AI and Data Mesh: A New Paradigm for Decentralized Healthcare Data Management Using Artificial Intelligence

Abstract

Purpose: This paper seeks to propose and evaluate a new architectural approach merging Artificial Intelligence (AI) and Data Mesh concepts to create a decentralized, privacy-preserving, and computation-independent environment for healthcare data

Methodology: This paper employs the design science research methodology to develop the decentralized data universe through domain ownership of data products, federated learning approaches, data privacy via blockchain concepts, and edge computing channels for inferencing. The proposed architecture is then validated through application to real-world case studies and synthetic inventions, presenting use in unified yet varied healthcare settings. Validation occurred through qualitative and quantitative assessment measuring performance against legacy systems for enhancements in security, interoperability, and redundancy.

Findings: The application of AI with Data Mesh correlates to clinically relevant activities with real-time healthcare data, research-related aggregate data crossovers without jeopardizing the proprietary data products of mandated research teams, and collaborative ventures for collaborative health claims processing that reduces fraudulent activities. Results indicated that using a decentralized system for healthcare data libraries significantly improves scalability, effectively enhances privacy protections for personal health information (PHI), and protected health information (PHI) as well as increases resiliency in direct comparison to cybersecurity and centralized service disruption risks. Decentralized redundancies also improve where the demand increases are irrespective of identity or identity-based scaling efforts.

Unique Contribution to Theory, Practice and Policy: This paper helps to bring a more comprehensive awareness of what healthcare data systems can be formed due to AI and Data Mesh applications. In practice, organizations ready to revamp their healthcare data governance and processing systems now have a scalable solution that blends with current regulatory actions while also focusing on cybersecurity considerations to maintain patient trust. Finally, current regulatory considerations must be altered to reflect an ethically sound sociotechnical solution brought on by these decentralized systems in the healthcare space that integrate AI but question FDA trust and algorithmic action on its own without human support during treatment efforts.