Search for a report, a publication, an expert...
Institut Montaigne features a platform of Expressions dedicated to debate and current affairs. The platform provides a space for decryption and dialogue to encourage discussion and the emergence of new voices.

Digital Tracing or Repetitive Lockdown: French Citizens, You'll Have to Choose

Digital Tracing or Repetitive Lockdown: French Citizens, You'll Have to Choose
 François Godement
Special Advisor and Resident Senior Fellow - U.S. and Asia

Originally published in French on 7 April 2020
When the epidemic first appeared in China, and then in Asia, the Western response was to prioritize individual freedom, and in particular the freedom of movement over the health crisis. This was perhaps conceivable if it went with a variety of preventive measures - some of which are already considered to be a first step towards restricting individual freedoms. For example, the wearing of masks, often perceived as a sort of useless manifestation of an "Asian-style" collective obligation. Or the use of digital tracing, after some success in local tracing (for example La Balme in Haute-Savoie),  where going up the chain of contamination allowed to shut it down using much less elaborate means.
Digital tracing is becoming a burning issue as the project of a government-sponsored mobile application could well see the light of day in the next few hours or days. Beyond the tracing of travel flows by mobile operators or by Google, which is already a reality, tracing can have two functions: the backtracking of a mobile phone user's contacts, and therefore of the risks of infection incurred; and the real-time tracking of movements to monitor the enforcement of lockdowns - or compliance with limits and exceptions set for the end of confinement.
One application seems to fit the bill: TraceTogether, a creation of the Singaporean government run by its Ministry of Health. The code is available in open source, thus usable by the whole world. The version of the app used in Singapore keeps user data on the phone rather than on a server, with login details assigned anonymously and regularly changed. Contact detection only works with the latest versions of Bluetooth. It determines the user's contacts (but not their location) with another user of the application with an accuracy of about 2 meters, and will later warn them of a risk that was incurred. At this point, the transmission of stored data becomes mandatory. There is still preservation of anonymity of contacts - while health authorities acquire a nominal view of these. Retroactive tracking can reveal a suspicion of infection but cannot confirm it: a medical test remains necessary. The introduction of TraceTogether however dates only from March 21, while an increase in cases of local contamination happened in Singapore on March 18, thus before the application could prove its worth.

Like all digital applications, TraceTogether comes with threats to privacy: although supposedly non-intrusive, it ends up identifying the entire population if two conditions are met: a large-scale download of the application, and an increase in potential infections, leading to the transmission of more and more contact data. Then the usual problems arise: trust in the system operators, the time limit for data retention (set at 21 days in Singapore), and loopholes in the privacy settings that maintain a separation between contacts. In addition, in France, there is a large percentage of mobile phones (25%) which do not have sufficient Bluetooth capability for effective detection.

A rapid identification of emerging or re-emerging chains of contamination is imperative. The problem of ending  the quarantine and resuming normal life is therefore inseparable from the issue of digital tracing.

These threats raise the question of adequate supervision. But is the threat any greater than that which we face today? There are two reasons to doubt it. The question of the means available - masks, tests, data and digital tools - has largely pre-empted the actual choices in the rising phase of the crisis. As a result, we are undergoing one of the most rigorous lockdowns in the world (after China), and thus an almost total amputation of our freedom of movement. As of March 29, Google's mobility data for France show a drop by 88% for shopping, 87% for public transport stops, and 57% for workplaces.

We are now approaching a flattening of the curve, with the prospect of a decrease that does not, however, mean the end of the epidemic. The endogenous infectious resurgences that are appearing in Asia show us, as many virologists believe, that there may be other waves, of unpredictable magnitude, before the arrival of vaccines that are themselves subject to the possible mutations of the virus. One can hardly fathom the economic and societal damages that a second large-scale lockdown within the next 18 months would cause. To avoid this, a rapid identification of emerging or re-emerging chains of contamination is imperative. The problem of ending  the quarantine and resuming normal life is therefore inseparable from the issue of digital tracing. It is not the only instrument that can be used, but we must stop the "all or nothing" discourse which seems to have repeatedly prevented the adoption of certain solutions.
Our mistrust towards data collection coexists with the daily acceptance of countless blows against the protection of our personal data. The problem exists - both with regards to how data is collected and how it is used. It is regrettable, for example, that the 2015 French Law on Intelligence, adopted under pressure just after the terrorist attacks, left judicial supervision of these mechanisms aside.
But can the quest for absolute protection of personal data justify the neglect of digital tracing tools? It is illusory to protect ourselves from contamination by others if these cannot also be protected from ourselves. The dilemma also exists with surgical masks: it is by protecting others that we protect ourselves. This is the essence of tracing: by accepting that my close contacts (within a margin that must be defined) are stored for identification in case of suspected contamination, I can know with some probability whether I have been in direct contact with the epidemic.
In France, there is a second obstacle: the administrative compartmentalization that hinders the use of digital data. This is particularly the case in the field of personal health, healthcare and medical research. The taboo surrounding personal health has led to a very low digital harmonisation of medical research. This is so true that, in order to optimize the workload of hospitals, there are now discussions about having their data managed by Palantir, the American digital data giant. Decades after its creation, the Shared Medical Record is still rarely used, as a variety of social insurance systems do not have a clear common computer base. Hospitals, doctors and health organizations use common grids mostly for reimbursements.
This should be of little importance in the case of a major epidemic, since it is not complex medical treatment data that are exchanged, but binary information on the infection (yes/no...) and contact identification data that is epidemiological rather than health-related. But the healthcare sector would nonetheless represent a more neutral actor in epidemiology than the Ministry of the Interior! In Israel, tracing was first entrusted to internal security, and then passed on to the Ministry of Health with an open source code application. This is also the case in Singapore.

With considerable variation, East Asian countries and a few others (notably Israel) have implemented one or more forms of digital tracing. In addition to the retroactive tracing of chains of contagion, real-time tracking of movements allows the effective isolation of a patient in the contagion phase, and the enforcement of partial lockdown. In France we are faced with a paradox: an unprecedented travel form has been set up, while thousands of police officers are in charge of controlling (and recording) it. Its efficiency is questionable. Are slips through the cracks the price to pay for not adopting digital tools?

The digital applications that France will adopt must be open enough from the outset to be suitable for visitors, starting with international carriers.

The modalities and the safeguards of both tracing and tracking differ among the different East Asian countries. Let us briefly review them:

  • As of today Japan is the most cautious, limiting tracking to anonymized data from telephone operators, and only for patients who are actually infected, in order to identify potential clusters. In addition, users can self-report their symptoms on a public health application. But Japan imposes isolated hospitalization for anyone presenting a symptom.
  • In contrast, South Korea and Hong Kong practice the most intrusive and coercive digital tracking. In Korea, tracing first targeted members of a religious sect that had travelled to Wuhan, and was followed by the mandatory tracking of all movements of its members. The obligation to track, without consent, is extended to all the people with, or likely to have, symptoms, based on a combination of telephone and location data; the use of credit cards; and video surveillance (in particular for the use of masks), all without judicial intervention. People under quarantine are tracked by an application managed by the Ministry of the Interior. In Hong Kong, the mandatory tracing of visitors and of people with symptoms is now accompanied by the mandatory wearing of electronic bracelets for arrivals and persons in quarantine.
  • Between these two options, Singapore and Taiwan represent a middle ground. In Taiwan, data on border arrivals are cross-referenced with data from medical consultations to identify potential patients. People in quarantine are equipped with telephones to monitor their location, and they are even publicly identified in the event of breaches of the quarantine.

None of the above is pleasant. But with the exception of Singapore (which will pose the problem of source code verification), the other countries mentioned are all democracies. None of these countries has had to adopt a total lockdown - but all of them understand the quarantine of certain categories of individuals (patients, newcomers...) as total isolation, and not a simple retreat into a family space that is very vulnerable to contamination. All of them, even Korea, which was hit by the mass accident of a religious sect, have much lower human tolls than ours. All of them have retained a large part of their economic activities, as Google's mobility data shows. We leave the Chinese case aside, as many problems remain and facts themselves are not sufficiently verified.
To sum up for our fellow citizens: we have completely lost our freedom of movement, a sacrifice that is essential if we are to break the epidemic curve. The economic price is huge. We know that an unconditional exit from the lockdown is a myth, and that many intermediate measures must be taken. We are faced with a situation that is both exceptional and of finite duration, even if it remains impossible to specify it.
As such, the step by step exit from the lockdown cannot afford to do without tracing and tracking instruments. The government is tackling this issue with a bit of a head start, and that is a good thing. For solutions to be both effective and accepted, there are several requirements: first, as soon as possible, a separate management from that of the Ministry of the Interior, and therefore an upgrading of the capacities of the Ministry of Health. Secondly, open and verifiable source codes, as well as guarantees on the length of conservation of data and its supervision afterwards (because urgency is presently uponus). To this end, citizens’ trust is crucial. Equipping the French people with mobile phones of a sufficient standard must be facilitated, in order to reach the critical mass of users, which epidemiologists estimate at two-thirds of the population. For the same reason, it is counterproductive, even if politically very tempting, to limit ourselves to a voluntary use of these instruments. This is the price to pay for the health of French citizens.
Ideally, European standards and norms - especially in the Schengen area - should be adopted quickly. In the short term, this amounts to mixing up totally different situations in terms of epidemics (Italy: 16,000 deaths; Denmark: 187), and thus delaying the whole process. And yet this same disparity is found within countries themselves. Asia combines other measures with a generalized closure of borders for people, which is not sustainable anywhere in the long run. The digital applications that France will adopt must be open enough from the outset to be suitable for visitors, starting with international carriers. European harmonization must be the next step.


Copyright : Odd ANDERSEN / AFP

Receive Institut Montaigne’s monthly newsletter in English