Imagine a new virus that is only known to be spread through the air by contact with other people. Imagine that in the first wave of this virus there are almost 30,000 deaths and hospitals are overflowing. Imagine that there are still no vaccines or rapid tests to warn of a positive result. Imagine that you have a tool, on your cell phone, to find out if any of the strangers you have crossed paths with are infected, if you have had a risky contact, if you could have been at risk. Imagine how useful that tool could be. Imagine, because the reality of Covid Radarthe application that was born to help contact detection in the covid-19 pandemic, was another.
Three months after the pandemic broke out, the government launched on the island of La Gomera the pilot test of a mobile application that allowed the user to be notified of their possible close contact. with a positive for covid-19. On an emergency basis, like most in those months, Indra was awarded the contract, which amounted to 330,537 euros. The results of the pilot were “positive”, as concluded the study published in the prestigious scientific journal ‘Nature’, and justified its “nationwide deployment”. By the end of the summer, the Government presented in society Radar Covid and awarded without advertising a new contract to the same company for the maintenance and implementation of what promised to be a key tool to help bend future pandemic infection curves.
Two years later, the positives confirmed by the application are just over 124,000.In addition, the Secretary of State for Digitalization has been warned for having violated eight articles of the data protection regulation and the bill amounts to 4.2 million euros, about 50,000 euros per code.
“We showed that, with 30% of people with the application, we were able to recognize up to 6.3 contacts per infection and that’s barbaric”, he points out. Álex ArenasD. in physics from the Universitat Rovira i Virgili and co-author of the La Gomera study. “But I think that the government did not insist too much for this application to be established”, he adds. For another of the authors, the specialist in Preventive Medicine and member of the General Directorate of Public Health of the Canary Islands, “I don’t think the government insisted too much on this application being established. Eva Elisa Alvarez-Leónshould have been implemented simultaneously and completely throughout the country” and this was not possible because it had to be adapted to “17 partially independent tracking systems”, one for each autonomous community.
The study published in ‘Nature’ already pointed out that some of its results had to be treated with “caution”, and warned that for Radar Covid to be successfully adopted “a major communication campaign” was needed. The government spent about €2.2 million. on publicizing the application. Even posters with QR codes to download it appeared on the seats of Renfe trains, under the slogan ‘Protect yourself and others’. An OCU survey in September 2020 indicated that 90% of the population had heard of Radar Covid. But it was not enough. In La Gomera there were volunteers explaining on the street how the app worked, something that did not happen at the national level. “A communication campaign was needed to make people understand what it was for, how it worked and what we gained,” insists Arenas.
The information leaflets talked about the fact that Radar Covid doubled the manual tracking capacity and, with 20% of users, the pandemic was reduced by 30%. The penetration rate was 21%, but it is not possible to know how many downloads are from the same user or how many have ended up uninstalling it.
The problem of autonomies
“The Spanish healthcare system is decentralized and competences are transferred to each autonomous community. Consequently, an adequate deployment requires that each autonomous community integrates its healthcare system with the application”. This was the conclusion reached by the authors of the Radar Covid pilot study, who warned that it was a “very difficult task”.a critical factor underpinning success” and that it had not been “validated in this experiment”. The two co-authors consulted agree that this has been the main problem.
The operation of Radar Covid is, a priori, simple. The cell phone remains active by emitting an anonymous and non-geolocated signal via Bluetooth technology. When there is a contact of more than 15 minutes with another phone, they memorize the signal of their contact. If the bearer of one of these phones confirms that he is covid-19 positive by PCR, the health center provides him with a 12-digit code that he will have to enter in the application. At that point, the application alerts all the handsets it has memorized as at-risk contacts. But the main funnel that the authors of the pilot study had already warned about occurred when it came to delivering the codes. Two years later, there is still no data on the codes that the autonomous communities have delivered to their positive contacts. In 12 of the 17 autonomous communities, the ratio of codes entered for confirmed cases is less than 0.5%, and in Extremadura, Comunitat Valenciana, Ceuta, Melilla or the Canary Islands itself it is directly zero. “That was a Christ, I sincerely believe that they did not take it too seriously”, laments Arenas.
“Each autonomous community had to integrate the tool in a different way into its own tracking system, and this required a dedication of human and technical resources that not all of them had at that time”, explains Álvarez-León, who adds that Radar Covid arrived when the second wave was raging. Alberto caught it at its peak, in October 2020. When he asked his family doctor, in Cadiz, for the code he had to enter in Radar Covid, he felt that he was being speaking Chinese. The doctor came out of the office to ask the staff if anyone knew where that code came from. They had all heard bells but did not know where the mass was.
“Haste and insufficient means always hinder the proper implementation of any novel tool.”
“Not all the communities were prepared, but there was a rush to make it available because the incidence was increasing again. Haste and insufficient means always hinder the correct implementation of any new tool”, laments Álvarez-León.
María, a woman from Madrid who tested positive with an antigen test one year later, in December 2021, contacted the telephone number provided by the Community to give notice. They never called her from her health center to do a confirmation PCR or to follow her up, so she was hardly going to be able to get the code to notify it in the application. A few days later, the Madrid Health Service changed the protocol and considered positives with a self-diagnostic test as true positives without the need for confirmation. By then, no one remembered Radar Covid any more and many of those who used it had even uninstalled it. And this despite the efforts of the Secretary of State for Digitalization and Artificial Intelligence, Carme Artigas, to continue defending its benefits: “It is not a failed project and I will not accept that it is said that the Secretariat has not done a good job”, she responded to the parliamentary groups in Congress in September last year. “Only if we had saved one life with Radar Covid I would be satisfied”, he added.
“It was not even heeded”
Arenas insists that the technology on which Radar Covid is based is very good, but the problem has been social: we are willing to install third-party applications on our phones to which we cede data that we do not even know, but when Radar Covid arrived, many doubts arose. “It was ignored, we didn’t even reach the minimums for it to have any representativeness.”, Arenas sentences, without any concealment.
“When a user did not find the desired response in their health service – they did not know how to get the notification code, they did not receive it quickly or they were not given access to a diagnostic test quickly – they generated more distrust of the application in their environment. And this impression that the application was useless spread and the population that initially installed it on their cell phones gradually uninstalled it”, says Álvarez-León. And this is confirmed by Diego, another of the users consulted by EL PERIÓDICO: “I ended up deleting it from my mobile because never warned me of any close contact and constantly asked me to update notifications, which I had already done”.
Sergio Carrasco is a lawyer with expertise in data protection and technology. On behalf of the NGO Rights International Spain, he denounced that Radar Covid did not comply with the guidelines set by the General Data Protection Regulation. He believes that there was a lack of willingness to make the project transparent from the beginning, with the government resisting to provide information. The process culminated in the sanction of a warning for failing to comply with eight articles of the data protection regulation. One of these pointed to the failure to carry out a timely data protection impact assessment on the pilot project, a mandatory document that sets out the potential risks to which personal data is exposed. “What should have been done is to reinforce the confidence of individuals precisely by giving access”,” he insists.
Another problem is that the application had a vulnerability in its design known to the secretary of state and was not remedied. It was possible, for example, to associate an IP with a positive test, and this could affect “all kinds of personal data”, according to the sanction. Carrasco insists that “the logical thing to do would have been to give access to the code from the beginning of programming.” so that vulnerabilities could be detected and fixed”.
“Your exposure is low” at the height of peak contagions.
Álvarez-León believes that there were points that were not “sufficiently studied”. For example, the distrust that could be generated in the population by a government application in terms of privacy or technical details that generated “negative user experiences”. As Brenda, who recalls how the application repeatedly pointed out to her “your exposure is low”, the message that appears automatically since it is downloaded, at the peak of contagion in the city where she lives.
Arenas insists that the idea of Radar Covid is to prevent waves, but when we are already in a peak of contagions “tracking does not work” and the application becomes meaningless. However, the two experts agree that what these applications offer is very useful and will be in the future, although “we still have years to go before it is socially consolidated,” as Arenas says. “An epidemic is not a medical problem, but a social problem: it is always determined by human behavior”, he adds. Carrasco believes that “we need to be more agile”, because “it is no use if at the end of the pandemic we have an application that works.”.
The government will let Radar Covid, whose maintenance contract expires in November of this year, die. In response to a parliamentary question, it insists that this type of application “showed its effectiveness in earlier phases of the pandemic”, but acknowledges that “are no longer essential at this point in time.”, although the data sow serious doubts that, at least Radar Covid, ever was.