Mieka Smart, DrPH, is an assistant professor in the College of Human Medicine at Michigan State University, Flint campus.
Smart also holds appointments with the Division of Public Health and the Department of Epidemiology and Biostatistics. He joined a COVID-19 task force focused on developing plans to protect Flint area residents from contracting and spreading COVID-19. Part of this plan involves contact tracing.
“My involvement with contact tracing started back when we first started to understand how COVID-19, based on mathematical modeling, would transmit rapidly across the U.S.,” Smart said.
Here’s how contact tracing works in a nutshell:
Once a person tests positive for COVID-19, state public health officials assign contact tracers. These contact tracers then open an investigation to document symptoms and identify close contacts of the patient.
Smart said once there’s a COVID-19 positive test result, the state should call you within 24 hours.
“That’s when the patient reports out when symptoms began and who they came into contact with,” Smart said.
Based on CDC guidelines, contact tracers call “close contacts” who were within six feet of a COVID-19 positive patient for at least 15 minutes.
Smart said one of the biggest hurdles is that “close contacts” are not necessarily picking up their cell phones, especially considering most calls display as “unknown.”
“We’ve been able to do successful contact tracing with 58% of people the COVID-19 positive patients came in contact with,” Smart said, noting they would like to see better numbers.
Symptom trackers and contact tracing
Online symptom trackers rely on self-reported symptoms that are collected from patients and stored in a database for analysis. In theory, data-driven technology like this can help identity early onset symptoms before a person develops a full-blown illness, said Jonathon Feit, chief executive of Beyond Lucid Technologies, Inc., a software company based in Concord, Calif. His company developed an AI-based symptom tracker that is currently being used at the Los Angeles County Fire Department.
Feit said he worked with fire department staff, who identified patterns of symptomology. This data allows the symptom tracker to monitor for potential COVID-19 cases among first responders.
Essentially, he said the software uses an algorithm that indicates the likelihood of testing positive for COVID-19. This helps prioritize testing, Feit said.
“If you have a limited number of COVID-19 tests or delays in the supply chain, it helps you triage who to test first,” Feit said.
Before delaying or prioritizing testing, though, Feit said, “you first need some data to able to do this wisely.”
The software provides this ability.
The value of geo-location in contact tracing
Large data sets from AI-based technology like this can reveal previously unseen patterns. This allows local health officials to pinpoint suspected outbreaks by geographic location. By tracking data in this way, the application may help prevent hotspots when personnel start to experience symptoms, Feit said.
Feit gave an example:
“The algorithm flags you on Saturday working in West Hollywood, then in East Los Angeles and you reported a headache on Friday, upset stomach on Saturday, now it looks like there’s a problem brewing.”
This location data from dispatched calls can be added to the employee health records. With this data, the technology can help identify which communities may experience an uptick in positive COVID-19 cases, based on first responders reporting their symptoms, Feit added.
“When you see people at the station starting to exhibit symptoms, you know it’s a sign there is something going on in the area they are covering.”
What healthcare professionals have to say
We asked Smart to weigh in on the usefulness of such technology. She said one benefit is that it captures symptoms as they are happening in real time. Capturing real-time data reduces the need to solely rely on memory about specific symptoms and day of onset.
“If they have good data on symptomology, it is helpful in terms of knowing peak period of transmission so that contact tracing can focus on the time when symptoms began,” she said.
It’s also critical to know if first responders made contact with COVID-19 positive patients, she added.
Terrence McGregor, NRP, LP, NCEE, with LA County Fire Department agrees, and points out that it bears some challenges.
“The biggest hurdle is getting that healthcare information reported to us from other hospitals and skilled nursing facilities,” McGregor said. “They don’t always provide us with that information, or in a timely manner.”
To close this feedback loop, Feit said they implemented a “pre-hospital health information exchange” to encourage hospitals to provide patient data back to fire departments.
When hospitals share information, McGregor said employees receive notifications if possibly exposed to COVID-19 positive patients. Employees then complete an exposure report to determine risk.
“If they are classified as medium or high risk, we require them to report symptoms for 14 days even if they have none.”