Alternative COVID-19 mitigation measures in school classrooms: analysis using an agent-based model of SARS-CoV-2 transmission

Abstract

The SARS-CoV-2 epidemic has impacted children's education, with schools required to implement infection control measures that have led to periods of absence and classroom closures. We developed an agent-based epidemiological model of SARS-CoV-2 transmission in a school classroom that allows us to quantify projected infection patterns within primary school classrooms, and related uncertainties. Our approach is based on a contact model constructed using random networks, informed by structured expert judgement. The effectiveness of mitigation strategies in suppressing infection outbreaks and limiting pupil absence are considered. COVID-19 infections in primary schools in England in autumn 2020 were re-examined and the model was then used to estimate infection levels in autumn 2021, as the Delta variant was emerging and it was thought likely that school transmission would play a major role in an incipient new wave of the epidemic. Our results were in good agreement with available data. These findings indicate that testing-based surveillance is more effective than bubble quarantine, both for reducing transmission and avoiding pupil absence, even accounting for insensitivity of self-administered tests. Bubble quarantine entails large numbers of absences, with only modest impact on classroom infections. However, maintaining reduced contact rates within the classroom can have a major benefit for managing COVID-19 in school settings.

Modelling the impact of school reopening and contact tracing strategies on Covid-19 dynamics in different epidemiologic settings in Brazil

“…Finally, large-scale testing of suspected individuals and contact tracing strategies are important to minimize the impact of school reopening, and more effective when combined with other mitigation strategies, similar to the predictions found for other countries [ [32] , [33] , [34] , [35] ]. However, our simulations indicate that they require a significant number of tests to do so, especially if there are already high levels of community transmission.…”

Investigation of airborne spread of COVID-19 using a hybrid agent-based model: a case study of the UK

“…The spatial and temporal scales of these models varied considerably. These models studied the infection spread over a variety of geographical locations ranging from a large building (school or hospital) to a country and the temporal scales of the modelling outcomes range from a couple of weeks to several months [34][35][36][37][38][39][40][41][42][43][44][45][46]. Agent-based modelling tools such as Covasim [29] and PanSim [33] have been used to model infection spread in cities and countries over a period of a year.…”

“…A number of epidemiological models of COVID-19 infection have been developed using agent-based modelling approach to investigate various non-pharmaceutical intervention strategies such as social distancing, usage of masks, contact tracing and subsequently these models also explored the spatial aspects of infection spread [26][27][28][29][30][31][32][33]. Particularly, it is used to investigate various levels of interaction details of the infection propagation in healthcare facilities [34,35], educational institutions [36][37][38][39], cities [40][41][42], states and countries [43][44][45][46].…”

Response to the COVID-19 Pandemic in Classrooms at the University of the Basque Country through a User-Informed Natural Ventilation Demonstrator

“…The COVID-19 pandemic has provoked a renewed interest in methods to predict the risk of infection spread [ 53 , 54 , 55 , 56 , 57 , 58 ], both probabilistic and deterministic, applied to a variety of settings (indoor, outdoor, public transportation [ 59 ], etc. ), as well as tools [ 60 , 61 , 62 ] for direct application.…”