Contact Tracing Apps Were Big Tech’s Best Idea for Fighting COVID-19. Why Haven’t They Helped?

When the Nevada Department of Health and Human Services launched COVID Trace, one of the nation’s first COVID-19 contact-tracing smartphone apps, on Aug. 24, state health authorities “strongly recommended” all 3 million-plus Nevadans download and use the app. But two and a half months later, adoption remained well short of that ambitious goal—the app has been downloaded just under 70,000 times as of Nov. 9, representing just under 3% of the state’s adult population. A total of zero exposures were registered in the app throughout the month of September, during which the state reported more than 10,000 new cases. One of the first positive test results logged into the app was submitted in early October by Nevada’s pandemic response director, who himself had contracted the virus. Nevada is currently reporting roughly 1,200 new COVID-19 cases daily, and the app doesn’t seem to be making a difference.

While researchers have worked for months to develop COVID-19 vaccines and treatments, contact-tracing apps like COVID Trace have been touted as one of the technology world’s most promising contributions to the fight against the pandemic. But seven months into the U.S. outbreak, such apps have made slow progress across the country, hampered by sluggish and uncoordinated development, distrust of technology companies, and inadequate advertising budgets and messaging campaigns.

“People are trying whatever they can think of, and this is one of those things,” says Jeffrey Kahn, a professor of health policy and management at Johns Hopkins University. “Whether it’s worth the investment, it’s really hard to answer that until there’s more information.”

Most U.S. contact-tracing apps are built by state governments, but underpinned by a Bluetooth-based exposure notification protocol released in May by Google and Apple in a rare joint venture. Smartphones running apps using the Google-Apple technology can exchange randomly generated identification numbers with other nearby devices; the apps then alert users if someone they’ve been in contact with later inputs a positive COVID-19 test so they can take appropriate measures, like getting tested. The idea was to augment, not replace, traditional contact tracing—a manual process in which human investigators interview infected individuals, then contact others with whom they recently spent time. (While many states have been ramping up their contact-tracing programs, many of these efforts have been overwhelmed by the sheer amount of viral spread in recent weeks.)

But five months after the Google-Apple project launched, apps using their protocol are available to the general public in only 10 states and Washington, D.C. Even in states that have rolled out contact-tracing apps, adoption generally remains low. Why?

Part of the problem, according to public-health experts, has been a lack of coordination by the federal government, which could have, for example, created a national digital contact-tracing solution and encouraged states to opt in. Absent direction or incentives from Washington, many states have chosen not to launch contact-tracing apps at all. Even states that have launched contact-tracing apps were initially wary of investing their limited resources in an unproven solution. Officials in New York, for example, told TIME they were interested in Google and Apple’s initial pitch as the pandemic battered the Empire State this spring, but first wanted to shore up their traditional contact-tracing program. The state eventually launched an app in early October.

Contact-tracing apps have also been slowed by state health departments’ lack of tech expertise, according to public-health officials and technologists. “From the perspective of an app developer that sat in those contracting queues in places like New York and California, states were utterly unequipped to start making procurement decisions on contact-tracing apps,” says Teddy Gold, executive director of Zero, a non-profit formed this spring to make pandemic response software. “You’d get sent from the public health department to the governor’s office, to the [chief information officer], back to a mayor’s office, back to the chief information security officer’s office. It was this Kafkaesque thing where no one had ever done this. No one had ever developed a contact-tracing app before. States don’t develop apps.” Dr. Norm Oliver, Virginia’s state health commissioner, agrees. “Public-health departments around the country, their strong suit is not going to be app development,” Oliver says.

At first, states using the Google-Apple protocol had to contract with developers on their own to build contact-tracing apps, an expensive and complicated process. In September, with only a handful of state apps online since the protocol’s May launch, the companies released Exposure Notifications Express, a basic, pre-built version that states can use instead of developing their own apps. Users in areas where the service is active can opt in after receiving a push notification, a feature likely to save marketing costs. A spokesperson for Google and Apple says the companies are seeing “momentum” in adoption among states thanks to collaboration with public health authorities and ongoing software improvements.

Some officials say the technology companies have done the best they can considering the constantly changing pandemic situation and the inherent complexities of public-health administration. “Apple and Google as tech companies, I think they were trying to find where they could best fit in to help, using technology that they knew was available without saying that they’re the public health authority, which they’re not,” says Jeff Stover, an executive advisor at the Virginia Department of Health. But getting a contact-tracing app up and running is still not as simple as flipping a switch, and some technologists and health experts are frustrated over delays in getting these apps into wide use. “We’re 200,000 dead people late,” says Gold. “It’s like the plane has crash-landed and everyone has died and the captain scrambles out of the rubble and he’s like, ‘Okay guys, in an emergency landing there are lifejackets under your seats.’”

Of course, whether an app takes a week or six months to build doesn’t matter if people aren’t downloading and using the resulting software—a problem in most states that have launched contract tracing apps. Alabama’s contact-tracing app, for example, was downloaded only 125,000 times between its release in mid-August and late October, a figure equivalent to just over 3% of the state’s adult population. Wyoming, which launched a contact-tracing app around the same time, has seen fewer than 5,000 downloads as of late October, equivalent to just 1% of the state’s adults. Apps in North Carolina and Pennsylvania had been downloaded by an equivalent of only 3-4% of their adult populations by the end of October, just over a month after launching. In New Jersey and New York, which both launched apps on Oct. 1, an equivalent of around 4% of resident adults signed up for their apps in less than a month.

That sluggish user adoption can partly be explained by people’s unfounded fears that contact-tracing apps are tracking their location or other personal information. “Concern about privacy is one of the things that’s suppressing adoption,” says Christian Sandvig, director of the Center for Ethics, Society, and Computing at the University of Michigan. That’s despite the fact that the Google-Apple protocol—which doesn’t track or share users’ locations or identities—represents the “gold standard” for privacy protection, Sandvig says. Some health experts have even argued that these apps were built with such an emphasis on privacy that they’re actually less useful in fighting the pandemic, in part because they don’t collect data like the locations where potential infections are taking place.

But many users may not see it that way, especially in an era when Americans’ trust in Big Tech is eroding and technology firms are catching flak from all sides of the political spectrum. In some instances, privacy concerns are even killing contact-tracing apps in the cradle—South Carolina, for instance, announced plans in May to deploy a Google-Apple powered contact-tracing app, only to shelve the plan the next month after lawmakers banned such software over privacy concerns.

In some states, sparse adoption may also be linked to a lack of advertising funding. In the two U.S. states with the highest adoption rates—Delaware (7.3% adoption) and Virginia (10.6% adoption)—officials have spent $0.11 and $0.18 per resident on advertising their apps, respectively. Officials in those states attribute their relative success to aggressive outreach efforts; in Virginia, that included a PSA featuring students from around the state as well as marketing materials in both English and Spanish. But in Wyoming and Nevada, where adoption sits at a paltry 1.1% and 2.9% of resident adults, respectively, advertising funds are scant. Wyoming isn’t spending any money at all to promote its app. The private-sector partners behind Nevada’s COVID Trace, who also paid to build the app in the first place, are spending a small amount on ads—around $0.03 per resident, plus ad inventory contributed by Google and a volunteer effort from Vegas-area performers—while the state is spending none of its own funds. The strategy has so far failed to bear fruit. “I feel like I’ve talked about COVID Trace every day since we launched it, and people will say ‘oh I didn’t know you did an app,’” says Julia Peek, Deputy Administrator of Nevada’s Community Health Services. “It’s like, ‘what are we doing wrong to promote this?’”

While it’s obvious that no U.S. state has achieved anything resembling a satisfactory adoption rate, it remains unclear how many people in a given population need to download and use contact-tracing apps to control viral spread. In May, researchers pegged that figure at 60%—far more than what U.S. states are seeing so far. But newer research from Oxford University and Google says that exposure notification apps could help reduce infections at any level of uptake. “There have been a lot of conversations in the past about [whether] you have to achieve minimum thresholds in terms of adoption levels,” says Larry Breen, chief commercial officer at Nearform, which developed a contact tracing app for Ireland as well as multiple U.S. states. “I’ve never accepted that as the right thing. As soon as you get the digital contact-tracing solution out into any cluster or group of people, it’s providing some level of protection.”

Even the highest adoption rates among U.S. states are far below those in countries like Ireland, where more than a third of the adult population downloaded the government’s contract-tracing app by early October, or Germany, which reached 27% adoption in September (though even both of those results fall well short of mass adoption). Breen says Irish officials and politicians have maintained a unified, consistent message promoting the app, which hasn’t been true in the U.S. “There’s a lot of confusion and different messaging coming out,” he says. Other countries, like Austria, have comparable adoption rates to Virginia.

All told, despite the relative success of states like Delaware and Virginia, agonizingly slow rollouts and uncertain public health benefits over the past few months have caused some experts to doubt the assumption that contact tracing apps can help bring the spiraling U.S. COVID-19 outbreak under control. “There’s an ultimate question here…which is, ‘Is this a great opportunity for software?,'” says Sandvig. “It may be that it is not.”