COVID-19: Artificial Intelligence To Back Up Data-Driven Response In Africa

By Reinhard Jacobsen

The epicentre of the devastating COVID-19 pandemic has moved from China to Europe and the United States. But it has not spared Africa. “It will victimise particularly communities living in poverty who have no protection and no access to health care,” says Prof. Dr. Mirjam van Reisen who chairs ‘Computing for Society’ at the Leiden University.

The public research university in the Netherlands has joined universities in Uganda, Kenya, Tanzania, Ethiopia, Nigeria and Zimbabwe in the Virus Outbreak Data Africa Network (VODAN) FAIR Data initiative.

Computer scientists across the continent collaborate to establish connected data to strengthen analytics to fight COVID-19. “With this project we can analyse the data to understand where and how the virus moves and assist governments to reach out to protect such communities,” adds Van Reisen. She believes the VODAN FAIR Data initiative is a crucial and necessary condition for other data initiatives.

FAIR stands for human and machine-readable digital data objects that are Findable, Accessible, Interoperable and Reusable. Accessibility is fully compliant with regulatory and governance frameworks, including personal data protection and ethical considerations based on the principle to Do no Harm.

Van Reisen noted that there are – understandably – many data initiatives in response to COVID-19. However, the aim should be to create a future-oriented infrastructure and capacities.

The European Commission, for example, has a great interest in ensuring that:

– data for science and other data is interoperable following the same guidelines. FAIR is the recognised guideline for the whole of the European Open Science Cloud and will generate 5% of all available funding for data stewardship. It makes sense to make African data initiatives convergent with this;

– the critical issue of data belonging to the data subject and stewarded recognising provenance of the data is critical to adhere to basic data relevant ethical principles but also to avoid littering the data or inclusion of fake data, which without pointing to origin cannot be managed. FAIR is the key solution to this;

– ensuring that data remain in jurisdiction of where data is produced is critical for equality in a world where data is increasingly the raw material for economic services, including in the sector of health care.

Van Reisen added that “the European Union already has leadership in this; it would therefore be very important that the European Commission aligns COVID-19 responses to this framework already established and strengthen it with an adequate response to the current crisis”.

“I am keen to ensure that something good emerges from this crisis; and the data architecture is key to the what the future will look like,” Van Reisen maintained.

The VODAN FAIR Data initiative is supported by the Philips Foundation, whose mission is to provide access to care for underserved communities. “With this donation, the Philips Foundation supports the establishment of a Virus Outbreak Data Network… to ensure that data-intelligence is available to support an effective response to the new corona virus,” said Margot Cooijmans, the Philips Foundation Director.

The Philips Foundation aims at reducing healthcare inequality by providing access to quality healthcare for disadvantaged communities through innovation, reflecting our commitment to United Nations Sustainable Development Goals 3 (Ensure healthy lives and promote well-being for all at all ages) and 17 (Revitalize the global partnership for sustainable development).

Another partner of the VODAN FAIR Data initiative is the Kampala International University (KIU). KIU Vice-Chancellor Dr. Mohammed Mpezamihigo said: “There is a fear for undetected spreading on the African continent which may lead to a sudden crisis. To help avoid this, both prevention and containment are critical. As one of the leading universities with a large medical campus and university hospital, we are committed to lead the initiative to build adequate capacity.”

KIU has already deployed ten computer and data scientists to work with medical experts. “As an international university with students from 42 countries, we are committed to make sure that all our efforts strengthen the global fight of the corona virus,” added Dr. Mpezamihigo.

KIU is a private, not-for-profit institution based in Uganda. The University which started in 2001, as a typical degree-awarding institution has now grown into the leading Private University in Uganda and is currently ranked as the fourth-best out of 65 universities in the country.

“Uganda is learning from the strategies that worked elsewhere to optimize the approach to protect people from COVID infection,” said Dr. Diana Atwine, Permanent Secretary in Uganda’s Ministry of Health. “Science will help us and our universities are helping the government.”

Data are critical. Dr. Atwine has asked universities to build the Uganda national COVID FAIR Data-point. This will be the first FAIR Data-point in which Data held in Uganda can be analysed with scientists globally. “This Data is ours, they will stay in our country, and those who visit our data will do so according to the laws and regulations in our country. This is the backbone of our knowledge that will help us navigate this crisis.”

Making data and metadata FAIR ensures that these data are discoverable on the Internet of FAIR Data and Services. Central to this approach is the establishment of FAIR Data Points (FDPs), for COVID-19 relevant digital data objects.

Opening up FAIR (meta)data by publishing them on a FDP allows algorithms to search these (meta)data, looking for patterns. The Internet of FAIR Data and Services is a distributed data discovery network; data are NOT moved, but algorithms going over the internet can find the data.

In a data-driven response, backed by Artificial Intelligence (AI), such as is required for COVID-19, it is of cardinal importance that Africa participates and is in the driver’s seat, the knowledge and capacity in Africa to build lasting architecture;

GO FAIR Foundation (GFF) is yet another VODAN partner. GFF was established in February 2018 as a separate legal entity under Dutch law in order to support the FAIR principles and metrics, and the GO FAIR International Support & Coordination Office (GFISCO).

VODAN is also supported by the Data Together: GO FAIR, CODATA, Research Data Alliance and World Data System. The first phase VODAN deliverables are focused on establishing a COVID medical data FAIR Data Point, reusing the global standards from the WHO, the World Health Organization, and the local laws and regulations of the jurisdictions where the FAIR Data Points are geographically located.

Simultaneously to the first FAIR Data Point which will be deployed in Uganda, other COVID data (biomedical research data, economic indicators, etc) will be prepared by FAIR data stewards, making them machine-ready.

Coordinated developers will work to co-develop and extend key FAIR functionalities of the FAIR Data Points (indexing, search and automated access control) and also provide easy front end solutions that make it easy for non-technical personnel (e.g., citizens, medical doctors) to create FAIR Data and deploy then in compliance with national regulatory frameworks, including personal (health) data protection.

The next phase will be focused on expansion, demonstrating federated queries across an emerging global network of FAIR Data Points including in China, Europe, the U.S. and Brazil. Unless permissions are explicitly requested and given, the data remain fully within the jurisdiction of national governance and regulatory systems and queries will be certified by the GO FAIR office. In this expansion phase there can be focus on deployment servicing other languages through interfaces.

Planned is also the Training of Trainers (ToT) to provide support to data stewards so that they can help combat the COVID-19 pandemic.

ToT will also help prepare ministries and institutions for publishing FAIR (meta)data; ensure that universities and institutions with medical data have capacities to assist the navigation of the crisis with quality data; enable data to be analysed in combination with data available globally; and ensure quality data for AI solutions to help deal with the pandemic.