Towards a Privacy-Preserving Health Data Sharing: Architecture and Critical Implementation Factors

Healthcare systems increasingly rely on connected devices and digital infrastructures, generating sensitive patient data. Traditional centralized approaches pose risks to privacy, scalability, and regulatory compliance. Hence, there is a growing need for decentralized architectures that allow meaningful data use without compromising individual privacy. This paper introduces a federated architecture for secure and privacy-preserving health data sharing. The design integrates data from personal medical devices, institutional systems, and regional platforms, using standardized interfaces and local processing capabilities. The critical implementation challenges and security requirements are also identified and enforced through on-device analytics, anonymization techniques, and layered access control. A proof-of-concept implementation demonstrates the practical feasibility of this model using an edge-based AI system deployed on a connected prosthetic device. The results show how decentralized data processing and privacy-aware mechanisms can enable secure clinical insights without requiring continuous data aggregation.