This article is part of Privacy in the Pandemic, a Future Tense series.
As Americans enter the second month of coronavirus isolation, experts are thinking through what systems we’ll need to have in place before loosening restrictions. The general consensus is we’ll need widespread testing in place, as well as methods to manage inevitable subsequent outbreaks. Some countries, like Singapore and Korea, have rolled out apps that track users’ locations and ping them if they’ve recently crossed paths with someone who has tested positive for COVID-19. Apple and Google recently announced plans for tools that could allow developers to build similar apps in the U.S. and elsewhere. This is the digital incarnation of contact tracing, a long-standing strategy in public health.
Long before COVID, contact tracers tracked the spread of other illnesses like gonorrhea, cryptosporidiosis, HIV, Zika, tuberculosis, and measles. So once COVID-19 was on the rise, local and regional public health officials enlisted nearby universities to help through old-school contact tracing. Shannan Rich, an epidemiology Ph.D. student at the University of Florida, began volunteering as a tracer after the Florida Department of Health reached out to her about a month ago. Her job: to reach people who have tested positive for COVID and any of their close contacts.
Sometimes the person she’s calling is aware they’ve tested positive, but other times she has to be the one to break the news. The next step is to ask questions to help them remember whom they’ve been around recently in hopes of identifying a potential source. “The goal is to figure out if we can link them to anyone else that’s a confirmed positive—maybe through an event they attended,” she says. That also helps identify anyone they may have exposed to the virus. While the scientific community is still investigating how long a COVID-positive person might be infectious, Florida’s contact tracing program currently defines that infectious period as two days before people first show symptoms up until the date of their interview with a contact tracer. Any “close contact”—anyone who has been within 6 feet of you for longer than 30 minutes during that determined infectious period—is notified that they’ve had interaction with a COVID-positive person. (In health care settings, where risk of transmission may be higher, the standard for “close contact” is sometimes adjusted downward to as little as 5 minutes of exposure.) The tracer will ask those exposed if they’ve experienced any symptoms and, if they’ve developed illness, encourage them to get tested. If they test positive, their contacts will be notified, hopefully breaking the chain of transmission. It’s not a perfect system, but according to statistics from the World Health Organization, contact tracing helped Chinese health officials identify hundreds of newly infected “close contacts” in a single day.
That crucial human element will be difficult to implement with any app, no matter how well designed.
This type of old-school detective work may sound straightforward: call people up and figure out whom else they may have seen. But with the science quickly evolving and standards shifting to accommodate new knowledge, contact tracers must navigate tricky situations, working to maintain individuals’ privacy while also making sure the right people are notified. As a general policy, tracers are not allowed to reveal the identities of people who are COVID-positive, even if people have guesses, and can only reveal the date they believe the exposure occurred. But Rich says notifying the people who live with a COVID-positive person has become easier with so many people self-isolating. Normally, she’d take down people’s info and call them individually, but since many people are at home, the infected person might instead waive their right to privacy and just put their housemates, spouses, and kids on speakerphone. But it gets more complicated when, say, a COVID-positive patient went to a health clinic and interacted with multiple nurses and patients during their visit, whose names they may not be able to recall. In that case, it’s the tracer’s job to figure out how to notify the right people without exposing the COVID-positive person’s identity.
The job also requires empathy, patience, and the ability to calmly explain next steps. Some of the people Rich has called up are panicked and anxious about being COVID-positive or having been exposed to someone who is. “It’s easy to interview them and ask them the main questions like a robot, but there’s an important educational piece,” she says. She explains to them what actions they should take, like quarantining, and can answer people’s questions about their personal circumstances, like how they can isolate themselves from the people they live with or how to access groceries. “There’s a lot of sensitivity involved in contact tracing, and the diversity of situations that people are in, how they respond to knowing about their status or exposure, is the human part of this work,” says Stacey King, a public health expert at the Harvard T.H. Chan School of Public Health, who is the co-lead of the Academic Public Health Volunteer Corps working with the state of Massachusetts to perform contact tracing. “Honestly, I think the human contact right now with a contact tracer is about much more than the health surveillance and the data. It’s also about helping them navigate all of it.”
That crucial human element will be difficult to implement with any app, no matter how well designed. For one, an app is no replacement for the educational aspect of contact tracers’ calls. An app might have an FAQ section, but it’s unlikely it could respond to the very specific questions or concerns people may have about their personal circumstances, and it definitely cannot provide reassurance in the same way a human voice might. It’s also less likely to have the flexibility of human experts thinking through difficult questions. “Every case is different. A lot of this cannot be just a decision a computer can make,” says Rich. “It’s all so new, and we’re making decisions by consulting with administrators at the state and national level. Like, we haven’t encountered this scenario—what should we do? There are lots of questions where we have to assess on a case-by-case basis.” For instance, Rich mentions they’ve considered scenarios in which a health professional may have been exposed. Their recommendations might differ depending on what kind of protective mask that professional was wearing at the time. (With widespread shortages, many health care workers are left with inadequate personal protective equipment.) Finding out that information requires talking with hospital infection control units, and a tracer might then consult with higher-ups, like epidemiologists from local health departments.
Rich also worries that contact tracing via app could result in an unnecessarily high number of “exposure” alerts. “If you were in a grocery store, and someone there was positive and they walked by you, would that indicate an exposure?” According to the current models for digital contact tracing, including existing apps like Singapore’s TraceTogether, yes, it would. Casting a wide net might encourage people to err on the side of caution and self-isolate, but may also result in a lot of alerts—and a drop in compliance as people become inured to them. Plus, “there are real psychological implications for informing someone that they’ve been exposed, and I think it’s important that people designing these apps think about that from the beginning,” says Rich. Some people are really freaked out, and getting an alert out of nowhere could be legitimately upsetting; Rich suggests that even just including a link to counseling resources in such alerts could be helpful.
Like coronavirus testing, the plans to roll out contact tracing have been driven primarily by states, counties, and academic institutions rather than the federal government. During a 2014 Ebola outbreak in Dallas, the Centers for Disease Control and Prevention worked with county and state officials to perform contact tracing. Currently, it appears the agency recognizes the need to expand contact tracing but has not yet announced any plans for a national effort. Meanwhile, contact tracers and staff organizing these programs have their hands full building capacity and running operations, but are still trying to make time to advise colleagues on starting their own programs. Harvard’s King tells me that some students at the University of Pennsylvania expressed interest in starting a contact tracing program, and she’s also heard from a graduate of her department who wanted to support a Native American community in California. “We’re organically hearing from different people. I don’t know how they hear about us,” she says. “I know other states have reached out to Massachusetts, and the state is planning to put together a toolkit and some other resources.” When I asked both King and Rich if they knew of national efforts to do contact tracing or records of which counties or states had programs, both said no.
Meanwhile, our neighbors in Canada launched a nationwide contact tracing program in early April and as of a week ago already had 27,000 volunteers. We’re nowhere close to that in the U.S.—Florida’s program currently includes about 100 tracers, Massachusetts has just hired about 1,000—and experts estimate we’ll need many, many more. Currently, epidemiology and public health experts estimate we’ll need between 100,000 and 300,000 contact tracers working to identify outbreak patterns if self-isolation policies are lifted. The city of Wuhan alone enlisted 9,000 tracers working in groups of five; each group was led by an epidemiologist.
Adding digital tracing apps to the mix will introduce even more variability to this already decentralized effort; it’s unclear which health departments might have access to this data and how it might fit in with existing tracking and contact tracing efforts. But if digital tools are deployed smartly, they could save investigators some precious time and effort. Rich mentions a new app called NextTrace, which asks COVID-positive individuals to fill out a form that asks many of the same questions a contact tracer might ask in an initial interview. “That could be powerful, potentially, in helping health departments focus on high-priority investigations,” she says, like spread in hospitals or nursing homes. In the push to develop apps, some companies have focused specifically on these “high-priority” locations—CarePredict, for instance, provides senior care facilities with bracelets that track people’s movement. If a resident falls ill, the facility receives a notification and can use the bracelet location data to determine which staff and residents should be tested and isolated, and which areas of the facility should get an extra-deep clean.
King agrees that digital tools could be helpful to inform public health officials’ response to COVID. “There’s no doubt that having that kind of data is the goal of contact tracing,” she says. But that data can only show us so much; people won’t necessarily download these tracking apps, nor will they necessarily bring their phones with them wherever they go. Stopping COVID will ultimately still require real people using their expertise, and King says the response from that community has been remarkable. After a call for volunteer contact tracers went out, “1,500 students responded within 72 hours,” says King. “I’ve never seen anything move so fast—to have so many people say, ‘Yes, we will help,’ and to have these massive institutions say, ‘Yes, we will deploy resources.’ ” Let’s hope that continues.
For more on coronavirus contact tracing, listen to Friday’s What Next: TBD.