January 12, 2021
A trip to the grocery store or a friend’s house may seem harmless enough, but the spread of COVID-19 is closely tied to people’s mobility. To help determine the level of travel that is safe for communities, as well as when an outbreak is likely to occur, researchers have developed a new type of mobility informed infectious disease model. It is already being used by the Centers for Disease Control and Prevention (CDC) to forecast COVID-19 deaths across the country.
“Overall mobility is measured by the number of places an average person visits in a day,” says professor Jeff Ban. “If a critical value is exceeded, our model will indicate an outbreak of more COVID-19 cases. Of course, public health measures such as social distancing and masks will change the critical value and other parameters of the model.”
Ban’s model — which he developed in collaboration with Yunfeng Shi, a materials science and engineering researcher at Rensselaer Polytechnic Institute — was inspired by the idea that simple chemical reactions, such as how molecules collide to form reactions, could be applied to forecasting COVID-19 transmission. Similar to how the likelihood of molecules colliding increases according to distance traveled, the researchers speculated that the risk of encountering someone infected with COVID-19 also increases the more a person travels.
Therefore, to control COVID-19 outbreaks, the average mobility over a period of time must be lower than a critical value, which varies by city and considers public health measures such as social distancing and city-wide mask usage. For example, the critical value was 30% of the pre-COVID average mobility for New York City and 60% of the pre-COVID average mobility for all other counties in New York in mid-March.
Although this is not the first COVID-19 transmission model that incorporates mobility, the researchers say it is unique in its simplicity and accuracy. The model incorporates two data sources: COVID-19 fatality data from Johns Hopkins University and mobility data from Google.