Computing Applications

Exit From the Covid-19 Lockdown: There’s an App for That

Can an app help people practice social distancing for safety?
In search of a viable contact tracing app.

In a display of boldness which is quite rare in Dutch politics, health minister Hugo de Jonge announced in a press conference on April 7 that the exit from the partial lockdown the Netherlands has been in since mid-March would be conditional on a new app for tracking and tracing people infected with Covid-19. After stating that at least 60% of all residents of the Netherlands should download the app voluntarily, he just stopped short of saying that; otherwise, everybody might be forced to download it.

That same week, an 'appathon' was held, an expert meeting in which seven commercial parties presented app designs for that purpose. The designs had to be open source and address privacy concerns preemptively, but the government offered little further guidance, and did not specify things like who would process the data, and how it would be protected.

The backlash from privacy action groups was massive. A group of IT experts and academics signed an open letter to the government, warning against rushing into 'techno-solutionism'.

The Netherlands' official privacy watchdog, the Dutch Data Protection Authority (Autoriteit Persoonsgegevens), said the app designs submitted were so rudimentary, it could not express an opinion on them.  

The basic idea for the app is that it would use Bluetooth signals from other smartphones to detect their proximity. If another phone comes closer than 1.5 meters (nearly 5 feet) for a specified length of time (typically 15 minutes), the app will register it as a high-risk contact. If the smartphone user later tests positive for Covid-19, the app can trace his/her high-risk contacts back over time, and send a warning to these persons to self-quarantine and get tested.

Since people infected with Covid-19 can be contagious up to three days before symptoms appear, early warning could significantly decrease the spread of the virus. Dutch health authorities believe such an app will be highly effective, indeed essential, for containing new outbreaks post-lockdown.      

With respect to privacy, Bluetooth is an improvement over GPS location because it does not register the location of either person in a high-risk contact, only their proximity. Also, the system could be set up without a central authority, only sending data from one phone to another, while alerts need not identify their sender.

But most apps presented in the appathon used unique identification numbers, and assumed a central authority to process the data.

In the real world, other problems will crop up. Even if someone downloaded the app voluntarily, would he/she promptly self-quarantine on an alert from an unspecified person, for some unspecified high-risk contact? If he/she doesn't, should the app notify the local GGD (public health service) about this? Even if using the app is not compulsory, would employers have the right to demand their employees download and use it? Would airlines have the right to refuse boarding to customers who can't show a green light from their app?

Holger Hoos, professor of machine learning at Leiden University and currently under lockdown with his family in Germany, signed the open letter because he worries that such an app "opens the door to decisions that will erode privacy permanently. Today it's corona, tomorrow there may be some other reason for contact tracing. Maybe you have been in proximity contact with this terrorist or criminal?"

Hoos points out that Apple and Google want to program this tracing technology deep into their operating systems. Contact tracing could become the default option, with all phones exchanging random tokens every few seconds, which can be linked afterwards, if deemed necessary. Says Hoos, "Sure, it's based on opt-in/opt-out, but what prevents governments from saying 'you can opt-out, but then you can't get into our country.' The U.S. Border Patrol already has the authority to look into your phone. This is deeply troubling, and should be outlawed in Europe."

Norway and Iceland have experimented with tracking and tracing apps, and the U.K. is working on one. Singapore deployed a tracking app in February, but that still relied heavily on manually tracing high-risk contacts, and just 16% of the population downloaded it. Even if the majority of people would happily sacrifice their privacy for public health, it is far from sure such an app will actually work.

In a recent study based on a large survey of self-reported social contacts and an epidemiological model, U.K. scientists estimated that with a close-contact distance of 2 meters (6.5 feet) and a minimum time of 15 minutes, 1 in 5 cases of COVID-19 would cause an undetected new case. Each new case, on average, would leave 36 high-risk contacts in its wake. 

The U.K study, however, assumes human tracking and tracing are taking place. There is no experimental data on how well a proximity app could accomplish such a task. Consider that the Bluetooth signal travels through glass windows and walls, while the virus does not; how many false alarms will that generate?

It also is not known how well proximity predicts risk of infection in the absence of physical barriers.

Even if the app is as good as humans at identifying high-risk contacts, for each new case, it will tell 36 people to self-quarantine, while only two or three are really infected. A system with that many false positives will quickly lose public trust.  

Epidemiologist Jaime Borjas Howard still thinks an app could be a useful, but not by itself. "In my view, this app should alert high-risk contacts to report to the GGD. The GGD could then use this app to reconstruct the chain of infection. To me, this is a tool to optimize tracking and tracing, not a panacea to eradicate the epidemic. So, the capacity of the GGD should also be increased."

To Borjas Howard, an initial large number of false positives is not necessarily a show-stopper; it could be part of the learning curve to optimize the app and the human tracking and tracing process.

After the surge of criticism, all seven candidate apps from the appathon were rejected, and Holland's deputy prime minister and minister of Health, Welfare, and Sport Hugo De Jonge is now assembling a group of experts to develop the app in-house. This plan also got a rather mixed reception, as the Dutch government has an impressive track record of expensive, failed IT projects. 

Arnout Jaspers is a freelance science writer based in Leiden, the Netherlands.

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More