How to Use Data to Supplement Manual Contact Tracing
Ever since COVID-19 first reached the U.S., the size of infected populations has fluctuated. One day, a state might see hundreds of diagnoses, and the next only a few dozen. To counter that variability, colleges and universities need contingency plans for potential outbreaks. One of the keys to containing an outbreak is contact tracing. But the manual nature of contact tracing can be a hindrance.
“They have no insight into how the patient is really doing,” says Matt Whitehill, a doctoral student at the University of Washington who is a part of a team developing an app to help monitor the coughs of quarantined COVID-19 patients. “They have to contact them and ask, ‘Do you have a fever? Do you have a cough?’ And these are all subjective measures.”
Over the past 10 years, the UW researchers have compiled an extensive database and algorithm that differentiates between roughly 20,000 cough sounds and other environmental noises. Since coughs are a common symptom of COVID-19, simply tracking the number of times patients cough per day can provide valuable insights into their recovery. “If the coughs per day over a two-week window are going down, they’re probably getting better,” Whitehill says. “It’s a good way for public health officials to triage or emphasize who may need medical attention.”
In a similar vein, University of California Irvine researchers developed an open-source codebase for a QR code–based app called TrackCOVID, which alerts people if they’ve interacted with infected people. The QR codes will be posted in high-traffic areas, such as university campuses, grocery stores and public transportation hubs.
“If you went to the same place as someone who was exposed within a certain time frame, you will be notified that you may have been exposed,” says Tyler Yasaka, co-developer of the app. “It actually doesn’t even tell you where you might have been exposed. It definitely doesn’t tell you who might’ve exposed you.