An Original Investigation into Secure Multi-Party Computation Protocols for Health Data Sharing Among Nigerian Hospitals

Collaborative clinical research and disease surveillance across Nigerian hospitals is hampered by patient data privacy regulations, institutional data silos, and mutual distrust between competing facilities, and secure multi-party computation protocols offer cryptographic solutions that enable joint analysis without data disclosure, but no such protocols have been designed or evaluated for the specific functional requirements and computational constraints of Nigerian hospital systems. This study conducted an original investigation into secure multi-party computation protocols for health data sharing among Nigerian hospitals, developing novel protocol adaptations and evaluating them on realistic Nigerian clinical data sharing scenarios. Three SMPC protocols (Secret Sharing with Semi-Honest Security, Oblivious Transfer-based Garbled Circuits, and Homomorphic Encryption using CKKS scheme) were adapted for three target applications: joint disease incidence analysis across six geographically distributed hospitals representing all geopolitical zones, federated clinical trial participant identification, and national pharmaceutical demand forecasting. Protocol performance was benchmarked on hardware representative of Nigerian hospital IT infrastructure (Pentium Gold servers, 4 Mbps inter-site links). Secret sharing achieved the lowest computation latency for disease incidence analysis (23 seconds for 10,000-record joint query across six parties) but required active simultaneous participation by all parties, impractical for unreliable connectivity conditions. CKKS Homomorphic Encryption enabled asynchronous participation but required 420 seconds for equivalent computation. An original Adaptive Protocol Selection Algorithm was developed that dynamically chooses the optimal SMPC scheme based on real-time network conditions and hospital availability, achieving 91 percent of optimal performance across 200 simulated network condition scenarios.