Analytical Study of the Research Gap in Federated Learning for Privacy-Preserving Healthcare Analytics in Nigeria

Federated learning enables machine learning model training across distributed healthcare data sources without centralising sensitive patient data, yet its application in the Nigerian healthcare system, where data governance is fragmented and privacy regulation is nascent, remains an unexplored research area. This study examined the research gap in federated learning for privacy-preserving healthcare analytics in Nigeria through systematic review and original feasibility analysis. A systematic scoping review identified only 2 publications from 2018 to 2024 specifically addressing federated learning in Nigerian healthcare, versus a rapidly growing global literature of over 400 publications. Original feasibility analysis assessed data infrastructure and governance readiness across eight public hospitals in the South South and South East zones through structured IT manager interviews. The analysis confirmed significant barriers: inconsistent electronic health record adoption (present in 37.5 percent of assessed facilities), absent data use agreements between facilities, inadequate bandwidth for federated model communication at 62.5 percent of sites, and absent data science personnel at all eight sites. The study proposes an original Phased Federated Learning Readiness Framework for Nigerian hospitals specifying infrastructure prerequisites, governance requirements, and capacity building milestones before federated learning deployment. Expert consultation with eight federated learning and health data science specialists validated the framework. Six priority research topics are identified. Recommendations include FMOH piloting a federated electronic records system as a foundational step, and NITDA funding federated learning research grants for Nigerian universities.

Keywords: federated learning, healthcare analytics, privacy-preserving machine learning, Nigeria, research gap