Wits MADIVA Ethical Health Data Sharing Model: Pioneering Framework for African Research Collaboration

Understanding Multimorbidity in Africa: The Rising Challenge Driving Innovation at Wits University

In sub-Saharan Africa, the epidemiological transition from infectious to non-communicable diseases has led to a surge in multimorbidity, defined as the coexistence of two or more chronic conditions such as diabetes, hypertension, and cardiovascular disease. This complex health burden affects millions, straining healthcare systems and highlighting the need for advanced data-driven solutions. At the University of the Witwatersrand (Wits) in South Africa, the Multimorbidity in Africa: Digital Innovation, Visualisation and Application (MADIVA) Research Hub is at the forefront, leveraging longitudinal datasets to uncover patterns and predict risks.

Wits researchers, including Professor Michèle Ramsay from the Sydney Brenner Institute for Molecular Bioscience (SBIMB) and Professor Scott Hazelhurst in bioinformatics, lead this NIH-funded initiative under the DS-I Africa Consortium. By integrating data from rural Bushbuckridge (Agincourt HDSS) and urban Nairobi (NUHDSS), MADIVA addresses gaps in representative African-ancestry data, moving beyond proxies like African American datasets.

Launch and Objectives of the MADIVA Research Hub

Established as part of a broader effort to harness data science for health discovery, MADIVA's protocol outlines four core objectives: curating and harmonizing integrated datasets, developing multimorbidity risk models via machine learning, creating interactive dashboards for clinicians and researchers, and building data science capacity across Africa. The hub draws on over 117,000 individuals from Agincourt and 90,000 from Nairobi, encompassing demographic, clinical, genomic, and verbal autopsy data.

Harmonization involves deterministic and probabilistic linkage, ensuring reproducibility through detailed documentation. Projects focus on precision public health, like automatic stratification to detect high-risk groups, and tools such as NLP for cause-of-death extraction from verbal autopsies. Pilot initiatives extend to Burkina Faso and South Africa, fostering regional collaboration.

The Imperative for Ethical Data Sharing in Collaborative Research

Health data sharing accelerates discovery but poses risks in resource-constrained settings: privacy breaches, data exploitation, stigma from sensitive conditions, and infrastructure limitations. In Africa, where multimorbidity data is underrepresented, ethical frameworks are crucial to balance openness with protection. MADIVA's policy emerges as a pioneering model, informed by a literature review of 232 publications (2010–2022) revealing patterns like hypertension-diabetes clusters but underscoring data gaps.

Challenges include negotiating across borders, aligning ethics approvals with consent, and complying with laws like South Africa's POPIA and Kenya's Data Protection Act. Wits' approach emphasizes participant rights, recognizing contributions from communities who 'generously gave their samples and data,' as noted by Prof. Ramsay.

Core Principles of the Wits MADIVA Ethical Framework

The framework adheres to FAIR principles (Findable, Accessible, Interoperable, Reusable) while prioritizing human subjects' rights. Key tenets include:

• Informed Consent and Risk-Benefit Balance: Data use must align with original consents; ethics committees approve extensions.
• Confidentiality and Anonymization: Primary identifiers removed, geospatial data re-coded, unique IDs assigned.
• Capacity Building: Prioritize African scientists' involvement; training in data protection mandatory.
• Benefit Sharing: Co-authorships, long-term collaborations for external users.
• Security: Data on Wits Core Research Cluster; breaches reported within 48 hours.

This ensures equitable access, preventing 'data colonialism' while advancing open science.

Photo by Artyom Korshunov on Unsplash

Governance Structure: From Custodians to Access Committees

MADIVA's governance is multi-tiered. Data Custodians (coinvestigators like SAMRC/Wits Agincourt Unit, APHRC) prepare anonymized datasets. The Coordination Committee (CC) approves analyses via Manuscript Concept Documents (MCDs) and sharing methods, requiring raw data partners' representation.

Data Access Committees (DACs) at Wits Health Consortium and APHRC handle long-term requests, evaluating merit, training, African inclusion (substantive role within 3 years), and risks. External requests follow a formal process: submission, CC review, HREC approval, Data Access Consent (DAC) or Transfer Agreement (DTA).

Implementation: Internal vs. External Access and Sanctions

Internal MADIVA members access via agreements on the secure cluster for approved MCDs. External collaborators submit requests; approvals lead to DTAs mandating no re-identification, purpose limits, and audits. Data embargo: ~24 months post-QC +12-month exclusive period.

Storage: Sensitive in SA repositories or APHRC portal; aggregates openly shared. Violations trigger access revocation, Hub expulsion, retraction requests, and legal reporting. Training ensures users understand POPIA/GDPR equivalents.Full MADIVA Policy PDF

For researchers eyeing research assistant jobs in bioinformatics, this model exemplifies rigorous governance.

Case Studies and Recent Publications Spotlighting the Model

The framework's efficacy is demonstrated in a 2025 Global Health Action paper by Daphine Tinashe Nyachowe (Wits PhD candidate), detailing lessons from policy development. Featured in Wits news (March 2026) and NIH's Global Health Matters (Jan/Feb 2026), it offers adaptable strategies for LMICs.

Prof. Ramsay: 'Managing that responsibly is interesting yet difficult.' Prof. Hazelhurst engaged legal expertise early. Outcomes include stratified models revealing multimorbidity clusters, aiding interventions.Read the full paper

Impacts on Capacity Building and African Health Outcomes

MADIVA boosts skills via workshops, mentorship, and pilots, prioritizing early-career Africans. Dashboards empower clinicians with patient insights and planners with aggregates. Risk models using polygenic scores predict diabetes onset, informing sub-county strategies.

In SA, it aligns with NDoH priorities, enhancing primary care. Broader: Over 1000 new African genomes added globally, reshaping polygenic risk scores.Sydney Brenner Institute

For aspiring academics, Wits offers paths in crafting academic CVs for such hubs.

Photo by Jonathan Gong on Unsplash

Stakeholder Perspectives and Challenges Overcome

Participants, researchers, and funders praise the balance: Openness fosters collaboration; protections build trust. Challenges like cross-border ethics (SA-Kenya) resolved via joint DACs. Stigma risks mitigated by aggregate sharing.

• Researchers: 'Complex negotiation yields robust integration' – Ramsay.
• Communities: Benefit from derived interventions.
• Funders (NIH): Extended phases enable impact.

Lessons: Embed legal/ethical input early; prioritize local capacity.

Future Outlook: Scaling the Framework Across Africa

MADIVA eyes expansion, with data in repositories post-embargo. Future: AI enhancements like LLMs for text processing, genomic integrations. For SA higher ed, it positions Wits as leader in ethical data science, attracting talent amid global genomic pushes.

Explore Rate My Professor for Wits faculty insights or higher ed jobs in genomics. As multimorbidity rises, this model offers actionable blueprint for ethical innovation.

Internal links to university jobs, career advice, and SA opportunities support aspiring researchers.