Bard professor studies whether social quarantine for COVID-19 is extreme
ANNANDALE-ON-HUDSON, N.Y. — The best methods for pinpointing COVID-19 transmission rates continue to elude public health and infectious disease experts in the U.S. and globally.
Now, it appears that mathematics may help connect the dots, particularly for small, rural areas such as Litchfield and Dutchess counties.
Three college professors, including Matthew Junge, assistant professor of mathematics at Bard College in Annandale-on-Hudson, N.Y., have been awarded a $60,000 Emergency Grant from the National Science Foundation (NSF) to develop forecasting models that better capture the geographic and social complexity of the COVID-19 pandemic.
Junge, lead investigator on the project, said his research team aims to develop network models and mathematical theory to test the robustness of some prominent models being used by governments to justify the extreme levels of intervention of the COVID-19 quarantine.
“Possibly we will provide some evidence that targeted intervention allows for many people to resume some semblance of normal day-to-day life,” said Junge, who will be working with Felicia Keesing, a biology professor at Bard and Nicole Eikmeier, a computer science professor with Grinnell College in Iowa.
While some of the research “will be purely theoretical,” said Junge, the goal is to provide insights to those who are developing strategies to mitigate the spread of the disease.
“Our models may help resolve questions about specific communities, like, ‘Can Bard College hold class in person this fall without risking another spike in infections?’ or ‘If Dutchess County reopens restaurants, what should the occupancy limits be?’”
The grant was awarded through the NSF’s Rapid Response Research program, which provides support for urgent scientific research that responds to emergencies and unexpected events. It includes funding for salaries, publishing costs and several undergraduate research assistants over a six-month period.
Junge explained that most existing models take a “zoomed-out” perspective — for instance, making statistical predictions using past infection and death counts. His team’s project, however, “zooms in,” he said, “and models individual connections in a community.
The zoomed-out models are better at answering questions like ‘How many Americans will die of COVID-19 by the end of summer?’” he said, whereas the research team’s models aim to more precisely pinpoint the pandemic’s geographic and social complexities.
One advantage of a network model, which tries to accurately describe the face-to-face interactions each individual in a society has and how an infection might spread, is that it is relatively easy to implement social distancing into the network.
“Mathematics are fairly adept at modeling the natural evolution of epidemics,” said Junge. “But most ‘off the shelf’ models were not built to describe the dramatic levels of intervention, such as business closures, travel limitations and social distancing, that we are living through during the COVID-19 pandemic,” said Junge.
“The grant brings together a biologist, computer scientist and mathematician as well as a few undergrad research assistants, to tackle this problem over the next six months. Felicia is an expert in infectious disease, Nicole in modeling real world networks, and I am experienced in network infection models.”
Once the research is completed, Junge said the hope is that the team’s research will offer “an alternate perspective from the zoomed-out models. This work tests their robustness, and could possibly help smaller communities — counties not countries — make policy decisions about managing disease spread.”