COVID-19 has spread too quickly for local public health departments to keep up with the intricacies of contact tracing. Some experts suggest automation, not phone apps, is the key to moving forward.
Contact tracing dates as far back as the 16th century, when physician Andrea Gratiolo used the method to show that the bubonic plague didn’t originate in a particular individual. In 1854, physician John Snow broke new ground in contact tracing, using “maps and records to track the spread of a disease back to its source,” and gave birth to modern epidemiology.
So if contact tracing has proven effective in detecting and slowing down the spread of disease for more than a century, why hasn’t it led to more success in the U.S. effort to control the COVID-19 outbreak?
Steve Bennett, former director of national biosurveillance at the U.S. Department of Homeland Security, said the type of technology used for contact tracing can make a difference.
Phone apps, for example, may not be the best choice. For one thing, research from software company Myplanet indicates that certain populations, such as older individuals, are less likely to trust such applications.
Perhaps more importantly, Bennett, who now directs global government practice at software company SAS, said the alert-based system of the apps doesn’t clearly identify what kind of contact individuals have with one another. Traditional contact tracing is based on interviews with individuals, who describe their activities in detail.
“You’re going to get a whole bunch of noise [with phone apps],” Bennett said. “It’s not going to be a real good decision to isolate people based on that.”
Sherrine Eid, global head of real-world experience and epidemiology for SAS, said she’s glad many companies have invested in contact-tracing tech, because better solutions can be found when many great minds work together. However, she said the phone apps may not provide the deepest level of information that would make contact tracing effective.
Eid also pointed out that privacy policies and regulations can be a hurdle for the apps.
“We usually develop [technology] quicker than our regulations can keep up,” Eid said. “That’s just human nature, I think.”
Nothing beats the tried-and-true method of contact tracing, Bennett said, but there’s a problem: the speed and scope of COVID-19 can turn the reliable process of contact tracing into a workforce numbers game that public health organizations may not be able to win.
Specifically, there are so many people getting sick at one time that it can make the classic interviewing method too slow. Bennett said agencies have tried to combat this issue by throwing more people into the mix. Research from the Association of State and Territorial Health Officials (ASTHO) has even estimated the number of staff that each state would need for successful contact tracing.
Unfortunately, gathering interview data is only the first step of contact tracing, and all the interviewing can result in data overload.
“Throwing bodies at the contact tracing problem can generate interview data, but you’re overwhelming public health epidemiologists who need to do the investigations,” Bennett said.
J.T. Lane, chief population health and innovation officer at ASTHO, said public trust in America plays a large role in the relative success of contact tracing. Depending on the state and county, citizens may complete interviews at a lower rate and may even be less likely to answer contact tracers' phone calls.
“To me, the biggest challenge is not the volume of data," Lane, a former state health official, said. "The biggest challenge is encouraging people and trying to get people to trust that their health department and the contact tracers are trying to protect the community."
From an investigative standpoint, epidemiologists need to be able to focus on augmenting interview data with other data, Bennett said. For instance, if an infected person got on a flight, one could then utilize flight records to further trace the outbreak. After one enriches the data in this way, the next step is deriving connections between all of the information at play and then identifying risk patterns.
But again, with COVID-19, scaling can be an issue. Bennett said this is where technology can help. Rather than have epidemiologists do investigative steps by hand, certain processes can be automated with software tools.
To Bennett’s point, the Centers for Disease Control and Prevention released a short guide on how digital tools can assist with various aspects of contact tracing.
Lane said systems that can keep up with taskloads and create reminders about following up with contacts can indeed be beneficial. However, such capabilities are not universal for health departments.
“Public health infrastructure has been chronically underfunded for decades," Lane said.
Another issue remains: the quality and consistency of the data. Both Bennett and Eid spoke about the lack of a national standard when it comes to contact tracing data.
“You could be in the same state and have 30 counties with 30 different ways of curating the data,” Eid said.
As a final concern, Bennett said the best contact tracing in the world won’t help much if you don’t have good testing to confirm who is sick. Testing becomes even more important when one considers how many COVID-19 carriers are asymptomatic.
Bennett believes the United States should have gotten more testing lined up when the country first locked down.
“The ability to test effectively has been our biggest failing,” Bennett said. “I think the United States will have a lot of soul searching to do for a long time. It’s a big policy question for whoever wins in November.”
Looking for the latest gov tech news as it happens? Subscribe to GT newsletters.