Learning from African health data without relocating it

FAIR data

FAIR stands for Findable, Accessible, Interoperable and Reusable for both humans and machines. So the data is findable, accessible, communicates with each other and is reusable. The data can be used by scientists to map disease states and distribution, for example. FAIR data is central to the VODAN network, which collaborates with data professionals from several African countries. Professor of FAIR Data Science Mirjam van Reisen is the international coordinator.

Data sovereignty

Van Reisen: “Data are very valuable. It is therefore important for countries to keep ownership of the original data and not shift ownership to a central location in another country. That's what we call data sovereignty. Countries around the world are working to ensure this, including African countries.” The VODAN network (Virus Outbreak Data Network) aims to learn more about virus outbreaks according to the FAIR Data principle, such as the cause and spread. Partners in the network can request data for research, generalizing the data. Data that can lead to identification of a patient are not used and remain safe in the clinics where they are patients.

An example: patient A is infected with the omikron variant of the covid 19 virus, is between the ages of 25 and 35 and is from country Y. Patient B does not have the virus, is also between the ages of 25 and 35 and is from the same country. Researchers then get the data, for example: in country Y, so many people between the ages of 25 and 35 have the omikron variant and so many do not.

Professionals from African countries take apart central systems and turn them into mini-services that are installed in connected health centers. These digital services are connected and can communicate with each other. “Compare it to video calling: you all sit in your own place, but communicate with each other through certain software. By not centralizing the data, the mini-services also require less energy per service. As a result, data from areas with less digital coverage can also be used.”, says Van Reisen.

In the case of the VODAN network, professionals from countries of member partners are trained on various aspects of FAIR Data, which was born in Leiden. Van Reisen: “For example, there are data clerks who enter the data. It's important that that data is accurate because it forms the basis. Data stewards then process that data into mini-algorithms, which together with complex algorithms can extract new findings from the data. This involves looking at techniques to prevent patient identification: this is the core of responsible AI. This is being investigated by PhD researchers and their supervisors, both from Leiden, Tilburg and affiliated African countries.”

Learning from each other

The knowledge, experience and data of African countries are very valuable to learn from. Researchers within the network can use the FAIR Data for vaccine research, but also to map and track disease patterns. Consider the example of omikron variant or mpox virus. Diseases that Africa is currently facing can come our way, due to the connected world or climate change, for example. “African colleagues and data can help us better understand where diseases come from, what the causes are and what the differences are in side effects of drugs or vaccines here or there. ”, says Van Reisen, “This is because it is not necessarily the case that side effects are the same all over the world. That's why it's very important to have that data in different places. Also, sometimes professionals don't enter all the data now because they know it goes to a central place. If the data stays in their own clinic, it helps with confidence that the data is being used ethically.”

FAIR data at the center of training

The VODAN network started in 2020 to fight corona with data, but developments are now moving fast. “Our goal is for FAIR data to become the foundation for universities. It should become the standard in all undergraduate studies, not only of data scientists, but more broadly. If you do an undergraduate or a master's degree, and certainly a PhD, you need to know how to process data in such a way that that data is reusable. And not only in LUMC, but also within African universities. That's why deans from different universities are also involved. At the end of August, we officially established the collaboration in The African University Network on FAIR Open Science. Here the president of the Pan African Parliament was also present as the chairman of this network. The trust that those data are used according to the right ethical formulas and the ownership remains with the original owner must be anchored both politically and scientifically.”