COVID-19 Projections for K12 Schools in Fall 2021: Significant Transmission without Interventions

Introduction

During summer 2021, the Delta variant of SARS-CoV-2 has spread with a substantially higher infectivity than previous strains. Millions of students across the US are scheduled to soon return to K12 classrooms, and vaccination among them is limited (middle and high school) or unavailable (elementary school). Policymakers, school systems, and parents are evaluating the potential impact of school openings and interventions such as masks and regular testing in K12 schools on the spread of SARS-CoV-2. This analysis assesses the benefits of layered interventions in schools.

Methods

An extended Susceptible-Infected-Recovered model was developed using R programming language to project infections over one semester (107 days) within a well-mixed student population (n=500) in which 0.5% of incoming students are infected and one case enters the school per week (e.g., infected outside school). Weekly RT PCR testing of 50% of students with sensitivity (specificity) of 85% (95%) is assessed. Students who don’t receive testing are never put in isolation, regardless of disease status. Students who receive testing but are uninfected are also never put in isolation. However, those who are infected and receive testing have a high probability of being put in isolation with compliance (70%, 90%) contingent on symptoms (asymptomatic, symptomatic); these students remain in testing until they are fully recovered. Infected students who test positive and are isolated can no longer infect susceptible students. Alternatively, infected students who do not test positive have the capacity to continuously infect other students until they eventually recover. Testing is random among members of the student population that have not already recovered and are not in isolation. Case-investigation of potential contacts is not conducted. We modeled three different school settings across different levels of incoming protection (through natural or vaccine-acquired immunity) of 30%, 40%, or 50%; for convenience these settings are denoted as “low”, “mid”, or “high”, reflecting different communities or grade levels (i.e., “low” = elementary school, ‘mid’ or ‘high’ reflecting middle or high school). Levels of protection were based on CDC reports that 30% of students in the middle school age-range are vaccinated, 40% of students in the high-school age-range are vaccinated, and prior infection among all primary-school children is approximately 10%.¹ A baseline effective reproductive rate (R₀) of 4.0 represented the context of increased infectivity from the Delta variant. We assume universal mask usage decreases infectivity by 50%.² A scenario analysis considered an R₀ of 5.0. All model parameters are provided in Table 1.

Results

Without testing and masking (R₀=4.0), more than 75% of susceptible students get infected within three months in all settings (Figure 1). With masks (R₀=2.0), the proportion infected drops to approximately 50% for “low” incoming protection settings (“mid”=35%, “high”=24%). Testing half the masked population (“testing”) further drops infections to 22% (16%, 13%). If R₀ is 5.0, 95% (93%, 88%) can be infected without mitigation and 70% (57%, 41%) with masking for the “low” (“mid”, “high”) setting. Assuming a conservative total of 10 days of school absence per 5 new infections, there will be an estimated 210 (510, 400) absent days for the school without any interventions or 140 (120, 76) days with masking and testing.

• Download figure
• Open in new tab

Figure 1: Projected Spread of COVID-19 across Epidemiological Setting and School Policies:

Each of the following panels presents the spread within the three different epidemiological settings of our model (“Low” = incoming protection of 30%, “Mid” = incoming protection of 40%, “High” = incoming protection of 50%). The Y axis represents the fraction of susceptible students who become infected throughout the semester (107 days). The “No Masking” scenarios reflect disease spread with an effective reproductive rate of four and the “Universal Masking” scenarios reflect an effective reproductive rate of 2, representing a 50% reduction in transmission from masking. Across each setting and masking scenario, two different testing policies are modeled: either no testing (“baseline”) or randomly testing half of students every week (“PCR_50”).

• Download figure
• Open in new tab

Figure 2 Testing and Isolation Pathways within Extended SIR Model:

Within our model, various pathways are followed among students who received testing.

Discussion

Without interventions in place, the vast majority of susceptible students among K12 schools will become infected and school absences will increase, followed by additional cases in communities as infected students transmit to household members. Universal masking can reduce student infections by 26-78%, and testing biweekly along with masking reduces infections by another 50%. Self-quarantine among exposed students or virtual-by-choice options may further reduce infections.³ Interventions should remain until most students are fully vaccinated. Jurisdictions need time to reduce inequities in vaccine uptake, and they may periodically need interventions when community spread is high.

The link between Delta and long-Covid or multi-inflammatory syndrome is unknown, and the risk of the most severe disease is lower among children but may exist.^4,5 However, there are mental health concerns,⁶ learning gaps,⁷ and other implications if schools remain closed. The multitude of direct public health and indirect social and economic benefits make layered interventions worth the investment.