For this year’s Digital Health Week, and in support of Transform Health Coalition’s campaign about digital and data governance, we hosted a panel discussion to explore country-level perspectives on data governance in the context of digital health. As an organization with many years of on-the-ground implementation experience in multiple countries, we wanted to complement the ongoing global-level policy conversations with a country-level view.
Last month, D-tree was fortunate to bring together a diverse panel of accomplished experts, representing views from the government, academic, international NGO, and private sectors from Zanzibar, Rwanda, Kenya and Guyana. Across this huge breadth of perspectives, one overarching theme still emerged, which was emphasizing that the point of data governance is to enable data to be used in impactful and responsible ways – in the form of routine decision-making, analysis and innovations such as AI (artificial intelligence).
However, even when data governance does enable data to be used responsibly, we heard how a lack of knowledge and skills for systematic data use can become a barrier for achieving the full potential of these tools in places like Zanzibar. Otherwise stated: while good data governance will enable data use, it is not always sufficient. There are additional requirements, such as building capacity to interpret and analyze data, that are necessary if we want data governance to generate health impact via data use.
We also heard an example of how data governance is an iterative process. Data governance frameworks are created and then implemented, but that is rarely the end of the story. Frameworks often need to be refined after they have been introduced, because that is the only time their effectiveness can be tested – in real life. Rwanda recently introduced its first Data Protection Law which provides strong safeguards for personal data, including giving individuals the right to decide how their data is processed. However, the team at the Centre for the Fourth Industrial Revolution (C4IR) in Rwanda found that this was making it very difficult for innovation research to be conducted using health data. This has resulted in the frustrating situation where there are Rwandan researchers who have the ability to create AI models that can be used to improve healthcare, but they cannot access the data to create those models. As a result, the C4IR team is working on extending Rwanda’s Data Protection Law to facilitate improved access to health data for researchers.
The example in Rwanda also highlights that effective data governance is a balance between allowing data to be used freely without restrictions, which can lead to privacy violations and exploitation amongst other harms, versus being overly protective of data to the extent that the data cannot be used in the ways that are most valuable. One of the speakers also pointed out that ‘balance’ is very much at the core of data governance, in terms of balancing what is ideal and what is pragmatic. This balancing act is what makes data governance challenging from a practical perspective – whilst what is ideal can easily be captured on paper, the ideal may not be possible or pragmatic in real life. Data governance should serve to define that balance, and enable accountability in the process, so that individuals and communities can call out governments, private sector actors or others, if the right balance is not respected.
We are very grateful to our four panelists for contributing their time and insights: Mohamed Habib Al-Mafazy, Head of ICT, Zanzibar MOH; Joris Cyizere, Acting Managing Director, Centre for the Fourth Industrial Revolution Rwanda; Karen Bett, Policy Manager, Data Equity and Inclusion, Global Partnership for Sustainable Development Data; and Dr. Jeb Weisman, Director of Global Health Informatics, Icahn School of Medicine at Mount Sinai. Finally, we also want to thank the Transform Health Coalition for providing the platform for the discussion during Digital Health Week.
In case you missed it, watch the panel discussion here.