (3) study selection, (4) charting the data and (5) summarizing and reporting results. The detailed PRISMA checklist is available in Supplementary Table S1.
We searched PubMed, Embase and preprint servers (ArRvix, BioRvix, ChemRvix, MedRvix, Preprints.org, ResearchSquare and SSRN) for designated terms for COVID-19, seroprevalence, HCWs and African countries. The complete search strategy is detailed in supplement 2. A hand-search was also conducted to add relevant articles not previously identified with the search strategy. Additionally, to put the HCW seroprevalences into context, general population seroprevalences were gathered from the SeroTracker database (Arora et al., 2021) for all countries included on 24 October 2021.

Selection and eligibility criteria
We screened title and abstracts of all publications returned in the search, guided by pre-set inclusion and exclusion criteria detailed as follows. We included articles published or in preprint between the first detection of COVID-19 and 19 April 2021 on HCWs irrespective of prior COVID-19 status or comorbidities in health settings in Africa with data on serological status. We excluded reviews and evidence that focused on non-HCWs or outside of Africa. HCWs were not defined in this scoping review, but we followed the definition of included studies, being mainly 'clinical and non-clinical care'. We did not restrict our search in terms of publication language, although search terms were all in English. Two researchers independently undertook study selection, and discussion resolved potential disagreement. Selection of studies for general population seroprevalence was restricted to the respective counties and overall study period.

Data extraction and quality assessment
Data from all selected studies were entered independently by two researchers into a form created for this purpose in Microsoft Excel. The following data were charted: author, publication year, title, journal, publication status, study period, city/country, sample size, sampling strategy, response rate, inclusion criteria, exclusion criteria, study design, sex, age, seroprevalence, asymptomatic rate, sample collection, testing strategy, testing rate, type of antibody, type of test, test name, factors investigated, factors associated, level of analysis and additional data such as reported triage system, dedicated COVID-19 wards, routine testing, IPC or personal protective equipment (PPE) availability.
The quality of studies was assessed by two authors independently using Joanna Briggs Institute (JBI) critical appraisal tool for prevalence studies (Munn et al., 2015) adapted with additional specifications (Arora et al., 2021). Bias levels were assessed using a 9-point ranking system with categorization of 8-9 low, 5-7 moderate and ≤4 high risk of bias (Galanis et al., 2020).

Selection of studies
The search identified a total of 102 articles; following the removal of duplicates and critical assessment of title and abstracts, 19 potentially relevant articles were identified for full-text evaluation ( Figure 1). Application of the pre-set eligibility criteria resulted in a final inclusion of 16 articles.

Quality assessment
Following the JBI critical appraisal tool, studies were rated out of a total of 9 points, and all were found to be of moderateto-low risk of bias (ranging between 5 and 8) (Supplementary Table S3). Risk of bias primarily arose from aspects such as test characteristics or non-reporting of response rates (Supplementary Table S3).

Risk factors
Ten studies investigated factors associated with seropositivity, eight of which found association reported by odds ratio (OR) or hazard ratio (HR) ( Table 1).

Gap analysis
None of the studies reported facilities having a COVID-19 triage in place. Only one specified existence of dedicated wards (Kassem et al., 2020) or that HCWs were specifically trained in COVID-19 IPC measures . Five studies raised concerns about insufficient PPE availability or asked to increase supply (Majiya et al., 2020;Mukwege et al., 2021;Etyang et al., 2021;Fwoloshi et al., 2021;Rusakaniko et al., 2021), whereby one study asked for PPE especially for HCWs with patient contact (Rusakaniko et al., 2021). Fifty percent of included studies stated a lack of routine testing of HCWs or recommended its implementation due to the high seroprevalence and asymptomatic rates.

Discussion
The overall seroprevalence among HCWs across 11 Africa countries varied widely reflecting recent data from a systematic review of seroprevalence of anti-SARS-CoV-2 antibodies in Africa, which combined the general population and specific working groups (Chisale et al., 2021).
Direct comparison with the general population is challenging and was not the primary aim of the current manuscript, but data available from the general population in the 11 included countries suggest highly varying seroprevalences similar to those of HCWs with rates between 0% and 44.6%. The seroprevalence among HCWs could be influenced by variety of factors that may differ between levels (national and local) and countries. National Public Health Institutes vary widely in breadth and depth (Africa CDC, 2019), especially in West Africa, where the strengthening of public health structures (Meda et al., 2016) and surveillance is urgently needed (Massinga Loembé et al., 2020). This need could be reflected in the suspected underreporting of SARS-CoV-2 cases. The Africa Centres for Disease Control and Prevention (CDC) together with Ministries of Health have recognized this challenge and taken action with their joint continental Strategy and the creation of Africa Task Force for Coronavirus (Africa CDC, 2020).
HCWs in Africa are potentially more often in contact with undiagnosed cases due to low surveillance and limited testing capacities (Chitungo et al., 2020). Furthermore, the clinical COVID-19 diagnosis is even more challenging due to other endemic febrile diseases such as malaria or typhoid fever, which also act as competing priorities for frontline HCWs (Maze et al., 2018). These competing priorities make it difficult for hospitals to implement triage and infectious disease treatment centres specifically dedicated for the COVID-19 pandemic (Ayebare et al., 2020). This aspect is reflected by the fact that only one included study reported having a COVID-19 ward and none to have a triage system in place.
Additionally, even where cases are accurately diagnosed, self-protection of HCWs is difficult due to IPC measure limitations in infrastructure and equipment, such as overcrowded settings (van de Ruit et al., 2020), in which proper distancing is more challenging and lacking supply or access to PPE material (Desai et al., 2019). This lack of PPE availability or the need for increased supply was confirmed by five of the included studies. However, in contrast to studies in the UK or USA (Nguyen et al., 2020), no significant association between IPC and seropositivity among HCWs in Africa was shown in our review (Abdelmoniem et al., 2021;Kassem et al., 2020;Mukwege et al., 2021). This lack of association could arise from the assessment procedure through questionnaire only. Observation should be used in conjunction with questionnaires to realistically measure the impact of IPC on infection mitigation (Sax et al., 2009).
The highest HCW prevalence rate was found in Nigeria (45.1%) and supports reports that this country has Africa's second largest population incidence (WHO, 2021b) This high rate was also reflected in the comparative general population data that showed Nigeria to be second highest after South Africa. Providing care with minimal precautionary measures was a reported factor that is likely influencing high seroprevalence rates among HCWs in particular (Olayanju et al., 2020).
Three of the included studies reporting a HCW seroprevalence <5% were conducted in April 2020, before the first wave peaked in Africa (WHO, 2021b).
Detection and subsequent isolation of SARS-CoV-2 cases is known to be especially challenging when asymptomatic rates are high. In this scoping review, asymptomatic rates were predominately >60% (Goldblatt et al., 2021;Kassem et al., 2020;Mostafa et al., 2021;Mukwege et al., 2021) and hence much higher than in a current meta-analysis (He et al., 2020). These high asymptomatic rates warrant extra vigilance in case detection strategies since asymptomatic HCWs can not only infect their peers, colleagues and families but also the vulnerable group of patients they are caring for. These high asymptomatic rates are in particular worrisome in the African region where continuous mass screening of HCWs is rarely done, asymptomatic infections are broadly undetected and hence HCWs are more likely to transmit the infections onwards. Consecutively, half of the included studies recommended the implementation of routine screening of HCWs to avoid hidden onward transmission. Routine testing in practice can hence mitigate the possibility of HCWs being a potential source of infection to the vulnerable patients and protect the scarce workforce of HCWs in low-resource settings.
In the African region, significant association was shown between seropositivity and working in operating rooms, obstetrics, outpatient clinic, gynaecology and emergency department (Mostafa et al., 2020;Olayanju et al., 2020;Rusakaniko et al., 2021). These risk areas are in accordance with results from Denmark that found emergency units to be a high-risk area (Iversen et al., 2020). One included study found that nurses and non-clinical HCWs were at higher risk of seropositivity than doctors (Mostafa et al., 2020), supporting our found association between seropositivity and lower level of education or the non-prioritization for IPC trainings. These associations show that non-clinical HCWs should also be included in training regardless of their level of patient contact. Interestingly, a study from Cairo found that exposure to a confirmed case at home had a significantly higher HR than exposure at work (Mostafa et al., 2021), suggesting that HCWs' non-occupational risk should also be considered. This finding supports the determined general population seroprevalence range being similar to that of HCWs. Seroprevalence was associated with higher age in three included studies (Rusakaniko et al., 2021;Mostafa et al., 2020;, which is concerning since higher age is known to be associated with higher risk for severe disease (ECDC, 2021).

Limitations
There are several limitations in the present scoping review. No grey literature of news and press releases were included. We searched in databases and public preprint servers to assess as much available scientific information as possible. As studies from only 11 countries could be identified, extrapolation of findings to all HCWs in the African continent could not be made. However, all African regions were represented, although data on risk factors mainly stemmed from Egyptian studies. Most studies used total population or random sampling, and none of the studies were rated to be of high risk of bias. No study explicitly included HCWs on sick leave or in quarantine leading to a potential underestimation. A valid and direct comparison between HCW and general population seroprevalence rates was challenging as neither methods nor timing of most studies were aligned.
Serological studies per se have some limitations. Given antibody kinetics, infections can be undetectable for up to 2 weeks from the onset of infection or after 3 months (Isho et al., 2020). In addition to antibody kinetics, infection through immunoglobulin G testing may also be undetected as some individuals may not develop antibodies (Mostafa et al., 2021). Overestimation, on the other hand, could be possible due to cross-reactive antibodies that have been detected in sub-Saharan Africa (Tso et al., 2020). Test characteristics and test validity also affect seropositivity. Ten of the 16 studies used tests that met the commonly set minimum performance criteria of sensitivity ≥90% and test specificity >95% (Biocentury, 2020). These varying tests are leading to potentially divergent results that stress the need for validation in study populations, especially in African countries, where the concern of potential lower specificity of SARS-CoV-2 commercial tests has been raised (Nkuba Ndaye et al., 2021). Confirmation through neutralizing antibody testing and interpretation with caution has therefore been recommended (Nkuba Ndaye et al., 2021).

Public health contribution
To our knowledge, this is the first scoping review including all preprint and published studies estimating seroprevalence and risk factors for SARS-CoV-2 seropositivity among HCWs from 11 African countries with evidence appraisal. The present scoping review has shown that data on seroprevalence in HCWs across the African continent are scarce and hence data on risk factors even scarcer. Even though there is no direct and valid comparison, this review offers the first hint that differences between seroprevalences of HCWs and the general population are not as prominent as in other regions (Nguyen et al., 2020). Little standardization of study protocols, including harmonization of tests, is done, and consecutively, seroprevalence rates and risk factors are only marginally comparable between studies and countries. This scoping review therefor stands as a roadmap to compiling and understanding available data on the COVID-19 burden in HCWs across Africa, a region in need for increased investment in research (Guleid et al., 2021). This roadmap provides the most up-to-date synthesis of all available research and its gaps. It is a beneficial resource for larger projects such as the planned representative national studies by the Africa CDC (Maeda and Nkengasong, 2021) and surveillance unity studies by the World Health Organization (WHO, 2020a). This serological data will help to assess the true burden and risk factors among HCWs across the African continent and hence provide knowledge to more strategically apply limited resources in response aspects such as IPC measures and vaccine programmes. Furthermore, serological data provides the possibility to make vaccination prioritization more efficient with adaptations such as prioritization of HCWs without detectible antibodies or single-dose vaccination of those who are seropositive (Krammer et al., 2021). Such a strategic development of the vaccination programme could better utilize limited vaccination resources given the high seroprevalence rate of HCWs found in some African countries.

Conclusion
This scoping review found a varying, but often high SARS-CoV-2 seroprevalence among HCWs in the African region. HCWs of older age and lower education, as well as nurses and those working in gynaecology, emergency department, outpatient clinic and operating rooms are at even higher risk. These results point to a clear need to target specific risks of HCWs, whether they arise due to their social position or due to increased exposure of certain functions/settings within the clinic. Further research is needed to better understand which factors lead to increased risk of certain cadres of health workers. Asymptomatic rates were high and half of the studies recommended routine testing of HCWs. Public health strategies should take these risk aspects and the partly high seroprevalence among HCWs into account when prioritizing IPC measures and high-risk groups for vaccine deployment.

Abbreviations
CI: Confidence interval HCW: Healthcare worker HR: Hazard ratio IPC: Infection prevention and control JBI: Joanna Briggs Institute NA: Not applicable NR: Not reported OR: Odds ratio PPE: Personal protective equipment PRISMA-ScR: Preferred Reporting Items for systematic reviews and Meta-Analyses extension for Scoping Reviews SARS-CoV-2: severe acute respiratory syndrome coronavirus type 2

Supplementary data
Supplementary data are available at Health Policy and Planning online.

Data availability
The data underlying this article are available in the article and in its online supplementary material.

Funding
No funding is received for this work.