The Journal of School Nursing
2025, Vol. 41(2) 197–200
© The Author(s) 2023
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/10598405231194538
journals.sagepub.com/home/jsn
Abstract
During the 2020–2021 school year, many schools adopted remote learning or a part-time in-person learning (“hybrid”) approach to reduce the risk of in-school transmission of COVID-19. The purpose of this work is to describe case rates of COVID-19 in schools practicing different learning modalities on rates of COVID-19 to support risk-benefit decisions in the context of respiratory disease outbreaks. We conducted a person-time-at-risk analysis of rates of COVID-19, as well as testing and test positivity rates among Colorado students. Schools practicing remote learning had a lower adjusted rate of COVID-19 cases compared to either hybrid or in-person learning modalities. Students attending a school with remote learning had fewer reported tests, and test positivity was higher for remote learning. Our analysis found that both case rate and test positivity were similar in hybrid and in-person learning modalities, indicating that hybrid learning modalities may not reduce the risk of respiratory disease transmission.
Keywordscommunicable diseases, screening/risk identification, quantitative research, elementary, high school, middle/junior/high school
During the 2020–2021 school year, many schools adopted remote learning or a part-time in-person learning (“hybrid”) approach to reduce the risk of in-school transmission of COVID-19. Hybrid learning may also have helped avoid in-school exposure by increasing spacing between building occupants and decreasing the impact of quarantine on students and educators (Parks et al., 2021). Schools varied in learning modality employed and may have changed from one learning modality to another based on local epidemiological trends and their capacity to safely accommodate full-time in-person learning (UNESCO, 2020).
However, disruptions to full-time in-person learning may have negatively impacted students’ educational advancement and social and emotional well-being (Christakis et al., 2020; Verlenden et al., 2021). Further, the effectiveness of different learning modalities at reducing the transmission of infectious diseases is incompletely characterized, especially for comparisons among hybrid, full-time in-person, and full-time remote learning (Goldhaber et al., 2021). The effectiveness of different learning modalities at interrupting transmission may also have varied across grade levels and school sizes. Understanding the effectiveness of different learning modality choices on respiratory disease transmission for different age groups is important to allow schools and communities to make risk-benefit decisions in responding to future outbreaks of respiratory diseases. The purpose of this work is to characterize the differences in case rates of COVID-19 between in-person, hybrid, and remote learning modalities for different age groups while adjusting for local epidemiological conditions.
We used a person-time-at-risk analysis to identify the rates of cases of COVID-19 among students enrolled in Colorado public schools under in-person, hybrid, or remote learning modalities between August 1, 2020, and June 1, 2021. We identified cases among students by matching K-12 enrollment records for all public and charter schools in Colorado to cases of COVID-19 reported to the Colorado Department of Public Health and Environment (CDPHE). This case identification was conducted as part of regular infectious disease surveillance work. Only aggregate case numbers by school were used in the analysis, and individually identifiable data was not used.
The learning modality for each week was ascertained at the school level based on de-identified geolocated device data (“mobility data”). This data was collected from a commercially available database of cellular device data and devices were not individually identifiable. For each school, a baseline number of devices present for at least one hour during the school day (8 a.m. – 4 p.m.) was defined using 2019 mobility data for the month of November. This baseline was compared to the maximum number of devices during the school day for each week under investigation. Schools with greater than 45% of their baseline devices present were classified as “in person,” schools with 25%–45% were classified as “hybrid,” and schools with fewer than 25% of their baseline devices were classified as “remote.” Schools with fewer than five baseline devices were excluded. This method was validated using a limited dataset of published district-level learning modality data (see Supplemental Material: https://drive.google.com/file/d/1_pSYEc8duUhjH-QqVNZKx0sGI6aqDlA9/view?usp=share_link).
Students identified as cases were censored on their infection date and did not contribute person-time to the subsequent person-time-at-risk analysis following this date. The infection date was estimated as 10 days prior to the disease report date, and case rates were adjusted for 14-day county case rates.
Schools were stratified according to CDE-reported enrollment, and grade levels were identified based on the oldest grade of students served; a school serving K-8 would be classified as a “middle school” under this scheme. Community case rates for each school were identified based on CDPHE data of the 7-day average case rate on the Monday of each week. All calculated case rates in the main analysis were adjusted for community case rates, with uncertainty expressed as 95% confidence intervals (Gardner & Altman, 1986).
Polymerase chain reaction (PCR) test rate and positivity data for tests reported to CDPHE were also calculated for each learning modality. Records were aggregated and anonymized prior to analysis. Analysis was performed in R version 4.12 (R Core Team, 2021).
This analysis included 390,945 enrolled students (65.6% of all public and charter school students in Colorado) and 124,195,197 person-days, including 554 schools in 180 districts.
Schools practicing remote learning had a lower adjusted rate of COVID-19 cases (18.3 cases per 100 K person-days at risk, 18.0–18.7 95% CI) compared to either hybrid (21.1, 20.5–21.7 95% CI) or in-person learning modalities (21.0, 20.6–21.4 95% CI). This trend was seen in larger schools (500 or more students) and middle and high schools, but not in smaller schools (499 or fewer) and elementary schools (Table 1).
Students attending a school with remote learning had fewer overall reported PCR tests (183 tests per 100,000 person-days) than schools with hybrid (233 tests per 100,000 person-days) or in-person learning (250 tests per 100,000 person-days). Test positivity was higher for remote (10.8%) than for in-person (8.6%) or hybrid (8.2%) modalities (Table 2).
Remote learning was associated with a lower case rate of COVID-19 than either in-person or hybrid learning in middle and high schools. Hybrid and in-person learning modalities had similar rates of COVID-19.
The lower case rate among remote students might be due in part to reduced testing resulting in relative underascertainment of cases, as suggested by higher PCR test positivity (Furuse et al., 2021). Remote students may have had reduced access to school-based testing and fewer incentives to test for COVID-19. This would result in an artificially lower case rate for remote students, even if they were infected with COVID-19 at the same rate. Nevertheless, other analyses have found that remote learning reduces COVID-19 transmission in communities, so it is unlikely that true COVID incidence (absent any testing bias) was identical between students learning remotely and students attending school in person (Goldhaber et al., 2021).
The differences in findings by grade level may be due to differences in the implementation of learning modalities among different grade levels. Differential testing practices may also partially account for higher case rates detected among older students. Alternatively, implementation of different learning modalities may have varied based on school size; supplemental analysis showed here a strong correlation between school size and grade level, with larger schools serving older students (Kendall’s tau-b p < .001). Serological studies have not found younger children to have lower rates of SARS-CoV-2 infection (Hobbs et al., 2021).
Our analysis had a number of limitations. We could not account for the effects of other non-pharmaceutical interventions, such as distancing, quarantine, mask-wearing, or increased ventilation, which may differ between schools and grade levels. The rigor of symptom screening and testing practices was also not assessed. Our analysis did not include educators, who are important vectors for in-school transmission (Gold et al., 2021). In addition, the learning modality of schools was indirectly measured, and we excluded schools with limited location data (35% of schools). Elementary schools had fewer devices on average, and younger students may not have carried devices regularly. Finally, we only analyzed data for reported cases, and could not ascertain cases among students who were infected with SARS-CoV-2 but were not tested with PCR, or whose tests were not reported to public health. However, the study period occurred prior to the widespread availability of at-home testing, so we anticipate the vast majority of tests to have been PCR and to have been reported.
Hybrid learning may have had operational benefits for schools in reducing quarantine by facilitating greater spacing between students and educators, as well as by supporting full-time remote instruction for students and educators unable to attend any in-person schooling. However, our analysis found that both case rate and test positivity were similar in hybrid and in-person learning modalities, indicating that hybrid learning modalities may not reduce the risk of disease transmission. School administrators and health staff should consider strategies other than hybrid learning to reduce the risk of respiratory disease transmission in schools.
We would like to thank Ginger Stringer, PhD of the Colorado Department of Public Health and Environment.
Dr. Brian Erly conceptualized and designed the study, drafted the initial manuscript, and critically reviewed and revised the manuscript. Mr. Jackson collected data and critically reviewed and revised the manuscript. Ms. Pilonetti critically reviewed and revised the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
Brian Erly is currently affiliated with the California Department of Public Health, Richmond, CA, USA.
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Brian Erly https://orcid.org/0000-0002-1860-7092
Supplemental material for this article is available online.
Christakis, D. A., Van Cleve, W., & Zimmerman, F. J. (2020). Estimation of US children’s educational attainment and years of life lost associated with primary school closures during the coronavirus disease 2019 pandemic. JAMA Network Open, 3(11), https://doi.org/10.1001/jamanetworkopen.2020.28786
Furuse, Y., Ko, Y. K., Ninomiya, K., Suzuki, M., & Oshitani, H. (2021). Relationship of test positivity rates with COVID-19 epidemic dynamics. International Journal of Environmental Research and Public Health, 18(9), 4655. https://doi.org/10.3390/ijerph18094655
Gardner, M. J., & Altman, D. G. (1986). Confidence intervals rather than P values: Estimation rather than hypothesis testing. British Medical Journal (Clinical Research ed.), 292(6522), 746–750. https://doi.org/10.1136/bmj.292.6522.746
Gold, J. A., Gettings, J. R., Kimball, A., Franklin, R., Rivera, G., Morris, E., Scott, C., Marcet, P. L., Hast, M., Swanson, M., McCloud, J., Mehari, L., Thomas, E. S., Kirking, H. L., Tate, J. E., Memark, J., Drenzek, C., Vallabhaneni, S., Almendares, O., …, Weng, M. K. (2021). Clusters of SARS-COV-2 infection among elementary school educators and students in one school district—Georgia, December 2020–January 2021. MMWR. Morbidity and Mortality Weekly Report, 70(8), 289–292. https://doi.org/10.15585/mmwr.mm7008e4
Goldhaber, D., Imberman, S., Strunk, K., Hopkins, B., Brown, N., Harbatkin, E., & Kilbride, T. (2021). To what extent does in-person schooling contribute to the spread of Covid-19? Evidence from Michigan and Washington. https://doi.org/10.3386/w28455
Hobbs, C. V., Drobeniuc, J., Kittle, T., Williams, J., Byers, P., Satheshkumar, P. S., Inagaki, K., Stephenson, M., Kim, S. S., Patel, M. M., Flannery, B., Alston, B., Bolcen, S. J., Boulay, D., Browning, P., Cronin, L., David, E., Hayden, T., Li, H., … Zellner, B. (2021). Estimated sars-COV-2 seroprevalence among persons aged <18 years—Mississippi, May–September 2020. MMWR. Morbidity and Mortality Weekly Report, 70(9), 312–315. https://doi.org/10.15585/mmwr.mm7009a4
Parks, S. E., Zviedrite, N., Budzyn, S. E., Panaggio, M. J., Raible, E., Papazian, M., Magid, J., Ahmed, F., Uzicanin, A., & Barrios, L. C. (2021). COVID-19–related school closures and learning modality changes—United States, August 1–September 17, 2021. MMWR. Morbidity and Mortality Weekly Report, 70(39), 1374–1376. https://doi.org/10.15585/mmwr.mm7039e2
R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/
UNESCO, McKinsey & Company (2020). COVID-19 response – Hybrid learning: hybrid learning as a key element in ensuring continued learning. https://unesdoc.unesco.org/ark:/48223/pf0000373767
Verlenden, J. V., Pampati, S., Rasberry, C. N., Liddon, N., Hertz, M., Kilmer, G., Viox, M. H., Lee, S., Cramer, N. K., Barrios, L. C., & Ethier, K. A. (2021). Association of children’s mode of school instruction with child and parent experiences and wellbeing during the COVID-19 pandemic—COVID experiences survey, United States, October 8–November 13, 2020. MMWR. Morbidity and Mortality Weekly Report, 70(11), 369–376. https://doi.org/10.15585/mmwr.mm7011a1
Brian Erly is a physician and medical epidemiologist. At CDPHE he developed surveillance systems for COVID in hospitalized patients and wastewater systems, while drafting policies and guidance around school health, pediatric vaccinations, and MPOX prevention.
Parker Jackson is the Founder and Lead Engineer at Vanadata.io, a startup founded to support the COVID-19 response.
Therese Pilonetti is the educational setting subject matter expert for the Disease Control and Public Health Response Division within the Colorado Department of Public & Health Environment. She has over 26 years of experience working in public health in Colorado.
1 Colorado Department of Public Health and Environment, Denver, CO, USA
2 Vanadata.io, Boulder, CO, USA
Corresponding Author:Brian Erly, MD MPH, Colorado Department of Public Health and Environment, 4300 Cherry Creek S Dr, Denver, CO 80246, USA.Email: brian.erly@cdph.ca.gov
