Universiteit Leiden

nl en

Fighting corona starts with sharing data responsibly

Gathering and distributing patient data can make an important contribution to containing the coronavirus. But if we want to be successful, we need better data. With this objective in mind, Leiden data stewards have joined the Virus Outbreak Data Network (VODAN).

What are the main symptoms of the coronavirus? Which groups are most at risk from the virus? And how long, on average, is a patient in intensive care? The answers to these questions have been filtering out from hospitals all over the world in recent days. These are crucial questions because they determine such things as who needs to be in quarantine and whether the hospitals have enough beds. 

Kristina Hettne (sitting, left) with her colleagues from the Centre for Digital Scholarship.

The recently set up Virus Outbreak Data Network (VODAN) wants to professionalise the collection and dissemination of data around the coronavirus and other epidemics, so that these kinds of important questions can be answered faster and better. VODAN is a partnership that has emerged from GO FAIR, a Leiden intiative that aims to make data in the broad sense more accessible for scientific research. At Leiden University, the LUMC and the Centre for Digital Scholarship of Leiden University Libraries (UBL) are partners in this international initiative.

‘Hospitals across the whole world are currently collecting patient data about the coronavirus, for example about the symptoms, treatment and course of the illness,' says Kristina Hettne, who works at the Centre for Digital Scholarship. ‘With VODAN we want to make sure that this data can be shared effectively so that others can learn from it. It is really important to share the data rapidly and efficiently. ’

The basic principle is that the data should be FAIR: findable, accessible, interoperable and reusable. This way, data can be used and compared by researchers and doctors everywhere in the world, and can even be used in artificial intelligence software or machine learning. Right now, much of the data does not meet these FAIR principles, as was apparent during the Ebola epidemic (see box).

Ebola: how not to do it

Just how much the VODAN initiative is needed is apparent from the Ebola outbreaks that ravaged Africa in recent years. Data about these outbreaks is difficult to find and to open, and it is absolutely not possible to compare it or reuse it. In the case of Ebola, it is ironic that the data is least accessible for users who were most affected by the epidemic.

‘It is also crucial that all the data is stored in the country of origin,' says LUMC Professor Barend Mons, who heads up VODAN. ‘In this time of crisis there is a huge call to store all hospital data centrally. But that's not realistic. For geopolitical reasons, China is highly unlikely to hand over its data to the US, and vice versa. Not only that, the privacy laws would prevent that happening.' 

Barend Mons

VODAN has to offer an alternative, which could consist of 'data stations' in all the participating countries, from where the data can be shared securely. Using algorithms you can have a kind of 'train' that passes through all the stations, gathering the data. That way the participating countries would each retain complete control over their data, but data exchange is still possible. 

‘If all the parties work together, we could, for example, examine why men become much more sick than women,' Mons gives as an example. ' Is that because of their testosterone level? And if so, what's the situation for men who for medical reasons were taking testosterone suppressants before they became sick? You only need to write an algorithm and you can then call up these data in the participating countries.’

The coming weeks and months will be crucial for the FAIR principle. The big data organisations and countless individual researchers have expressed solidarity with VODAN, but we still have to wait to see whether that's enough. Will this be the definitive breakthrough for FAIR? Or, in this period of crisis, will we fall back on old reflexes? 'One thing is certain,' says Mons. 'It won't matter a jot for the algorithm that the data isn't all stored in one place.' 

Text: Merijn van Nuland
Mail the editors

This website uses cookies.  More information.