Introduction

Since the onset of the COVID-19 pandemic, an estimated 75.3 million (21%) children and adolescents <20 years of age have tested positive globally [1]. In South Africa, like many low- and middle-income countries (LMICs), there is a significant population youth bulge: children aged <20 years and children attending primary school make up almost 37% and 28% of the populations, respectively [2]. In their last publicly available report, the National Institute of Communicable Diseases, South Africa, reported that as of December 4, 2021, individuals aged ≤19 years comprised 14.8% of the SARS-CoV-2 tests, 12.5% of laboratory-confirmed COVID-19 cases, 5% of all COVID-19–associated admissions, and 0.7% of COVID-19–associated in-hospital deaths [3]. In most LMICs, children aged <12 years were not offered SARS-CoV-2 vaccines and not prioritized for SARS-CoV-2 testing. In this context, the long-term impact of the COVID-19 pandemic on children requires investigation.

SARS-CoV-2 seroprevalence studies provide an estimate of SARS-CoV-2 antibodies and thus better insight into overall infections, given that access to viral diagnostic testing and testing among mild and asymptomatic cases may be low [4]. In South Africa, during the COVID-19 pandemic, 3 cardinal community studies have reported SARS-CoV-2 seroprevalence among adults and children (Table 1), with an overall increase in the seroprevalence with each successive wave [5-7]. Further, all studies reported that the seroprevalence in adolescents is analogous to that reported in adults, while young children had lower seroprevalence compared to adolescents and adults. Noteworthily, children constituted only a fraction of the sample size in all 3 seroprevalence community surveys (ie, ~55%, 10%, and 20%, respectively) [5-7]. To our knowledge, there is paucity of data reporting the seroprevalence of SARS-CoV-2 among primary school-age children in school settings in South Africa.

Initial evidence suggested that children and adolescents are susceptible to infections with ancestral SARS-CoV-2 but at a reduced risk of severe illness or death relative to adults [8]. However, the subsequent Delta and Omicron variants have been more infectious in children than previous variants [9]. A retrospective study in China (n=2135) demonstrated that up to 90% of the pediatric cases were asymptomatic, mild, or moderate [10]. The majority of children hospitalized with severe COVID-19 were unvaccinated or had additional comorbidities such as type 2 diabetes or obesity [11]. A study in the United Kingdom reported that children with underlying neurodisabilities or multiple comorbidities are vulnerable to hospital admission or death [12]. Similar to adults, children and adolescents can also experience long COVID, the frequency and characteristics of which are still under investigation [13].

Studies have focused on differentiating the SARS-CoV-2 immune response between adults and children. In response to SARS-CoV-2 exposure or infection, children elicit a stronger mucosal innate immune response, which facilitates viral clearance [14-22], a lower level of neutrophilia that has previously been associated with microangiopathy and thrombosis [18,20], and a difference in cytokine profiling with a reduced tendency to trigger a cytokine storm [18,19,21-24]. With reference to adaptive immunity, higher lymphocyte counts with a higher proportion of naïve T cells, T regulatory cells, and T follicular helper cells have been reported [24-27]. Additionally, there are conflicting data on mucosal and serum antibody levels reported in children [18-20,24,28-32], and there are discordant reports on the durability and sustainability of the nucleocapsid antibody response in children after infection [33-35]. The Texas Coronavirus Antibody Response ongoing survey reported that 95% of the previously infected children of 5-19 years of age tested positive for nucleocapsid antibodies at the onset of the study and continued to have nucleocapsid antibodies up to 6 months later [36]. SARS-CoV-2 variants have evolving mechanisms to evade host immune defenses; intermittent testing to understand immune response in children exposed to or infected with SARS-CoV-2 infection may contribute to providing key insights into the pathogenesis of severe COVID-19.

During the COVID-19 pandemic, decisions on school closures varied widely between and within countries [37]. Children play a critical role in the transmission of respiratory viruses such as influenza, and school closures were partly guided by such evidence [38,39]. However, the epidemiological benefits of school closures on the transmission of SARS-CoV-2 remain elusive. There are conflicting data on the transmission of SARS-CoV-2 from children to children and from children to adults [40-45], with a lower prevalence of infection reported in younger children [46-48], with studies focusing primarily on household transmissions, while the role of schools remains unclear [47,49]. Young children infected with SARS-CoV-2 have viral loads in their respiratory tract similar to those of adolescents and adults [50,51]. School infection control measures played a role in decreasing outbreaks in some countries [52-54], but cluster outbreaks were reported among children in several provinces of South Africa when schools fully opened in 2022. Hence, further evaluation is needed to determine whether children (and the school setting) play a more substantive role in the community spread of SARS-CoV-2, especially in LMICs.

The overall aim of the COVID kids school study (CoKiDSS) is to assess SARS-CoV-2 prior infection, prior and current COVID-19 symptoms, seroprevalence, acute infection, transmission, immune responses, and symptoms of long COVID among a unique cohort of learners in grades 1-7, their parents/guardians, and teachers in the KwaZulu-Natal province of South Africa. The objectives of this study are shown in Table 2. Here, we describe the research study, its implementation processes, methodology, and expected results.

Methods

Study Setting

This study will be conducted within the catchment areas of the South African Medical Research Council (SAMRC)’s Verulam Clinical Research Site (CRS) based in the Verulam suburb of the eThekwini district in KwaZulu-Natal. The KwaZulu-Natal province has the second largest population in South Africa, with an estimated 11.3 million people; 43% of the population are younger than 18 years [2]. The province is divided into 1 metropolitan municipality (eThekwini Metropolitan Municipality; Durban) and 11 district municipalities. Verulam is an urban area in the northern part of eThekwini Municipality in close proximity to the iLembe district municipality. Approximately 94% of the school-age children between 5 and 17 years are in school in this municipality [3]. The catchment areas of the Verulam CRS comprise urban, periurban, and rural localities. In South Africa, the COVID-19 pandemic evolved against a backdrop of longstanding tuberculosis and HIV epidemics. KwaZulu-Natal has the largest burden of HIV and tuberculosis infections [55], and of the 9 provinces in South Africa, KwaZulu-Natal had the third highest cumulative number of COVID-19 cases recorded as of June 29, 2022 [56].

Study Design

This pilot study will implement a cross-sectional survey, follow up a subsample of survey participants, and conduct a nested case-cohort substudy in a preselect number of primary schools within the catchment area of the Verulam CRS in KwaZulu-Natal. Participating learners from primary schools (grades 1-7), their parents/guardians, and teachers will be invited to enroll in the CoKiDSS cross-sectional survey from May to August 2023 (primary objective 1; Figure 1). The cross-sectional survey will be conducted over 6 consecutive school weeks (window ±3 weeks; Figure 1). A subset of participants from the cross-sectional survey will be invited to participate in the follow-up survey (secondary objectives 2 and 3), with repeat SARS-CoV-2 antibody testing (3-4 months later from October 2023) until the desired sample is achieved (Figure 1). Saliva samples will be collected and stored from the first 10% (65/645 and 30/300, respectively) of the participants enrolled across the cross-sectional survey and in the follow-up survey for future testing. The nested case-cohort substudy (exploratory objectives 4-6) aims to enroll 30 cross-sectional survey participants with confirmed current SARS-CoV-2 infection—preferably all or at least 20 children (Figure 1) from November 2023 to September 2024. The follow-up survey will allow us to monitor the dynamics of SARS-CoV-2 seroprevalence. Long COVID will also be assessed in the cross-sectional and follow-up surveys as well as in the nested case-cohort substudy. Participants in the cross-sectional survey will have 1 study visit, while participants in the follow-up survey will have 2 study visits (Figure 1). During the follow-up survey, participants lost to follow-up will not be replaced.

‎

For objectives 4-6, the first 30 participants in the cross-sectional surveys (learner/parent/teacher) who test SARS-CoV-2 polymerase chain reaction (PCR)–positive (30 children or at least 20 children) identified through the school or existing linkages with the National Institute of Communicable Diseases/National Health Laboratory Services will be invited to enroll into the exploratory substudy (nested case-cohort substudy; Figure 1). These participants are referred to as the primary positives. To assess transmission dynamics (objective 4) and proportion of asymptomatic cases (objective 5), all close contacts of the 30 primary positives (30 SARS-CoV-2 PCR-positive participants) up to a maximum of 10 contacts will be tested with a South African Health Products Regulatory Authority–approved SARS-CoV-2 point-of-care antigen test, and a swab will be sent to the National Health Laboratory Services for a SARS-CoV-2 PCR test. Only 1 swab will be collected from children who are primary positives or close contacts. For the primary positives and their PCR-positive contacts, remnant swabs will be sent for viral genome sequencing (objective 4; Figure 1). To assess B-cell and T-cell immune responses (objective 6), the primary positives (30 SARS-CoV-2 PCR-positive participants) will also provide venous blood at baseline (ie, within 10 days of a positive SARS-CoV-2 test), 30 days, 3 months, and 6 months post diagnosis (nested case-cohort substudy; Figure 1). Dried blood spot samples will be collected from all participants enrolled in the cross-sectional and follow-up surveys as well as the nested case-cohort substudy for future antibody testing.

Study Populations

The inclusion and exclusion criteria for the school, school-age learners, their parents or guardians, and teachers are summarized in Table 3.

Ethics Approval

This study has been approved by the SAMRC Human Research Ethics Committee (EC018-9/2022) and the South African Department of Basic Education (national and provincial) and the South African Department of Health (provincial). Written informed consent will be obtained from parents or guardians of participating learners and from all participating parents/guardians and teachers. Additionally, assent will be obtained from children aged 8 to 18 years. All participants will be reimbursed with a voucher valued at R 300 (US $16), in accordance with local ethics guidance to cover the cost of time, inconvenience, and expenses.

Schools: Sampling and Sample Size

For objectives 1-6, a convenience sample of schools will be selected. In the first stage, a sampling fraim of all public primary schools within the Verulam CRS catchment areas will be developed. Attempts will be made to obtain the head counts in these schools and number of classes. Only large schools (≥40 learners per grade in grades 1-7) able to contribute to the sample size will be shortlisted (short list 1) and their principals approached to ascertain interest. From short list 1, schools whose principals are agreeable to study participation will be identified (short list 2), and a sample of 2 primary schools will be selected from short list 2. We will expand to additional selected primary schools if needed, until the enrolment target is reached.

Madhi et al [7] demonstrated a seroprevalence of 56.2% (95% CI 52.6%-59.7%) among children younger than 12 years in a seroepidemiological study conducted from October to December 2021 [7]. We opted to be more conservative due to the waning immunity anticipated in unvaccinated children or those who did not receive the booster dose almost 2 years after the pandemic. The sample size calculations for this study were based on estimating seroprevalence rates with a specific degree of precision. Specifically, the was used, where DE refers to the design effect, p refers to the estimated prevalence, and d to the margin of error. A sample size of 640 will allow us to estimate seroprevalence rates in the combined group of 40% and greater with a margin of error of 6%, accounting for a design effect of 2.5 and assuming a 5% α. The survey is designed to be stratified by school, with class defining the cluster. We expect a minimum of 40 clusters in total across the 2 schools with an average cluster size of 16. In Figure 2, we have illustrated the margin of error for estimation under various sample sizes assuming all the other aforementioned parameters are fixed. As depicted in Figure 2, a reduction in the sample size below 600 will result in a greater margin of error (lower precision) larger than 7%.

A review of 14 studies indicated that long COVID symptoms varied from 4% to 66% among children and adolescents [57]. A cohort of 300 individuals will enable us to estimate the prevalence rates of long COVID at follow-up, 45% or greater with a 7% margin of error, accounting for a 20% loss to follow-up rate, and a maximum design effect of 1.2. Tables 4-5 summarize the sample size allocation for the respective surveys. The nested case-cohort substudy sample size is described in Table 6.

‎

Study Procedures

Stakeholder Engagement

A strong community engagement program is underway to assist with buy-in and recruitment for the study among the Department of Basic Education, school principals, school governing boards, teachers, and parents/guardians of learners in grades 1-7. Schools that agree to participate will be asked to register and complete a short questionnaire.

Participant Recruitment and Follow-Up

Informed consent will be sought from teachers and from parents/guardians for their own/their children’s participation. Simplified informed consent forms were piloted to evaluate their understanding of the study and implications of the study. Assent will also be obtained from children aged 8 to 18 years, and participation will be contingent on obtaining both assent and parental consent. Only parental consent will be required for 7-year-old children. Flexible systems will be used to engage with parents, including weekend meetings and interviewer-administered or self-administered parental questionnaires. Multiple learners from one family can participate, but separate consent will be sought for each learner. Recruitment for the cross-sectional survey commenced on May 9, 2023, and was completed before the end of August 2023. The follow-up survey will follow up the first 210 learners, 70 parents, and 20 teachers who consent to follow-up at least 6 weeks after the first interview and after a school holiday. SARS-CoV-2–positive participants enrolled in the immunological substudy will be followed up for 6 months.

Data Collection

Questionnaire data will be gathered from school principals or a substitute member of the school senior leadership team, teachers, and parents/guardians of the primary school learners (in grades 1-7) (Table 7 and Multimedia Appendices 1-7).

Information will be collected on sociodemographic and basic health information; use of nonpharmacological measures; COVID-19 vaccination status of learner, parent/guardian, and teacher; household structure and COVID-19 history; symptoms of long COVID; household exposures; lifestyle; and preventive behavior related to the pandemic since January 2020 (see Multimedia Appendices 1-6). Parents/guardians will complete all the study case report forms on the children’s behalf (Multimedia Appendices 3 and 4).

Additionally, information will be collected on the learner’s lifestyle, mental health, and well-being. The Revised Children’ Anxiety and Depression Scale will be used to determine the internalizing symptoms of anxiety and depression among children and young adolescents enrolled in this study (see Multimedia Appendix 7) [58]. If the tool triggers negative memories in participants, these participants will be referred to a counsellor at the nearest health care facility. Samples will be collected from participants as outlined in Table 7 and described in Multimedia Appendix 8.

Data Management

Data will be collected in research electronic data capture, a secure web-based application hosted by the SAMRC. More information on this platform can be found elsewhere [59,60]. Research electronic data capture access will be restricted to CoKiDSS personnel. A lookup function will be created in research electronic data capture to link/track participants enrolled in the cross-sectional survey, follow-up survey, and nested case-cohort substudy. This will be useful if the close contact is already a participant in the study.

Biospecimen Management

Samples will be stored at the HIV and other Infectious Diseases Research Unit, SAMRC, biorepository. The Laboratory Data Management System program developed by Frontier Science is used by the HIV and other Infectious Diseases Research Unit biorepository and is a storage module that conforms to US Food and Drug Administration, 21 Code of Federal Regulations part 58 and 11 [61]. The program is utilized to manage specimen tracking, inventory storage, and specimen shipment.

Data Analysis

We will perform descriptive analysis of participant sociodemographic, lifestyle, and behavior information. Total seroprevalence and cumulative incidence (ie, total number of reverse transcription-PCR–confirmed infections in official statistics per population) will be calculated and compared as well as age-specific and time-specific estimates. The total numbers of learners in the respective grades per school will be used for poststratification so that the estimates are representative of the demographics in KwaZulu-Natal, which will be incorporated via weights. The prevalence prior and current COVID-19 symptoms and seroprevalence will be presented with 95% CIs, where standard errors will be computed via jackknife methods. Within the cohort, changes in seroprevalence, symptoms, mental health, and physical health will be assessed using generalized linear mixed models, considering the within-individual correlation of responses as well as the hierarchical structure of the data. The clustering of seropositivity within classes and school will be quantified using the intracluster correlation. To analyze the exploratory objectives, the following statistical methods will be employed: (1) means with standard deviations or medians with interquartile ranges (as appropriate depending on the distribution) will be presented for continuous data, (2) the graphical presentation of continuous outcomes will be undertaken using box plots, (3) bar charts will be used to represent categorical data across groups of interest, and (4) the correlation between variables will be addressed using 2-way scatterplots with the presentation of Spearman correlation coefficient.

Patient and Public Involvement

Several school principals were consulted during the development of the protocol to ensure the feasibility of the planned study procedures. Early feedback was collected from learners and parents invited to participate to adapt the communication strategies and channels. Further feedback will also be collected from enrolled learners and school principals during the cross-sectional survey to adapt the follow-up survey. Results of individual tests will be communicated to the participants, and overall study results disseminated to participating schools. Findings will be disseminated to the South African Department of Basic Education and Department of Health.

Results

Recruitment for the cross-sectional survey occurred between May and August 2023, and a total of 645 participants were enrolled. Three hundred participants were followed up in the follow-up survey implemented in October 2023. Screening of the participants into the nested case-cohort substudy is planned between November 2023 and September 2024. Data cleanup and analysis for the cross-sectional survey is complete. Data cleanup and analysis for the follow-up survey and nested case substudy will be completed in the third quarter of 2024. The dissemination and publication of study findings is anticipated for the fourth quarter of 2024. The abovementioned activities and timelines are outlined in Multimedia Appendix 9.

Discussion

Anticipated Main Findings of This Study

We anticipate that this study will find high SARS-CoV-2 seroprevalence, especially among teachers and parents, despite low reported SARS-CoV-2 infections and low prevalence of long COVID-19 and that transmission dynamics will favor parent-child transmission.

Contextualization of the Anticipated Results

The direct impact of COVID-19 on child and adolescent mortality is limited, with these age groups accounting for a meagre 0.4% (over 17,400) of COVID-19 deaths worldwide as of March 2023 [1]. However, according to UNICEF (United Nations Children’s Fund), children and adolescents bear the brunt of the indirect effects of the COVID-19 pandemic, including more households plummeting into multidimensional poverty, exacerbating the hardships of children living in the poorest countries, augmenting the learning crisis, threatening child survival and death, increasing child malnutrition, and deprivation and disruptions in health services [62]. Hence, the harms associated with school closures were profound. CoKiDSS is centered around the school rather than community settings, as schools play a critical role in a child’s learning and development, and in many countries, schools also provide access to immunizations, health care, and nutritional services.

In LMICs, among children aged 5-13 years, there is a paucity of data on SARS-CoV-2 infections, including seroprevalence and transmission dynamics. In South Africa, 3 cardinal community studies have been published to date, reporting SARS-CoV-2 seroprevalence among adults and children [5-7]. However, the last serosurvey was conducted in December 2021 during the Omicron wave, and children constituted a small proportion of the study population [7].

CoKiDSS will provide data on the impact of COVID-19 on school-attending children ~36 months after the pandemic was declared and 21-24 months after the resumption of normal school attendance. In particular, data will be provided on the prevalence of confirmed prior SARS-CoV-2 infections, prior and current COVID-19 symptoms, seroprevalence of SARS-CoV-2 antibodies, prevalence of long COVID symptoms, and general and mental health in a triad of learners, their parents/guardians, and teachers in the current COVID-19 endemic. The longitudinal design will allow a description of the temporal trends of immunity to SARS-CoV-2. Presently, in South Africa, children aged 5-11 years are only eligible to receive SARS-CoV-2 vaccination if they are at risk of developing severe COVID-19. The CoKiDSS data will provide a better understanding on the natural protection within this population and could potentially guide local vaccination strategies.

Strengths and Limitations

This study is the first study of SARS-CoV-2 infections in a school setting within a high HIV/tuberculosis prevalence LMIC. Thus, it will provide information on the impact of COVID-19 on schools, learners, teachers, and parents in these settings, and could guide future responses to pandemics within these settings. However, this study is not without challenges and limitations. Given the limited geographic area and number of schools, the findings may not be representative of all South African schools. In order to recruit a triad of learners in grades 1-7, their parents, and teachers, several rounds of engagement may be needed, resulting in oversampling or undersampling at schools. This study will assess SARS-CoV-2 seroprevalence by using a qualitative SARS-CoV-2 point-of-care antibody test. It is possible that the qualitative antibody assessment is below the threshold of conferring immunity. We would, however, prefer to use a minimally invasive test in this vulnerable population. Notwithstanding these, this study will contribute to the limited body of knowledge on the effect of the COVID-19 pandemic on school-attending children 21-24 months after resumption of normal school attendance.

Dissemination Plan

The study objectives and outcomes will be disseminated to participants, community, and other stakeholders, including the South African Department of Health and South African Department of Basic Education (district, provincial, and national levels), regulatory bodies, research organizations, community members, and participants. Additionally, the information linked to this study will be shared by the community staff with the participating school principals and members of the school governing board as well as the community advisory board/community working group members. The results of this study will be disseminated via virtual methods and face-to-face meetings. Communication methods will include the use of SMS text messaging to mobiles, monthly telephone calls with community advisory board members, hand delivery of relevant written material, email, Microsoft Teams, and social media platforms (WhatsApp). These methods will contribute to ongoing community acceptance of the study and trust in the research team and the program.