Navigating Data Challenges in African Health Research

In the realm of modern health research, particularly across diverse African populations, the ability to share and integrate health data ethically has become paramount. Researchers at the University of the Witwatersrand (Wits) in South Africa have stepped up to address this critical need with a pioneering model for responsible collaboration in multi-site studies. This approach, born out of the DS-I Africa initiative, tackles the complexities of combining genomic, demographic, and clinical data from multiple locations while upholding participant rights and legal standards.

The model emerges from real-world hurdles faced by scientists working on multimorbidity— the coexistence of multiple chronic conditions—using data from rural Bushbuckridge in South Africa and urban Nairobi in Kenya. As health challenges like diabetes, hypertension, and cardiovascular diseases rise across the continent, such collaborative efforts promise breakthroughs, but only if data handling is done responsibly.

DS-I Africa and the MADIVA Research Hub at Wits

The Data Science for Africa (DS-I Africa) programme, funded by the US National Institutes of Health's Fogarty International Center, positions Wits as a leader in leveraging data science for health improvements. Within this, the Multimorbidity in Africa: Digital Innovation, Visualisation, and Application (MADIVA) hub, co-led by Professor Michèle Ramsay and Professor Scott Hazelhurst, exemplifies Wits' commitment to innovative higher education research.

MADIVA draws on the Africa Wits-INDEPTH Partnership for Genomic Studies, incorporating longitudinal data from Health and Demographic Surveillance Systems (HDSS). This includes over decades of records from the Agincourt site in Mpumalanga Province and Nairobi, enabling machine learning to identify multimorbidity patterns. For South African universities, this underscores the value of interdisciplinary teams blending genetics, bioinformatics, and law.

• Genomic data from community cohorts
• Demographic surveillance tracking vital events
• Clinical records on chronic disease prevalence

Key Challenges in Multi-Site Health Data Collaboration

Multi-site studies amplify benefits but introduce hurdles. Professor Ramsay notes, "This ended up being far more complex than anybody anticipated." Primary issues include ensuring ethics approvals align with informed consent, navigating cross-border legalities, and negotiating data access with originating research groups.

In Africa, where data infrastructure varies, additional barriers like limited storage, fears of exploitation, and stigmatization risks loom large. The Wits team identified that combining datasets from South Africa and Kenya required harmonizing diverse consent forms, de-identification protocols, and governance structures to prevent misuse while enabling open science.

For higher education institutions in South Africa, these challenges highlight the need for robust training in data stewardship, as seen in Wits' bioinformatics programmes. Explore research jobs at universities driving such innovations.

Components of the Wits Ethical Data Sharing Model

The model, detailed in the paper "Balancing protection of participants and other stakeholders with openness: African lessons from the MADIVA data sharing and access policy," outlines a structured framework. Developed with input from PhD law student Daphine Tinashe Nyachowe, it balances three pillars: participant protection, open science promotion, and institutional safeguards.

• Controlled Access: Data requests reviewed by a governance committee, requiring ethical approvals and data use agreements.
• Temporary Embargoes: Local researchers get priority for publications, typically 12-24 months.
• Security Measures: Anonymization, encryption, and audit trails to ensure confidentiality.
• Equity Focus: Prioritizing African-led analyses and capacity building.

This practical blueprint ensures ethical health data sharing supports multi-site collaboration without compromising trust.

Photo by Daria Nepriakhina 🇺🇦 on Unsplash

Legal and Ethical Foundations Underpinning the Model

Rooted in South Africa's Protection of Personal Information Act (POPIA) and international standards like the GA4GH Framework, the model integrates legal expertise early. Nyachowe's thesis work examined cross-border transfers, emphasizing dynamic consent models where participants can specify future uses.

Ethical review boards at Wits played a pivotal role, aligning the policy with national guidelines. This addresses historical concerns in African research, such as neo-colonial data extraction, by mandating benefit-sharing discussions. For South African colleges and universities, it sets a precedent for embedding ethics in data science curricula. Check career advice for roles in this growing field.

Real-World Application: Integrating Agincourt and Nairobi Data

A prime example is MADIVA's analysis revealing common multimorbidity clusters: cardiovascular-metabolic diseases dominate, with risk factors like age, female sex, and low socioeconomic status. Machine learning stratified data automatically, uncovering transferable high-risk patterns across sites.

From 232 reviewed publications (2010-2022), gaps emerged—over-reliance on diaspora data and limited translational research. The model facilitated secure integration, yielding insights like early diabetes predictors, vital for South Africa's National Health Insurance rollout.

Implications for South African Higher Education

Wits' model positions South African universities as global leaders in ethical data science. It fosters interdisciplinary training, with bioinformatics courses producing experts in responsible data handling. As funding like DS-I Africa emphasizes African solutions, institutions must invest in data platforms.

This aligns with national priorities, enhancing research output and attracting international partnerships. For academics eyeing jobs in South Africa, Wits exemplifies how higher ed drives public health advances.

Insights from Wits Researchers

Professor Ramsay emphasizes participant generosity: "The data comes from people who have generously given their samples and data, so managing that responsibly is interesting yet difficult." Hazelhurst highlights legal integration: "about understanding how to share data from a legal perspective and an ethical perspective."

Nyachowe's contribution underscores the role of law in science, a growing niche for South African graduates. These voices reflect Wits' culture of ethical innovation.

Photo by Wulan Sari on Unsplash

Future Outlook: Scaling the Model Across Africa

Upcoming MADIVA outputs include machine learning papers in BMJ and Nature, predicting diabetes progression. DS-I Africa calls for extended funding, as "science takes time." The Wits Open Data Vault launch supports this by advancing transparent sharing.

Broader adoption could transform African health research, reducing silos and accelerating discoveries. South African universities should adopt similar policies, training via research assistant roles.

Actionable Steps for Researchers and Institutions

To implement similar models:

• Form interdisciplinary teams early (bioinformatics, ethics, law).
• Develop site-specific data policies with embargoes.
• Invest in secure platforms like federated learning.
• Prioritize community engagement for trust-building.

For career growth, visit higher ed career advice, rate my professor, or explore higher ed jobs and university jobs. Wits' blueprint offers a roadmap for ethical health data sharing worldwide.