Study population and setting
This was an ecological study of 520 US counties between July 1 and September 14, 2021, comparing pediatric COVID-19 case rates between 1) counties with school masking requirements for all students and 2) counties without any masking requirements for students. Counties with heterogeneous school masking policies were excluded from the analysis. Only counties with at least 21 days of case data after the start of the school year were included. Pediatric (<18 years) case rates per 100,000 population were aggregated by week and numbered relative to the start of the school year in each county (i.e., weeks were numbered -3 to 2). Crude comparisons were made via t-test; additionally, the authors used multivariable linear regression adjusting for age, race, ethnicity, pediatric COVID-19 vaccination rate, community transmission rates, population density, social vulnerability score, proportion uninsured, proportion living in poverty, and community COVID-19 vulnerability score.
Summary of Main Findings
Of the 3,142 counties initially sampled, 520 were retained for analysis after the inclusion criteria were applied. The average change in weekly pediatric COVID-19 incidence from one week before to one week after school opening was 18.5 per 100,000 higher in counties without school mask requirements (34.9 per 100,000) than in counties with school mask requirements (16.3 per 100,000). After adjustment for possible confounding variables, school mask requirements were associated with a 1.31 per 100,000 lower daily COVID-19 pediatric case rate (95% CI: 1.11 to 1.51).
The authors were able to adjust for a wide array of ecological-level variables that may have confounded results.
The analysis did not adjust for other county-level differences in SARS-CoV-2 transmission mitigation measures in schools (e.g., ventilation, physical spacing of students, teacher and staff vaccination rates, testing and quarantine measures). These measures may have correlated with mask requirements at the county level; if these measures were partially responsible for lower COVID-19 case rates, the results here may overestimate the beneficial effects of mask requirements. Additionally, the inclusion criteria allowed for classification of a county if there was information on mask requirements in any single school within the district. This may have led to considerable misclassification if schools for which there was no mask requirement information had conflicting mask policies. Only 520 counties were included in the analysis, and these counties may not be representative of the wider U.S. population. No data were available on COVID-19 case rates by student age; there may have been variation in the associations between mask use and COVID-19 incidence by age groups. Finally, this was an ecologic study that did not examine individual-level mask use or COVID-19 outcomes.
This study adds more evidence derived from ecological, observational data to suggest that masks are effective in reducing SARS-CoV-2 transmission among children in schools.