Erin Brodwin writs in STAT that software engineer Tim Brookins didn’t build his Bison Tracker app with the coronavirus pandemic in mind. He was just looking for a fun way to chart the annual migration of fans from his home state of North Dakota to Texas for the national championship football game.
But in a surprising turn of events, the backbone of the Bison Tracker app has become one of the first digital contact tracing tools to be deployed on the state level — and Brookins is one of the lead engineers shepherding the effort.
Brookins, an engineer at Microsoft, fell into the volunteer role somewhat by chance. As the pandemic picked up earlier this year, he texted an old colleague — North Dakota Gov. Doug Burgum — to see how he could help and soon found himself in a meeting with Burgum and other state officials. Now, Brookins is developing two contact tracing apps as part of the state’s Covid-19 response.
ADVERTISEMENT
The approach will serve as a crucial test for whether digital tools will play any sort of permanent role in the pandemic response. While apps can’t replace health care workers, they may be used to bolster their efforts — so long as enough people use them. So far, though, that hasn’t gone as planned: Early excitement over cutting-edge technology has given way to a largely lackluster role for contact tracing apps. Many states have opted against adopting the tools as part of their arsenal. And the handful of states that are launching such apps — including North Dakota — haven’t seen them gain much traction.
“This is not some revolutionary thing that changes the game,” Brookins said. “Both are incremental improvements on what is and will continue to be a human contact tracing process, and they can both make it incrementally better.”
North Dakota could be an ideal place to test the waters, since it has both highly urbanized populations and rural areas where cellphones, smartphones, and internet access are less common.
“Most of the people in my area either don’t have interest in these apps or don’t have a cellphone or a smartphone,” said Jennifer Galbraith, a contact tracer with the North Dakota Department of Health who oversees a rural, southwestern part of the state. That doesn’t make the apps pointless, but it does limit how useful they can be, she said.
Right now, just 4% of North Dakotans are currently using the first app Brookins created, a tool called Care19 that keeps track of a Covid-19 patient’s geographic history.
“Do we wish it was 40%? Yes. Is it still useful? Hey, we tried,” Brookins said. He also knows it will be difficult to get people to start using the second app he and his colleagues are working on, which will inform people who may have been exposed to the virus.
“We know it’s hard enough to get people to download one app, and it will be harder to get them to download two,” Brookins said.
So far, the Care19 app has proved somewhat helpful to contact tracers like Galbraith, even with the limited uptake. During phone interviews with infected patients, for example, Galbraith will ask Covid-19 patients to open the app and walk her through their recent whereabouts, paying close attention to whether they’ve had close contact with strangers so that she can locate them and instruct them to isolate. It’s a simple way to jog people’s memories. When a patient says they don’t have the app, it doesn’t interfere with Galbraith’s process.
“Ultimately it would be easier for contact tracers if everyone used these apps,” Galbraith said. “But it’s more important that someone is conscious of where they’ve been and who they’ve been in contact with.”
The rollout of Care19 in North Dakota has also given Brookins and his colleagues a chance to work out the technical challenges that have hamstrung some digital contact tracing efforts, including early issues enabling Bluetooth-based apps on Apple and Google phones.
In the early days of the pandemic, one obstacle in particular made it difficult to get digital tools off the ground. Battery protections baked into the operating systems of both Apple and Google devices automatically shut down Bluetooth whenever a user’s screen went dark, making them a futile tool for tracking people’s proximity. Bluetooth would have been more ideal to use from a privacy standpoint, since it doesn’t require invasive location data but can instead estimate when two phones are nearby by having them exchange signals or “pings.” Location-based tools, on the other hand, track people’s whereabouts based on their GPS coordinates and are generally regarded as more invasive.
Without a way to solve the Bluetooth issue, Brookins went to work on a GPS-based app using code pulled from his Bison Tracker. That app, which became Care19, uses location data to create a list of all the places someone has visited in the past two weeks. Users can manually adjust their location in the case of inaccuracies, and the app automatically deletes the data every 14 days.
“We don’t always get it right, but it’s easy to check it every day or two,” Brookins said.
In May, Apple and Google released the technology that enabled apps to use Bluetooth without getting automatically turned off. Brookins is using Apple and Google’s backbone, called an exposure notification, to develop his second app. It will use the pings smartphones exchange when they’re nearby — in this case, within 6 feet for at least 15 minutes — to alert iPhone and Android users that they’ve been near an infected person.
The companies have left it up to states to coordinate their approaches in a way that ensures the apps work across state lines. To do so, each health department must make sure its app can share certain data — namely the random numbers used to log when someone has been exposed to an infected person.
Having state-level apps that can’t talk to each other presents a significant problem. Say an infected person from California traveled to Nevada, for example, and used only the coronavirus app developed by California’s public health department. Only California could trace that person, leaving Nevada health officials in the dark.
Outside experts say this issue can be easily solved by having health authorities coordinate their approaches and use the same central data repository.
“This is not a technically difficult problem,” said Edouard Bugnion, vice president and adjunct professor of information and data systems at research institute and university EPFL in Switzerland, adding that it would make sense for states to coordinate their approaches with the federal government.
South Dakota and another undisclosed state have already licensed the technology behind North Dakota’s apps. But even if every state deploys a Covid-19 tracking app, and all of those apps can talk to one another, digital tools still won’t be able to replace other critical elements of human contact tracing, such as the ability to collect detailed contact information and demographic data. Apps also can’t assess an infected patient’s medical history and or make sure that person has all the necessary supplies they need to safely quarantine, such as food and medications.
“You need to be able to say, ‘OK, Mrs. Smith, I see you have asthma. You’re going to have to be very careful and watch out for these symptoms,” said Claudia Williams, chief executive officer of California health information network Manifest MedEx and former senior adviser to the White House Office of Health Information Technology.
“That’s something an app just can’t do,” Williams added.