Throughout US states and much of the rest of the world school closures were implemented in March and April of 2020 in efforts to stave off the COVID-19 pandemic. In-person schooling has been thought to present a significant risk of infection to students, teachers, and the broader community. Remote learning however is not without its costs. In-person learning helps support the emotional and social development of young children. Reports are now indicating that the youngest learners and their families are finding remote learning particularly difficult. Schools often provide more than just a place to learn and are frequently sources of support for families providing much needed nutrition.
We use Covasim model, a detailed agent based model of COVID-19 to evaluate the tradeoffs of different proposed strategies for reopening schools and identify the requirements to avoid significant community transmission. We explored 7 school reopening strategies including the use of face masks, physical distancing, cohorting students into distinct classrooms, screening, testing, and contact tracing, as well as changes to student schedules to reduce the number in school at the same time. Our work shows that phased school reopening starting with elementary schools, A/B day scheduling, and classroom cohorting presents the lowest risk strategy to include in-person learning.
This work also uses SynthPops, our open-source data-driven generative model of multilayer contact networks, to model detailed school network structures of contact between students, teachers, and additional school staff. We use census and municipal records to model realistic contact networks in different physical contact settings. For our latest schools work, we updated the model to generate more detailed classroom networks under different cohorting strategies.
Read the latest paper.