Primeur Magazinein this video is focusing on the session about COVID-19 research data. There were four speakers and one moderator. Helen Johnson from the European Centre for Disease Prevention and Control (ECDC) was the first speaker. ECDC is an overarching centre looking at infectious diseases, connecting data, and giving advice on a European scale. ECDC is overarching the national centres. Each country in Europe has such a national centre. In Belgium, it is called Sciensano; in the Netherlands it is called RIVM; Germany has the Koch Institute. ECDC is responsible for the overarching collaborations.
Helen Johnson talked about modelling COVID-19 from data to forecasts. She was looking especially at the possibility of forecasting: What will happen? What will be the evolution of the pandemic in the coming 30 days? ECDC has built a Monte-Carlo based model that is supported by lots of data. Essentially what the researchers want to get out of it is the evolution of the number of confirmed cases, the number of deaths, etc. for the projection of the coming 30 days. The researchers use a large amount of data to feed the model.
To build the model, ECDC uses lots of existing knowledge from scientific literature. The researchers also need to have current information, as up-to-date as possible. This information is often difficult to get since there is no system to automatically collect the information. Sometimes, it even takes too long before information from the national centres is delivered to ECDC. Therefore, the researchers scrape the data from the Internet and later on, ask the national centres to validate the information.
The researchers try to represent the response measures taken by the different countries. Countries can decide to introduce a lockdown, to forbid the gathering of large groups for a certain time, for example. ECDC also tries to model this information. The model is in use already for quite some time but it is evolving further and further. The latest item that was added was a rather crude model of a face mask. Basically, one assumes that is has the same effect wherever you use it.
ECDC also has to model the behaviour of the people and how well they are responding to the measures. This is compared with Google data about mobile phone use. The model also needs to be calibrated. It has to be as accurate as possible. This means that the most recent data on daily confirmed COVID-19 cases and daily deaths are inserted into the model. The calibration is hierarchical. The researchers calibrate the whole model. Data that comes in from one country influences the quality of the model as a whole. The model is improving as more data are being inserted and calibrated.
As for the results, Helen Johnson gave examples for two countries. One country shows a peak and then a slow rise again. Another country shows a more stable evolution. The last 30 days show the predictions. The measures that have been taken are also shown to visualize what the length of a lockdown can mean in terms of causing a positive effect on the COVID-19 spread. How good is the model? It is working but the latest data and reports date from mid September. There was a very high peak in most countries in Europe in the week that followed. For some countries, this was modelled very nicely. For other countries, like for instance the Netherlands, the result was rather poor. However, ECDC is working on the model and it will improve overtime, so that the predictions will become better and better.
Another aspect is the modelling of data used by researchers. There are researchers who are working on vaccines, and others are working on the spread of the disease. A lot of research is done around COVID-19. The scientific researchers need data and tools. One of the organisations in Europe that is looking after this is ELIXIR, a research infrastructure for life sciences. Niklas Blomberg, ELIXIR Director, gave a presentation on this topic. He explained that a research infrastructure is a collection of tools like data bases, computational power, HPC and supercomputers, models, data from measurements, and data from collaborations. The idea is that the research infrastructure provides support for researchers in that area. In this way, researchers can concentrate on the scientific models they are producing. The underlying technology is handled by the research infrastructure.
ELIXIR is situated in the life sciences. It is not only focused on COVID-19 but their research ranges from biodiversity to genome research. There are 23 countries that are member and participating in ELIXIR. Hundreds of institutes around Europe are working together, sharing data and tools among researchers. For COVID-19, a European Data Portal has been set up. ELIXIR is working together with another international governmental organisation called the European Bioinformatics Institute (EMBL-EBI).
There are already a lot of data and resources available but they are scattered and disconnected. The idea is to collect them all in the COVID-19 Data Platform in order to combine the data with the tools. People networking is also important because for some research people need to work together who did not work together before. Since traveling currently is difficult, it is essential to find a good way to have researchers network with each other. After a few months, the network has been well established as a trusted research infrastructure.
ELIXIR is one of the main players in the network. Other projects and centres are also supporting the research. For instance, CINECA provides support with a supercomputing project that is used by researchers working in the COVID-19 field. The goal is to create a European Health Data Space so this is not a one-time effort. The European Health Data Space will become an infrastructure where all the research data are available for European health. This will be implemented in Horizon Europe, the next Framework Programme.