Sam Aldridge was sceptical when she first heard about Australia’s contact-tracing app. The 30-year-old, who works at a design studio in the country’s second city Melbourne, was wary of her government’s track record of data breaches and concerned about how data from the app, Covidsafe, would be held on an Amazon Web Services server.
But the overwhelming desire to see her life return to normal prompted her to download it anyway. As she put it: “I’m at the point where, if it’s going to mean we can start going out again, I’m in.”
When the app is open, her iPhone logs the Bluetooth signal of nearby phones, from the friends she’s met to strangers she might sit next to on the bus. If Aldridge tests positive for Covid-19, Covidsafe can decide who else might be at risk, warning them to get tested or enter quarantine to halt any further spread of the virus.
Much is riding on the efficacy of the app, which was launched on 26 April 2020. Contact-tracing technology has been sold as a life-saving tool and a ticket out of lockdown. Australia’s prime minister Scott Morrison said the easing of the country’s restrictions was dependent on how many people run Covidsafe on their phones.
Covidsafe places Australia among a small group of early adopters, where government-endorsed contact-tracing apps are already in use. India’s Aarogya Setu app was made mandatory in mid-May for workers and people living in “containment zones”, but these rules have since been relaxed. Singapore, Iceland and Norway have also launched similar apps.
As we work on The Correspondent’s Track(ed) Together project, we’ve found at least 80 contact-tracing systems in place or in development around the world. Faced with a difficult choice, millions of people are making the same decision Aldridge did – to download apps they are not entirely comfortable with because the technology promises to liberate them from lockdown.
Despite the early flurry of activity to develop these apps, we’ve noticed that enthusiasm has waned and the rhetoric surrounding the technology has shifted significantly in the past month. In countries where these apps are already in use, the technology is facing serious challenges: mounting problems with low download rates, accuracy and usability are mounting. Governments are waking up to the realisation that technology cannot offer an easy way out of this pandemic.
Pandemic surveillance? There’s an app for that
To see this technology as a panacea is problematic for various reasons. But first, what do we know about the 80 or so contact-tracing apps we’ve identified so far?
Some apps have been built transparently, with clear privacy policies and open-source software, so everyone can check how they work. But in many more instances, this information is missing, making it difficult to find out what happens with users’ data.
Bluetooth contact-tracing technology has proved popular across the world, and we’ve tracked at least 40 apps using a system where phones exchange Bluetooth “handshakes” when they cross paths. These “handshakes” create a record of who has met whom – and if one user tests positive for coronavirus, who else could be infected.
However, there is disagreement about how the data these apps collect should be stored. Some countries have opted for a centralised approach, where every phone is assigned its own ID code and anonymised records are sent to a centralised database. This approach means “contact-matching” – the process of deciding who may have come into contact with an infected person – takes place on a central server. Advocates for centralised systems argue that it allows health experts access to information about how the virus is spreading.
Other countries have chosen instead to make their apps compatible with a decentralised system now offered on Android and iPhones by Google and Apple. Here, the data is encrypted and stored on individual phones and not on a central server, with contact-matching being done on the phone itself. The proponents of this approach say it makes it harder for hackers or governments to access a person’s list of contacts.
Beyond Bluetooth, we also found at least 11 apps that use GPS as a basis for contact tracing.
Problems with accuracy and efficacy take the shine off contact-tracing apps
Despite the eagerness for life to return to normal, governments are struggling to persuade significant numbers of their country’s population to download contact-tracing apps. In a country of 25 million people, as of 5 May Australia’s app has been downloaded just five million times. Some experts think about half, or even more, of the population is required for these apps to be effective. As of 29 May, Aarogya Setu has been downloaded more than 100 million times, but that is a small fraction of the 1.3 billion people living in India.
Even when downloaded, they are not necessarily used. Apple does not let Bluetooth apps with centralised storage systems run in the background of its devices, meaning the app has to be open on the screen to work properly. This drains the battery of the phone and can interfere with listening to music or podcasts, making it a nuisance for users.
In countries where this centralised model is in use, Apple’s restrictions have been blamed for the apps’ poor functionality. According to The Norwegian Institute of Public Health, as of 28 April 2020, only 20% of the 1.5 million people who downloaded the app were actively using it.
There are also doubts about the accuracy of these technologies. GPS is imprecise, especially in and around high-rise buildings. Many phones are not suitable for Bluetooth contact-tracing and sometimes it’s difficult to measure the distance between people – apps have been known to register “a meeting” even if people are on opposite sides of thin apartment walls.
This imprecision could lead to false positives, which in turn could swamp health authorities. Instances when a contact isn’t recognised by the technology – false negatives – are also likely, potentially leading to a false sense of security and greater risk-taking behaviour.
As with concerns about accuracy, there is much doubt about whether contact-tracing apps actually work. “We haven’t seen any evidence to suggest that the app is being more useful than traditional methods of tracing yet,” said Meru Sheel, infectious diseases epidemiologist at the Australian National University, who adds the app could still prove to be effective if Australia sees a second wave of infections.
‘The use of these apps will have consequences for how the pandemic is fought. You have to be sure that it’s not in a negative way’
Paul-Olivier Dehaye, a Belgian data protection activist, regards the rollout of contact-tracing apps sceptically. He says it’s important to acknowledge that these apps are one big experiment – one on entire populations. “The use of these apps will have consequences for how the pandemic is fought. You have to be sure that it’s not in a negative way.”
Advocates promise these apps will enable contact tracing to be done more quickly and at scale, but according to Dehaye, that’s not a given. “Suppose you have to trace 50 contacts. Will you test them all? Isolate them all? Are there enough resources for that? Maybe there are many contacts that run little risk, but you miss high-risk people because they didn’t use the app or the technique didn’t work properly.”
The Switzerland-based mathematician Dehaye also thinks that focus on contact-tracing apps has crowded out the development of alternatives, which he calls a missed opportunity.
Not everyone is wary. In Latvia, where there are plans to launch one of the first apps that will use Google and Apple’s new API, Andris Berzins argues that if implemented correctly, the technology has the potential to save lives, even though he also acknowledges that there’s no evidence to support that view yet. “The honest answer is until we put it out there and prove that it’s used, some people will have to take a leap of faith,” said Berzins, who’s part of the coalition developing the Latvian app.
His comment alludes to the idea that the technology has not yet had the chance to prove its effectiveness. But for sceptics, the question remains if we’re placing too much trust in – and handing over our data for – unproven technologies. These concerns were voiced by the Council of Europe which asked in a statement: “Considering the absence of evidence of their efficacy, are the promises worth the predictable societal and legal risks?”
Government doubts reflected in strategy
And so, as doubts have grown in number and volume, governments which once seemed emphatically in support of contact-tracing apps are now using much more cautious language. “The language from public officials has been toned down,” wrote Josh Taylor, a reporter at the Guardian Australia, less than a month after Australia’s Covidsafe was launched. “No longer is [the app] the key to freedoms, but an add-on to existing contact-tracing methods, to work in concert with social distancing rules and continued testing to keep a lid on outbreaks.”
Irish ministers, who once called the app “a key element of the next phase”, also now talk of technology playing “a role, not a major role”.
“There was a lot of hype about apps in the beginning and it seemed like they could be a big part of contact tracing. But as time goes on, politicians have become more reticent,” said Antoin O Lachtnain of the non-profit Digital Rights Ireland. “There have been a lot of technical problems and concerns about privacy that have meant that the app uptake [around the world] hasn’t been high. At the same time, there is no clear evidence that apps actually help with contact tracing in practice.”
In the Netherlands, a similar shift in attitudes has been observed. The ministry of health first presented the app as a precondition for easing the lockdown, but now, as Dutch society begins to open up again, there’s still no app and hardly any further debate about it. The emphasis has changed to manual contact tracing, with the app only intended to be used as “digital support”.
Even in countries such as Iceland, where the government-endorsed app had been downloaded by 40% of the population, the results appear underwhelming. An official in charge of contact tracing said recently that the app had been “useful” but “wasn’t a game-changer”.
Ardern: ‘It will be a helpful supplementary tool.’
The new, more cautious tone in Ireland, the Netherlands and Iceland echoes statements issued by New Zealand’s prime minister Jacinda Ardern, who has always been cautious about the merits of contact-tracing apps. "I don’t wish for us to rely on [a contact-tracing app] being the answer, because it never will be,” Ardern said in May. “It will be a helpful supplementary tool."
A move to QR codes
When New Zealand launched its app on 20 May 2020, it marked a departure away from the contact-tracing technology introduced elsewhere. NZ Covid Tracer uses a QR system instead of GPS, location data or Bluetooth. When a person arrives at a restaurant or office, they scan the QR code near the entry or exit so the visit is registered. The details of their recent movements are kept in a “digital diary” and stored on their phone for 31 days before being automatically deleted.
Without needing to send their location history, the user of the app can review all the places where they have been if made aware by a healthcare provider that they’ve come into contact with an infected person.
“The QR codes are technically simpler than Bluetooth, so there are fewer technical challenges, and can be deployed quickly,” said Andrew Chen, a research fellow at the University of Auckland’s Koi Tū Centre for Informed Futures, adding that QR codes can make it easier for businesses to maintain their contact-tracing “registers”, which are now required by law.
Chen: “The notion of using contact-tracing registers is particularly important for helping manual contact tracers do their job more comprehensively, and so [a QR code system] makes sense in a country like New Zealand, where it has been made clear that our contact-tracing processes will be driven and controlled by people, rather than by a decentralised app.”
After Singapore’s TraceTogether app, which was applauded as an early model for digital contact tracing, reported stubbornly low download rates, the country introduced its own QR code system. SafeEntry requires people visiting offices, supermarkets, malls and schools to scan a QR code on entry to assist contact tracing efforts. Singapore’s GovTech department did not reply to The Correspondent’s email asking whether SafeEntry was designed to replace TraceTogether.
QR code systems aren’t necessarily more privacy friendly. While New Zealand’s system anonymises data through encryption and limits access to it to health officials, in other places location logs are centrally stored in combination with phone numbers and national identification numbers. Some QR-scanning schemes don’t even seem to have a privacy policy.
And as with other methods of contact tracing, QR scanning cannot guarantee 100% accuracy – the information you get simply tells you who was in the same venue at the same time. In larger spaces, there is no knowing if those people did indeed come into contact.
We found at least 16 use cases of QR tracking. In Switzerland, a QR-scanning app, which uses a technology called GastroCovid, is used in some restaurants, so other patrons who have dined in close proximity to an infected person can be found.
In some cases, such as Thailand and Malaysia, governments have made QR scanning a precondition to exiting the lockdown. In Thailand, millions of people are using the ThaiChana app in order to enter shopping malls and stores.
Tried and tested: the return to manual contact tracing
Some countries, such as Belgium, are rejecting contact-tracing apps all together – at least for now. “There is no need for an app for contact tracing, it can be done manually and it has been around for years,” telecommunications minister Philippe De Backer told Belgian television channel VRT on 23 April. De Backer has said it is not “feasible” to persuade an entire country to download an app.
‘Human contact tracing is so old and so well known, we know what we need to do, we know how it works, we know that it’s useful.’
Instead, Belgium has focused its efforts on human contact tracing, announcing it was hiring 2,000 people at the end of April. “Human contact tracing is so old and so well known, we know what we need to do, we know how it works, we know that it’s useful,” said Anne-Mieke Vandamme, Belgian epidemiologist at KU Leuven University.
Although a centralised contact tracing app is currently being trialled on the Isle of Wight, the technology’s importance to the UK’s “track and trace” system is also being downplayed as the government tries to shift focus to its newly recruited army of 25,000 human contact tracers. Lord Bethell, a junior minister of innovation at the Department of Health and Social Care, said: "We have therefore changed the emphasis of our communications and plans to put human contact tracing at the beginning of our plans and to regard the app as something that will come later in support."
The UK did use human contact tracers for six weeks between January and March but stopped the scheme on 12 March, arguing the system was no longer effective when the disease was so widespread. In contrast, early aggressive deployment of human contact tracers has been suggested as one reason why sub-Saharan Africa has seen less explosive transmission than Europe and the US. Countries such as Rwanda, South Sudan and Uganda were able to quickly repurpose existing infrastructure, such as teams of trained contact tracers developed in response to the Democratic Republic of Congo’s Ebola outbreak in 2018.
Now Massachusetts is one US state attempting to replicate this approach by budgeting $44m to hire 1,000 people and working with the non-profit Partners in Health, whose doctors have led responses to Ebola, Zika, drug-resistant tuberculosis, among other highly infectious diseases in some of the world’s poorest countries. Japan’s success curbing its coronavirus infection rate has been attributed in part to using this technique since January, when the first cases in the country were identified.
Sometimes, there’s just no app for that
When the coronavirus arrives in a country, its government usually responds by looking to the technology sector for fast, cost-effective solutions that can deliver immediate results, and there will always be companies who promise to deliver. This pattern existed before the coronavirus, and it will outlive it. But as countries grapple with a threat that could leave hundreds of thousands dead, the promises of solutionist tech – the idea that all problems can be solved with the right app or algorithm – have mushroomed.
However, as the shift back to manual contact tracing demonstrates, technology is not always the answer. There’s not always “an app for that”. Contact-tracing tech has been cast as the star role in national efforts to curb the virus, but in this crisis, it is, at best, a supporting character alongside comprehensive testing and human tracing regimes. “No contact-tracing app can be used on its own; they [should all be used] in the context of a lot of other measures,” said Vandamme.
But governments have been distracted by the promise of easy solutions in the form of contact-tracing apps. The problem is, right now is a very dangerous time to be distracted.
You can read the Dutch version of this article on De Correspondent.
Dig deeper
Stay up to date with Covid-19 surveillance
Every Monday, Morgan Meaker shares the latest on coronavirus surveillance from our global database of tech experts and journalists.
Covid-19 surveillance tech explained: 6 ways governments are monitoring the virus – and you
Desperate to control the spread of the coronavirus and lift restrictions, governments around the world are turning to technology. We list the six most important technologies being used and explore the risks and limitations of each. We still need your help to get the global picture.