The news on Friday (April 10) that Apple and Google are partnering to simplify coronavirus contact-tracing is a big deal. Just like that, read the headlines, more than 3 billion people globally might have an effective warning system if they come into contact with newly diagnosed COVID-19 patients. Unfortunately, that’s not the case. Putting aside how many devices actually carry the right Bluetooth technology, there are two critical factors that stand in the way of this being effective.
Bluetooth contact tracing uses a relative signal strength indicator to detect when one device is near another, and for how long. Your phone collects unique identifiers for those other phones you are near throughout your day, those other phones do the same for you. Your phone also downloads unique identifiers for those newly testing positive for COVID-19. If there’s a match, you receive a locally relevant alert—monitor for symptoms, get tested, self-isolate—without breaching your privacy.
Apple and Google have joined forces to remove the core technical issues associated with a bluetooth contact tracing app. This will provide an API initially, and core OS functionality beyond that, for national programs to build upon, making their apps as effective and user-friendly as possible, working in the background, preserving battery life, safeguarding data privacy.
Let’s take a look at Singapore. The country’s TraceTogether was launched last month and has become the public catalyst for a global push towards the privacy-friendly use of bluetooth proximity tracking to alert users who may be at risk. The push to such platforms in Europe and the U.S. reference Singapore as an example of what good looks like. The software behind TraceTogether has now been made open-source, as Singapore encourages other countries to follow its lead.
Singapore was lauded for its fast approach to COVID-19, built around rigorous contact-tracing for new patients, although it is now suffering a second-wave of confirmed infections. Deploying a government-backed system in Singapore is very different to Europe and North America. The culture is more compliant and citizens are subject to far more surveillance and control than we see in the west. Even so, only around one million citizens, some 20% of the population, has downloaded the app.
“In order for TraceTogether to be effective, we need something like three-quarters—if not everyone—of the population to have it,” the country’s development minister Lawrence Wong told local media early this month. “Then we can really use that as an effective contact-tracing tool.” Not only do citizens need to install the app, they need to get it up and running on their devices. Some of those technical challenges are resolved by Apple and Google—the decision to install, though, is not.
This is the first critical issue with making such voluntary contact-tracing apps work—inertia versus compliance. China is the only country that can claim highly-effective digital contact tracing thus far, and that is because it did not require any citizen compliance. The country simply deployed its existing surveillance state. That’s not an option elsewhere. As Oxford University researchers in the U.K. (which is also building a national bluetooth app) have warned: It only works “if used by enough people,” a figure thought to be around 60%. TraceTogether’s million installs is “a record for a [Singapore] government app,” highlighting the scale of the challenge.
Beyond the highest level numbers, there are other issues—the percentage of older citizens with capable smartphones, for example, exacerbating the age disparity we are already seeing with COVID-19. Older or less healthy citizens worry about becoming infected. Younger, healthy citizens less so. Given behavior and mobility, younger citizens are likely a major cause of new infections. Not only are they far less likely to install an app, given fewer health concerns, they are also much more likely to be asymptomatic or to suffer a mild infection and not seek testing.
And that brings us to the second critical challenge—testing. Bluetooth tracing apps link to national health efforts to test and monitor the population. You cannot have an app at scale where anyone can just tick a box to say they have symptoms and then all those they’ve been near immediately self-isolate. That would be chaos. Instead, a formal test would flag a user’s infection and the app would then alert those they had been near. If most people do not get tested, it does not work.
Much of the focus on the Google and Apple announcement has been on technical privacy and ease of use. It is beyond doubt that for such bluetooth contact-tracing to work, it needs this move by the smartphone OS giants—but that doesn’t make it happen. The best analogy is to suggest that better battery technology in of itself created the electric car industry—a major factor, yes, but it was far more complicated than that.
The risk with the type of publicity greeting Apple and Google’s move is that it has welcomed the good news aspects and overlooked the “no news” aspects. There is a mountain to climb to get even 60% of national populations to use an app or enable an OS function, never mind 75% or more. Let’s not forget, the idea for bluetooth contact tracing is almost a decade old—the FluPhone app was touted by Cambridge University back in 2011. It did not catch on, despite WHO estimating that “290,000 to 650,000 respiratory deaths occur each year associated with seasonal influenza.”
Beyond the technical enablers, the privacy safeguards touted by Apple and Google should also be welcomed, taking the platform out of government hands to protect privacy, limiting access to data. In that regard, an ACLU spokesperson welcomed an approach that “appears to mitigate the worst privacy and centralization risks,” while warning that “there is still room for improvement [and] any contract tracing app [should] remain voluntary and decentralized, and used only for public health purposes and only for the duration of this pandemic.”
The DP-3T team working across eight European universities, including University College London, has been campaigning against centralized tracking and tracing apps, arguing that data must be kept on devices. As such they welcome the Google and Apple news.
Michael Veale, lecturer in digital rights and regulation at the university, told me this “enforces a decentralized approach, where no personal data leaves your device—The shape of the protocol is very much the same as ours,” he explained, although the tech giants can go much further in the actual implementation, coding it into the OS, removing the need to keep phones on or play with bluetooth settings.
That said, Veale also warns that take-up below 60% “would endanger the quality of the data and its core functionality, so that’s where the focus needs to be,” adding that “if testing is not widely available, I think an app provides more false security than actual support fighting the virus… Making it solve societal problems in practice, however, needs much more than just a technical protocol.”
ACLU also acknowledged these two critical challenges to making such a system work, “widespread, free, and quick testing,” as well as generating unprecedented levels of trust in a government-backed platform to push people into installing the app, with “such contact tracing likely to exclude many vulnerable members of society who lack access to technology.”
Unless and until governments can either develop or mandate a system that deploys this across the majority of its population, and then backs it up with the rigorous testing regime that is stitched into the core concept of operation, such apps will be helpful but not game-changing. As countries around the world progress the development and launch of their apps, all of which will benefit from the Apple and Google move, you can expect to see take-up and testing as the twin pillars of success.
To make this effective, we may need to accept some form of mandatory proximity tracing as the pandemic evolves. This is already being explored in the U.K., where its own version of a bluetooth contact-tracing app might be mandatory for those returning to work, according to the Sunday Times today (April 12). As unlikely as that sounds now, think how much has changed from a tracking perspective in recent weeks.