Network-based modeling of COVID-19 in Washtenaw County, MI
Like many in the modeling community, I focused my efforts in 2020 on studying the COVID-19 pandemic in my local area. On this page is a short summary of these efforts. This work is in collaboration with my supervisor Denise Kirschner and a team of modelers at University of Michigan led by Marisa Eisenberg.
What is COVID-19?
COVID-19 is an infectious disease caused by the novel coronavirus SARS-CoV-2. It is thought to spread mainly from person to person through respiratory droplets. This disease has spread globally, infecting millions of people and causing a global pandemic. Most US states, including Michigan, have issued stay-at-home or shelter-in-place orders to slow the spread of COVID-19 and reduce the burden on our healthcare system.
For more information on COVID-19, see the CDC or WHO coronavirus pages.
What is our model?
We used a network-based modeling framework built on a synthetic population dataset for Washtenaw County, in which individuals interact through sharing households, workplaces, schools, and group quarters, as well as through casual contacts. Synthetic datasets developed by RTI International are available for all US counties and states through the University of Pittsburgh Public Health Dynamics Laboratory. Previously, we used this modeling framework to study the epidemiology of tuberculosis (see our recent paper on medRxiv). Recently, we have adapted the model to simulate the progression of COVID-19 using the model diagram used by the University of Michigan Epimath COVID-19 Modeling group and by adding essential businesses and hospitals to the model. Further, we have incorporated recent social distancing measures and executive orders. The model is calibrated based on counts of confirmed cases, hospitalizations, and deaths in Washtenaw County.
How was the model used?
Using our network-based model, we were able to explicitly simulate the effects of school and workplace closures, reductions in casual contacts, hospital-related transmission, and varying levels of infectivity and susceptibility. We used this model to evaluate re-opening strategies and predict how such strategies would affect the second wave of infections in Washtenaw County.
Our work was published in the Journal of Theoretical Biology, and can be found here.
Public data sources