Healthcare digital twins are emerging as powerful tools for simulating patient conditions and operational workflows in real time. This paper explores the architectural and technical foundation necessary for building high-fidelity digital twins—those capable of accurate, synchronized, and responsive modeling. It identifies key challenges, including fragmented data, latency, poor semantic alignment, and identity inconsistencies. To overcome these, the study proposes a five-layer architecture integrating real-time data processing, data mesh principles, and master data management (MDM). Through case studies involving heart failure monitoring and hospital operations, the research demonstrates improvements in fidelity, latency, and interoperability. The study concludes with strategic guidance for healthcare organizations and outlines future research topics, including automated twin generation and federated implementations. By aligning infrastructure with intelligence, the proposed model advances the promise of high-fidelity digital twins from concept to clinical reality.
Keywords: Digital twin, Healthcare, data mesh, master data management (MDM), real-time processing
