“Let’s use the telecom data of all Belgians to stop the coronavirus”

We regret that Belgium does not apply digital practices proven elsewhere in the era of the Covid-19. We could speed up decision-making, make it more precise or dynamic and, above all, prepare Belgium’s toolbox for the next global pandemic, which may concern a much more lethal disease than Covid-19.

For 8 years, we have led the fight for “Data for Good”: raise awareness the general public and governments the value of public or private data and their potential for the public good. Our vector of action is entrepreneurship where we seek to bring together technical teams, data holders, public institutions and actors in the field around well-targeted issues.

Sebastien Deletaille and Frederic Pivetta

The first is an entrepreneur in the field of technology and health; the second is eBig Data / AI entrepreneur for societal purposes and Managing Partner at Dalberg Data Insight

In 2014, we had the opportunity to help the United Nations to fight Ebola with anonymized telecom data. Indeed, Unicef ​​teams relied on our models to calculate population mobility, predict geographic spread and estimate the effectiveness of quarantine areas. This work has been awarded the MIT Innovators Under 35 Europe award.

Since 2015, we have deepened our expertise in fighting malaria in Africa or Zika in Brazil; with telecom, private and public data; with new partners including Telefonica, Orange, the Gates Foundation, ministries of health or USAID; to prepare the health system for the next epidemic or pandemic.

Armed with these experiences, we regret that Belgium does not apply the same digital practices in the era of the Covid-19. We could speed up decision-making, make it more precise or dynamic and, above all, prepare Belgium’s toolbox for the next global pandemic, which may concern a much more lethal disease than Covid-19.

Ebola and coronavirus, two examples of viral epidemics

Ebola and the coronavirus are obviously two separate infectious diseases. From a statistical point of view, they share a high transmission rate: an infected person will transmit the disease on average to 2.5 other people (these rates change over time and space). Fortunately, diseases differ in mortality rates. WhileEbola has a fatality rate of around 60%, the coronavirus is between 1 and 3%, depending on age groups.

To fight an epidemic, public health institutions produce epidemiological models and action plans based on a series of parameters such as the risk of exposure or spread of the disease. The objective is to adapt the reaction to the context: a ticket clerk at the station must be made aware differently than a person who works at home, control at Zaventem airport is more important than that at Hurpes station ( Hainaut), ditto for a traveler returning from a risk zone. To have targeted plans, authorities need granular and dynamic data and models, which cover not only travel but also human contact.

An effective preventive approach requires that the focus is on people who move around a lot, in populated places, and near or in connection with groups of at-risk populations (for example, nursing homes).

We must have an ambitiously modern management of the crisis and take advantage of it to prepare for the future.

Lwhen a person is infected, the priority goes first to raising awareness of the “social graph”, namely the persons with whom the patient has been in physical contact since his infection; then raising awareness of the areas connected with the places this person has frequented. We are interested not only in the main home and the workplace but also in the “third places” where the patient likes to spend time (for example at the local cafe, the Saturday bookstore …).

Without control, all epidemiological models predict an impressive number of deaths. It is therefore essential to limit the spread.

How to do? What have we learned?

What have we learned in Africa and Latin America? Use telecom data and involve the private sector.

The basic idea is simple. Infectious diseases (including Ebola, Covid-19) are transmitted from person to person and spread through movement and physical contact. Hence the interest of telecom data: with enough history, the call and geolocation registers make it possible to calculate the social graph and the maps of real movements. These calculations are performed without knowing an individual’s name and can be grouped into “mobility hubs”.

By superimposing the maps thus obtained with epidemiological data (for example, known cases and foci of infection), you identify the places and people where you must launch your preventive actions. In the case of Ebola, this has slowed the spread of the disease and quantified the impact of quarantine measures.

To be successful, several skills are required: managing large telecom data flows, developing algorithms, developing agreements and processes allowing privacy to be respected while providing key indications at the epidemiological level, perpetuating these platforms with local institutions.

Our call to political action in Belgium

Unlike China or the African continent, Belgium has not often been exposed to epidemics. We have only known epidemics of influenza for which we have an impressive medical infrastructure and a great discipline of preventive vaccination. Today, we have a hundred cases of Covid-19 in Belgium. There is still time to take action and take preventive action. When we have more than 10,000 cases, the targeted approach will give way to mass action, like quarantine zones in Lombardy.

Are we ready to respond to an out of the ordinary crisis? Are we ready to learn from the investments that Western countries, including Belgium, have made in Africa or Latin America?

We believe thatwe must go much further and not miss the digital opportunity. We must have an ambitiously modern management of the crisis and take advantage of it to prepare for the future. Covid-19 is a serious epidemic, but its mortality rate is relatively low compared to other viruses. Belgium can and must be innovative. She is small, highly digital and has the means in terms of infrastructure and knowledge.

Concretely, an ecosystem like AI4Good, from AI4Belgium, could work with government for:

  1. Use telecoms data from all Belgians for the past three months and ensure full compliance with the General Data Protection Regulations (GDPR), for example, under the supervision of members of the privacy committee.
  2. Identify the people posing the greatest risks amplification of the epidemic (for example, people with high mobility or who have a large social network). Contact these people by SMS and provide them with prevention advice.
  3. Identify the places posing the greatest risks amplification of the epidemic or the deadliest risks.

In addition, the GDPR provides a regulatory framework (Article 9 (2)) where the use of personal data is authorized in the interest of public health.

With the Covid-19, we have a duty to prepare the toolbox that will be used in a much more serious crisis. As the mortality rate is lower, we have the opportunity to show the Belgian public that we have the data, the algorithms, the communication and the action plans to react. It’s not about spreading fear, it’s about showing that we take public health seriously. The data exists, the regulatory framework too, all that is missing is political action.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.