Participation to the study
Children’s participation in the study was lower than expected and may be partly explained by the problem of the recruitment period, which occurred during the lockdown. Reaching parents and children was difficult, resulting to misunderstandings and fears about the study. The two excluded schools because of a low participation rate had also a low socio-economic index. Some parents did not speak French, and even if we translated flyers into different languages, they refused to let their child participate. The consent form was the only way to inform parents in details about the study, but it was written in complicated terms as requested by our ethical committee, and led to some fears. For example, in the consent form, one of the possible side effects described for the finger prick test was a bruise, which was mentioned as an ecchymosis.
Most parents expressed their refusal to let their child participate, using arguments that showed a lack of understanding of how the study was being conducted (e.g. “we do not want our child to be vaccinated”). Investigators tried to explain the process of the study in simple terms, suitable for the general population, to reluctant parents. Some parents subsequently agreed to their child’s participation, others did not.
Participation of children was higher in the schools where the participation staff was high. The only person available to explain the study simply and clearly was the teacher or the school principal. When the teacher was motivated to participate, children were perhaps more motivated to participate. School staff were mostly motivated to participate in every school. Not everyone who worked in the school could be present on the day we conducted the test, which may explain some refusals. Adherence to the study for the purpose of knowing the presence or absence of antibodies to the SARS-CoV-2 virus was one of the reasons that motivated staff members to participate. Contrarily, for those who had been infected by COVID-19, some did not wish to participate because they already knew the presence or absence of antibodies. Some school staff stated that they were interested in knowing the presence/absence of SARS-CoV-2 antibodies.
Motivation of school was not the only criterion for children to participate to the study.
In the protocol, we decided to establish 3 inclusion criteria to select schools. One was the incidence level of SARS-CoV-2 based on May 6th 2020. When we computed the participation according to the incidence level of SARS-CoV-2, we estimated the incidence based on the inclusion day. Participation of children and school staff was higher where the incidence level of SARS-CoV-2 level of SARS-CoV-2 was lower. The fear of the virus can explain the lower participation where the incidence level of SARS-CoV-2 was higher.
Participation of children and school staff was lower in large schools. Small schools are known to have more social contact and therefore, as children and school staff participated, others were more inclined to participate.
Children’s participation was the lowest (< 50%) in schools with low socio-economic index (ISE ≤ 7), but it wasn’t true for school staff. Socio-economic environment seems to have an influence on children’s participation. Although we needed written consent from children and parents, the information should be more accessible and understandable, with simple terms suitable for children and parents. Every school staff didn’t come from the same communal area. The Socio-economic index was calculated for every child within the school and children usually come from the same communal area as the school. As a result, the socio-economic index and the participation of school staff were not correlated.
A low socio-economic index shows that the different variables that decrease this index are more important in these schools. The different variables that decrease the socio-economic index of the schools are: proportion of people with a nursery or primary school degree, proportion of persons with social assistance, proportion of people with a manual work and proportion of people working in the lowest level of tertiary sector.(16)
One of the variable concerns the obtention of social assistance, in order to determine the socio-economic index. In Belgium, the CPAS (“Centre Public d’Action Sociale”) is a social assistance that aims to ensure the right to social integration for people who do not have sufficient income and who meet the various legal conditions. In order to benefit from this social assistance, the person must meet the six legal conditions, which are: Belgian nationality, permanently living in Belgium, being of legal age, insufficient resources, willingness to work (unless health reasons prevent this) and having used up all the different social rights (the CPAS only operates as a last resort).(19) The ISE represents mainly the children and not the staff in a school.
Seroprevalence in children and in school staff
The seroprevalence of SARS-CoV-2 antibodies in school staff was higher than the seroprevalence in children, with 25.4% (95% CI [20.0;31.4]) and 20.7% (95% CI [18.1;23.5]) respectively. Seroprevalence among children was higher in schools with high seroprevalence among school staff. The weak intraclass coefficient correlation (ICC = 0.08 in children; ICC = 0.12 in children with school staff) suggests that seroprevalence didn’t present classroom clusters. The classroom size did not appear to have an impact on the seroprevalence of SARS-CoV-2 among children in this study.
An initial study in Iceland showed that 6.7% of children under 10 years of age (38/564) were tested positive for SARS-CoV-2 by nasopharyngeal and oropharyngeal tests, compared to 13.7% (1,183/8,635) of individuals older than 10 years of age.(3)
In England, a study in primary schools reported a seroprevalence of 11.2% (91/816; 95% CI [7.9;15.1]) in children and 15.1% (209/1,381; 95% CI [11.9;18.9]) in staff. The weekly transmission of SARS-CoV-2 in children was 4.1 per 100,000 persons (12.5 in adults).(20) A prospective cross-sectional study analyzed infection clusters and outbreaks in staff and students. Staff had a higher incidence over children and the risk of an outbreak increased by 72% for every 5 cases per 100,000 in community incidence.(20) Children appear to be more likely mild or asymptomatic, and then are less being tested for the virus. (20, 21)
Another small study conducted in a Belgian primary school, 20.6% (13/63; 95% CI [10.6;30.6]) of children tested positive with throat washing sample, and the majority of them were asymptomatic. In that study, seropositivity in adults was 27.1% (32/118; 95% CI [19.1;35.7]), also higher, as in our own results.(22)
The course of COVID-19 was often milder with a better prognosis in children than in adults, and deaths were exceptionally rare in children.(23–25) The difference between adults and children to develop COVID-19 remains unexplained, even if immunity and innate responses may play a role.(26–28) Children can develop an immune response to the virus without virologic confirmation of SARS-CoV-2 infection, suggesting the possibility that immunity prevents the onset of SARS-CoV-2 infection.(29)
Children less than 10 years old were estimated to be significantly less infected by SARS-CoV-2 compared to teenagers or adults.(30) It seems important to make a distinction between children and adolescents, as the results do not appear to be entirely similar. Several studies have shown that seroprevalence in children and adolescents would differ.
Two Belgian studies compared seroprevalence in Belgian primary schools to secondary schools. Both studies reported similar antibody seroprevalence in children, but it was lower than seroprevalence in adolescents. In the first study, 6.6% (95% CI [1.2;12.1]) were seropositive in primary schools compared to 12.2% (95% CI [7.2;17.1]) in secondary schools.(31) The second study performed this comparison over 3 periods: from December 2020 to January 2021, in March 2021, and from May to June 2021. Seroprevalence was comparable in both groups but lower in primary schools. Seroprevalence was: 11.0% (95% CI [7.6;15.9]) in primary schools versus 13.6 (95% CI [9.9;18.5]) in secondary schools, 17.1% (95% CI [13.3;21. 9]) in primary schools versus 18.0% (95% CI [13.6;23.8]) in secondary schools for the second period; 15.4% (95% CI [12.2;19.6]) in primary schools versus 17.2% (95% CI [13.1;22.7]) in secondary schools for the third period.(32)
Seroprevalence and communal cumulative incidence
Throughout the lockdown period, investigators worked closely with the local health promotion teams and schools. Before, during and after the study, the local physicians provided data concerning infected children and school staff. Specifics measures against COVID-19 were set up for primary schools (see Additional file 1 for more details).(33, 34)
Previous studies reported that children do not appear to play a major role in the transmission of SARS-CoV-2.(3–9) Results of our study show that the seroprevalence among children and school staff within schools was higher than the incidence of COVID-19 positive cases in the communal area, because the communal cumulative incidence was underestimated. At the beginning of the pandemic in Belgium, only people with severe symptoms (e.g. temperature ≥ 38°C) were tested. All possible contacts in adults were tested after mid-June 2020.(33)
The lateral flow test that we used during our study had a sensitivity of 68.8% and a specificity of 95.8%.(17) The positive predictive value is expected to be high, but we have no way to estimate the seroprevalence itself.
The lateral flow test in children and school staff was operated at inclusion but a second test one week after was not allowed per protocol, so positive IgM with negative IgG could not be assessed for new infections.
Presence of antibodies, reflects a previous infection/contact with the virus. It doesn’t show the actual number of persons with SARS-CoV-2 infection at that time. This might also explain why our seroprevalence of antibodies in children wasn’t correlated to the communal cumulative incidence.
Seroprevalence and size of the school
Size of the school was considered as a way to evaluate social contact in children, which is an important factor in the virus transmission.(6) The findings suggest that the seroprevalence of SARS-CoV-2 antibodies was lower in larger schools, both in children and in school staff. These results suggest that close contacts are maybe more important in small schools, even if number of contacts is expected to be higher in larger schools.
Seroprevalence and socio-economic index
Seroprevalence in school staff was higher in schools with a high socio-economic index, and seroprevalence wasn’t correlated with ISE in children. A first reason is that the socio-economic index is calculated from children data and not from staff data. Another reason is that most teachers and school staff live outside the communal area. For children, seroprevalence didn’t correlate with socio-economic index as opposed to participation rates that were highly correlated with ISE.
Our study has some limitations. Information about the study was only provided in paper forms with complicated terms and maybe not adapted to the general population, leading to low participation. Recruitment was made thanks to school principals because of the restriction distance measures at that time. Distance recruitment increased misunderstanding of parents and children about the purpose of the study.
Seroprevalence was not analyzed two times but only at the beginning of the study, based on the protocol. A comparison between the beginning and another time point during the study would allow to analyze the evolution of the situation during the study, but also during the different measures taken at that time against COVID-19.
Another limitation is that the ISE is linked to a school and not a child; we don’t know if all participant children from low-ISE schools are living in poverty or not.
Nevertheless, our results are consistent with other studies. Our main findings are that children from schools with a low socio-economic index were less allowed to participate. Children had a lower seroprevalence than school staff and there were no classroom clusters, suggesting that they are not the transmitters of SARS-CoV-2. Our results strongly suggest that information strategies have to be more oriented to low socio-economic index in order to increase their participation in the disease control. Well-being and mental health of children should be at the center of our decisions in order to limit further consequences in their development and scholarship. Data on the virus infection and transmission are important for decision-making to control the disease, while considering the numerous consequences that may result from these decisions.
Children have been and are still being strongly affected by the COVID-19 pandemic. Understanding the transmission of SARS-CoV-2 in schools could limit the long-term consequences on children’s scholarship, development and well-being. Taking socio-economic index into consideration is an important notion to analyze the virus transmission as the pandemic has reinforced social inequalities. By integrating all the lessons learned from this pandemic, we will be better prepared to deal with future health problems.