Background
In March 2020, the COVID-19 pandemic caused widespread closures of schools throughout the U.S., impacting up to 55 million students in 124,000 schools [
1,
2]. Studies suggested that these closures may have negatively impacted student learning and the mental health of children and families, [
3‐
5] and due to the importance of in-person education for children, families, and communities, many schools reopened for face-to-face instruction in the 2020/2021 academic year [
6,
7]. However, due to their potential for high population density and frequency of close contact between students or staff, risks for introductions and transmission within the school setting persist while transmission of SARS-CoV-2 is ongoing in the community, particularly in populations where vaccination remains low [
7,
8]. Instances of widespread secondary transmission of SARS-CoV-2 have been documented in school settings when comprehensive prevention strategies are not implemented [
9]. To ensure that schools remain open, can continue to operate as safely as possible during the COVID-19 pandemic, and to inform implementation of prevention strategies, better understanding of the risk behaviors of staff and students relevant to in-school transmission is imperative.
Prevention strategies such as mask use and physical distancing have been identified as significant protective factors for reducing SARS-CoV-2 transmission in schools, while risk factors have included close-contact indoor sports, eating meals in close proximity, and attending social gatherings with persons outside the household [
10‐
14]. Similarly, the effectiveness of prevention strategies in the community are well documented, while activities such as dining in restaurants, attending church, and group fitness classes have been identified as risk factors for transmission [
15‐
25]. However, information on the prevalence of risk behaviors among school populations and how these behaviors impact risk of transmission in educational settings is sparse, and few studies have formally assessed the direct relationship between individual behaviors and SARS-CoV-2 transmission in schools. Understanding the extent to which students and school staff engage in risk behaviors may contribute to developing policies and practices that could help to reduce introductions and transmission of SARS-CoV-2 in school settings in order to limit school closures and maximize opportunities for in-person learning.
To describe the extent of transmission of SARS-CoV-2 in schools, the Centers for Disease Control and Prevention (CDC) partnered with the Georgia Department of Health and Cobb & Douglas Public Health to investigate transmission in a Georgia public school district. Results of this investigation showed that teacher-to-teacher and teacher-to-student transmission were a central element of transmission chains in this district [
26,
27]. To further inform school and community efforts to reduce SARS-CoV-2 transmission in schools, this secondary analysis utilized data collected during the school transmission investigation to describe the prevalence of risk behaviors relevant for SARS-CoV-2 transmission in this population of unvaccinated students and school staff and to identify associations between these behaviors and testing positive for SARS-CoV-2 in the era prior to widespread vaccination availability.
Discussion
This analysis builds on a previous investigation of in-school SARS-CoV-2 transmission to examine prevalence of risk behaviors among 717 students and 79 school staff originally identified as contacts of a COVID-19 case in a school setting. High rates of risk behaviors for school and community transmission were reported in this population despite high community incidence and lack of vaccine availability. A large proportion of participants reported unmasked time at school indoors, and most students who participated in sports did not wear masks during these activities. The majority of participants reported visiting indoor locations in the community, and large numbers reported social gatherings with people outside their household, including groups of ≥6 people.
This analysis also identified factors associated with SARS-CoV-2 positivity in this population of unvaccinated staff and students exposed to COVID-19 in school, including self-reported behaviors and school-reported characteristics of exposure. Nearly 10% of participants tested positive for SARS-CoV-2. Across student age groups, the strongest associations with SARS-CoV-2 positivity were with participation in school sports, especially unmasked time in sports, exposure to an index case during sports (middle/high only), and exposure to a staff index case (elementary only).
These results highlight behaviors that might increase or modify the risk of in-school transmission and introductions of SARS-CoV-2 into schools and help to demonstrate the connections between the school environment, in-school behaviors, and out-of-school behaviors. Since people who work in or attend schools are part of the community at-large, risk behaviors out of school could lead to increased opportunities for introductions in school and vice versa when community vaccination is low [
8,
32,
33]. For instance, the high proportion of respondents who reported dining indoors at restaurants or having large social gatherings has implications for introductions in the school setting, particularly due to the high level of community transmission in winter of 2020-2021 and unvaccinated status of staff and students [
27]. Correspondingly, although in-person education has not consistently been shown to increase community transmission, [
8,
34‐
37] this investigation and others have found substantial transmission from positive contacts to their families [
26,
27,
38,
39].
Few studies have examined behaviors in and out of school among a school population in the pre-vaccine era, which highlights the importance of these data in the literature. In October 2020, 65% of surveyed US middle and high school students reported consistent mask use in classrooms among their peers, and only 28% reported consistent mask use during sports or extracurricular activities [
40]. In four studies from May-November 2020, 74-89% of surveyed US adults reported wearing a mask in public, 80-89% tried to keep 6 feet apart, 66-82% avoided restaurants, and 38% avoided socializing with people outside their household, with lower rates in rural areas [
22‐
24,
41]. However, it is challenging to directly compare these studies to the current results due to varying study populations and the potential for changes in behavior patterns over the course of the pandemic.
One of the key results from this analysis was the strong association between participation in indoor sports and SARS-CoV-2 positivity. Indoor sports (basketball, wrestling) were the most frequently reported by participants, likely due to the winter sports season, and associations with SARS-CoV-2 positivity were very robust across age groups. Among middle/high school students, this result was consistent between self-reported sports behaviors and the exposure source identified by the school district. This was a surprising result for elementary school students, who do not have formal interscholastic sports leagues and were not generally identified as having predominantly sports-related exposures by the school district; however, the association with self-reported sports participation was very strong despite small sample sizes. This could be explained because, unlike in middle and high school, students in elementary school stay together in a cohort throughout the day including activities like recess, so the school district may have identified the classroom as the predominant exposure location due to the long duration of time in this setting. However, this designation does not preclude additional exposures during other activities. Thereby, although it is not known whether participants played sports directly with the index case, the strong associations between SARS-CoV-2 positivity and self-reported sports participation among this population of children with a known exposure suggests that playing sports unmasked conferred an additional risk to sitting in the classroom.
These findings of potential increased risk associated with sports are consistent with previous reports, which provide a growing body of evidence that there is limited ability to prevent transmission in unvaccinated individuals during high-intensity, close-contact indoor sports [
12,
27,
32,
42‐
44]. When community transmission is high, other athletic activities could be considered where comprehensive multilayered prevention strategies can be implemented, including correct and consistent mask use, vaccination of eligible staff and students, adequate physical distancing, avoidance of large crowds, and improved ventilation [
7]. Students can thereby continue to experience the physical and mental health benefits of school athletic activities while mitigating the risk to themselves and others [
45]. Mask use during sports did appear to be protective in this unvaccinated population, an important finding that could be incorporated into future prevention practices; however, results should be verified with larger studies.
These results also support previous findings that school staff are central to transmission in elementary schools [
8,
26,
27,
46,
47]. Close interactions between teachers and younger students are necessary for learning but provide more opportunities for transmission among unvaccinated staff or students, particularly if mask use is not consistent. Despite school policies requiring mask use indoors, our findings and others indicate that these policies may not be followed with 100% fidelity, particularly among younger students [
14,
40]. This issue is compounded by the finding that risk behaviors in the community were highest among staff. Interventions to reduce risks of staff-related transmission include vaccination of eligible staff and students, activities to reinforce appropriate mask use, reducing unnecessary in-person interaction among unvaccinated staff, and taking measures to reduce community exposures.
As a final note, the high prevalence of risk behaviors identified in this investigation underscore the importance of comprehensive school, state, and local policies to reduce transmission, in keeping with guidelines to prioritize schools remaining open for in-person instruction over nonessential activities [
7]. At the time of the investigation, vaccination was not available to staff or students, and Georgia COVID-19 regulations did not include any universal mask mandate or prohibit dining indoors in restaurants or the operation of indoor gyms and bars, [
48] despite demonstrated efficacy in reducing community transmission [
15‐
19]. The frequency of these behaviors in our population are therefore not unexpected given the proximity to holidays, cold winter weather, and ongoing effects of isolation, particularly among staff who may live alone or have responsibilities outside of the home [
49,
50]. Similarly, although CDC recommends limiting sports and extracurricular activities when community transmission is high, [
7] the Georgia High School Association did not impose restrictions on school sports or require mask use during sports at the time of the investigation [
51]. Without this guidance, it may have been challenging for local school boards to independently limit sports or require mask use during athletics. Implementing structural policies at the state or local level during periods of high transmission would likely improve adherence to behavioral recommendations, reduce community acquisition of SARS-CoV-2, and therefore reduce introductions of COVID-19 into schools [
16,
52]. Improving vaccination rates among eligible populations may also reduce introductions into schools and in-school transmission. However, until vaccination is available to persons of all ages, continued adherence to in-school prevention measures such as appropriate mask use will continue to be important to prevent in-school transmission.
This investigation had several limitations. Enrollment was limited to known contacts of a positive SARS-CoV-2 case and occurred in a single school district, which constrains generalizability to other populations. Similarly, the investigation occurred prior to widespread vaccine availability, and both behaviors and risk factors have likely shifted since this time. The sample size of SARS-CoV-2 positive participants was also small, which limited the ability to conduct statistical comparisons, and the use of self-reported survey data rather than direct observation of risk behaviors increases the chance of recall bias and social desirability bias. Due to the sampling design of recruiting only people exposed in school, it was not feasible to assess associations between SARS-CoV-2 positivity and out-of-school behaviors due to the risk of selection bias. Furthermore, for all self-reported in-school behaviors, it is not known whether the participant was exposed to the positive case during those activities, so this investigation could only identify associations but cannot determine causality. A sizeable percent of the population also could not be reached or declined participation, with the leading reason for refusals being belief that testing was not needed for asymptomatic contacts. This may indicate that contacts who refused were different than participants regarding behaviors for SARS-CoV-2 prevention. Finally, although SARS-CoV-2 positivity was associated with sports participation, comparisons between individual sports activities or settings could not be conducted due to low sample size and the limited number of sports in season during the investigation. Future studies could attempt to discern which sports activities are associated with the highest risk for transmission.
Despite these limitations, this investigation is one of the more comprehensive reports of school staff and student behaviors relevant to COVID-19 in the literature to date and identifies several characteristics and behaviors associated with probable SARS-CoV-2 transmission in school settings. These findings may be valuable to guide implementation of interventions in and out of schools to improve the safety of staff and student populations. Furthermore, COVID-19 vaccination is still not available for children under 5 years, and vaccination rates remain low in many US counties and internationally. Additional studies are needed to assess the impact of behaviors in school and the community on transmission in school populations, including the role of vaccination status.
Acknowledgements
We would like to thank the students, families, and school staff who participated in this investigation, principals and administrators at Marietta City Schools, the Georgia Public Health Laboratory, the CDC COVID-19 epidemiology task force, and the COVID-19 Georgia K–12 school team, including Abirami Balajee, CDC; Rebecca J. Chancey, CDC; Deanna Crosby, CDPH; Morgane Donadel, CDC; Cherie Drenzek, GDPH; Catherine Espinosa, CDC; Mary E. Evans, CDC; Katherine Fleming-Dutra, CDC; Catalina Forero, CDC; Kaitlin Forsberg , CDC; Jenna R. Gettings, CDC; Jeremy A. W. Gold, CDC; Esther Kukielka, CDC; Janet Memark, CDPH; Kiren Mitruka, CDC; Sam Moeller, GDPH; Jasmine Y. Nakayama, CDC; Yoshinori Nakazawa, CDC; Michelle O'Hegarty, CDC; Caroline Pratt, CDC; Marion E. Rice, Katelin Reishus, GDPH; Grand Rivera, MCS; CDC; Gurleen Roberts, CDPH; Roxana Rodriguez Stewart, CDC; Raquel Sabogal, CDC; Emanny Sanchez, CDC; Ebony S. Thomas, GDPH; Katerine Topf, GDPH; Snigdha Vallabhaneni, CDC; Andres Velasco-Villa, CDC; Mark K. Weng, CDC
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.