Last week I wrote about a set of predictions that I got spectacularly wrong
. Today, to soothe my fragile ego, let’s talk about a case where I was a closer to the mark.
In April, after talking with public health experts, I wrote about why Bluetooth-based approaches to finding new cases of COVID-19 seemed likely to fall short
. A combination of human behaviors and technical challenges seemed likely to limit the effectiveness of a pandemic response scheme that relied on our phones passively pinging each other wherever we went. At the same time, I wrote, a more old-fashioned approach — using human beings to make phone calls and identify people who may have been exposed to the disease — seemed much more promising.
Somewhat hilariously, at least to me, the very next day Apple and Google announced that they were building a Bluetooth-based COVID tracking system
. (Originally they branded it as “contact tracing,” but after talking to more people they switched to the more accurate “exposure notification.”) My qualms about its effectiveness aside, it was clearly a good-faith effort to make a real difference in reducing the spread of COVID-19, and over the next several weeks Apple and Google worked diligently to refine their approach based on the feedback they were getting.
The first phase of the project, which allows whitelisted public health apps to make enhanced use of the Bluetooth features of your phone, launched on May 20th with three US states on board
. The hope was that more states would sign on, creating a more effective patchwork of notifications across the country. But as David Ingram reported at NBC News over the weekend, that’s not what’s happening
States that had committed to using contact tracing apps or expressed interest are now backing away from those claims. The few states that have rolled them out have seen only tepid responses. And there are no indications of any momentum for the apps at a national level.
California Gov. Gavin Newsom was an early supporter of contact tracing apps, saying on the same day that Google and Apple announced a joint project that he wanted to ”build that capacity and partnership”
as part of the state’s plans. Two months later, the nation’s most populous state isn’t using any apps or cellphone tracking technology, said Ali Bay, a spokesperson for California’s Public Health Department.
Among the challenges: it’s hard to get people to download these apps; there has been no endorsement of the approach — or coherent response strategy — from the federal government; and it’s unclear that Bluetooth-based approaches have been all that effective even in countries where COVID-19 is better under control. (Norway abandoned its contact-tracing app on Monday after it reached just 10 percent adoption
.) It seems possible that the second phase of the Apple-Google plan, in which exposure notification becomes a system-level feature of iOS and Android, could give exposure notification new life.
In the meantime, though, it’s hard not to feel like the tech giants’ work to date has been mostly for naught. They lived up to their end of the bargain, building a workable system with meaningful privacy protections in a very short period of time. But then the federal government all but went missing, leaving the pandemic response largely to individual states and their governors, and much of the promise of exposure notification evaporated as a result. (One thing the federal government did do was promote public-private partnerships to increase testing capacity, and that effort also stumbled as this piece about a Utah initiative by Robert P. Baird lays out
So what is
working? I’m pleased to say that human beings making phone calls have largely lived up to their promise. In the New Yorker
, Ben Wallace-Wells looks at the experience of Massachusetts, which rapidly spun up an effective contact-tracing scheme
known as the Community Tracing Collaborative. There aren’t a lot of feel-good stories in pandemic news these days, but this is one of them. When Massachusetts set up its collaborative, more than 22,000 people applied, and in the past few months they’ve effectively identified a huge chunk of potential new exposures. Wallace-Wells writes:
Throughout the spring, the Massachusetts contact-tracing program got faster. It took between three and four days for the C.T.C. to learn of a positive test, but after investigators had that information they were able to reach seventy per cent of cases, and contact tracers were then able to speak to seventy-four per cent of those cases’ contacts. This still meant that nearly half of potential contacts never spoke with anyone working for the tracing program. Crystal Watson, a health-policy expert at Johns Hopkins, who co-authored a report in April on contact tracing, told me that these numbers were better than she had expected. “Even if you’re only getting half of all contacts, that’s a lot better than we would have been doing without the program,” she said.
Wallace-Wells reports that the federal government has allocated funding for contact tracing nationwide, but only at about one-sixth of the $3.6 billion experts say will be necessary to get the job done. Among the missing pieces: funding for what the public health world calls “care resource coordinators,” who help solve practical problems for people who are sick and need to isolate — getting them food, for example, or medicine. Public health experts say problem solvers like these are necessary to ensure that sick people remain isolated, but cash-strapped public health agencies are reluctant to pay for them.
The lesson in all this is not that Apple and Google should not have built an exposure notification system, or that the system they built is a failure. Rather, it’s that Silicon Valley was only ever going to be able to provide one or two pieces of a comprehensive response to COVID-19 — and elected officials and public health agencies have mostly failed to provide the rest. In April, it was easy to see that Apple and Google’s approach alone couldn’t stop the spread of the disease. What I couldn’t foresee back then was just how alone they would be.