SDS Faculty Member Turns Thanksgiving Problem into COVID Solution

February 18, 2021

Since the COVID-19 pandemic hit the United States in March of 2020, Mike Porter, Associate Professor of Data Science and Systems and Information Engineering, has taken a deep dive into research surrounding the virus. 

Specifically, Porter created a model to assess the risk of gatherings and is also working with a team to track COVID-19 through wastewater surveillance. 

Risk of Gathering Model

When Thanksgiving was coming up, Porter’s family, like many, could not decide whether or not to gather. 

“We had multiple sides of the family and many groups felt uncomfortable with people coming in from different states and different locations and all the health concerns,” Porter said. “That got me thinking -- how do I quantify this? Is there a probability, a number that could help me understand and make decisions to go to something or not?”

From there, Porter started to put together models and equations to make estimates about the probability of someone having COVID in a group of  people from the reported case count data. 

“Taking this another step forward, I wanted to know ‘Can I estimate the likelihood that someone has COVID, but doesn't know it?’ Because this is the real concern.”

Porter wanted to create a tool that anyone could use to assess their probability of contracting COVID-19 from going to a gathering. After sending out information amongst the School of Data Science team, Porter was connected to online School of Data Science MSDS student, Darren Frye, who has a background in building apps and dashboards. 

“We teamed up and started to put some of these ideas together and our goal was to get an app together before New Year's,” Porter noted. 

Porter and Fry achieved their goal and created the app before New Year’s, which can be found here

When trying to make a decision about attending a gathering, individuals can plug in information about where they are from, where their relatives and friends are coming from, 

“The model looks back over the last several days and takes into account the number of positive cases in that area,” Porter explained. “It makes an estimate on the percentage of people in each region that have COVID and don't know it, combines this with the expected number of people at the event,  and produces a probability that there is at least one person at the event that is unknowingly infectious.”

For more information on the equations and models Porter used, see here

Wastewater Surveillance

Currently, Porter is teaming up with Heman Shakeri, Assistant Professor at the School of Data Science, and Brent French, Professor of Biomedical Engineering at UVA to track the spread of COVID-19 using wastewater surveillance. 

This project is funded by the Ivy Foundation COVID-19 Translational Research Fund. 

Porter broke down the research process. 

“UVA’s  facilities management team uses a device that goes down into the sewer and pumps up sewage wastewater into a container. That gets taken back to a lab that Brent French runs, they process the sample, and produce a score that quantifies the level of SARS-CoV-2 RNA in the sample,” Porter explained.

How does this data help? Porter noted that using the location of the manhole from which the wastewater was extracted, the team can figure out which areas in Charlottesville are contracting more COVID-19 than others. 

“If we get a strong signal in one of the manholes, we know that there's COVID in all of the upstream locations,” Porter explained. “If we're thinking about a community like in Charlottesville, our plan for this project is to be able to report on which parts of town are seeing the highest rates. For example,  ‘the west side of town, seems to be having an expansion in COVID, while in the south side of town, COVID is staying low.’”

A distinguishing factor of using wastewater surveillance is that it can detect the prevalence of COVID-19 before any symptoms may be evident. 

“People tend to shed COVID in their wastewater before they even have symptoms, and even if they never have symptoms, so it can provide somewhat of an early warning system of maybe up to up to two days or three days, which is significant when it comes to COVID.”

Once COVID-19 is detected in an area via the wastewater, Porter added that they can use this information to test other manholes within the area and get more specific locations, such as neighborhood or housing area, where the levels are high. 

“Part of our goal is to combine traditional epidemiological modeling with data from wastewater testing,” Porter said. “The wastewater can tell us about COVID trends earlier than the positive cases counts, so we are trying to combine those two data sources and models together in a new way that can provide more power and more predictability or forecast stability.”

This method of COVID-19 detection could be especially powerful in catching asymptomatic cases. This detection would prompt community testing of a particular area, which would help individuals realize when they have COVID-19, whether or not they are presenting with symptoms. 

Porter added that the eventual hope is to create an alert system, notifying a particular neighborhood or area when COVID-19 is detected in the wastewater. 

For more information, see here.