Child pages
  • Abstract: Spatially-explicit agent-based modeling of COVID-19 transmission in elementary schools, Y1
Skip to end of metadata
Go to start of metadata

This spatially-explicit ABM model uses real school plans and school activity schedules to simulate COVID-19 transmission in elementary schools. The model allows users to evaluate a range of non-pharmaceutical interventions (NPIs) which are being considered by school and school districts, including mask wearing for all or selected student cohorts, different class sizes, relocating lunch from cafeteria to classrooms, use of plexiglass desk dividers, shortening class time, various ventilation scenarios to reduce viral load especially in shared spaces, and staggered attendance. The agents (students, teachers, support personnel) can have different health status (healthy, exposed, infected -- asymptomatic, pre-symptomatic or symptomatic), which depends on the frequency and duration of close contact with infected agents, and on duration of viral load exposure in school spaces. The goals of the model is to tune non-pharmaceutical interventions to specific student cohorts, school resources and circumstances, to ensure that the limited resources are used optimally and that weak spots and higher-risk activities are identified early. The model has been initially implemented as a set of Python scripts and Jupyter notebooks, using ABM Mesa and Mesa-GEO libraries. A web-based version of the model has been run over 300 times by school principals and other school administrators in San Diego county. However, the online version of the model is limited to a range of pre-computed scenarios as it takes 20-30 mins to run a single model scenario on Comet. The goal of this allocation request is to optimize model computations, make sure that additional scenarios can be requested by users via a job submission portal, and manage computed scenarios for subsequent runs. An additional goal is to expand several model components, in particular: added infection risks from using school buses; more detailed HVAC modeling given additional information about air flows and the installed HVAC systems; school infection risks as dependent on community transmission in their service areas; and influence of testing and vaccination on day-to-date school management.
The model will contribute to scientific understanding and forecasting of public health risks and strategies for school reopening during the pandemic.

  • No labels