From Static Records to Virtual Digital Patients: An Artificial Intelligence-Driven Framework for Medical Data Assetization and Governance.

Abstract

Healthcare data remain difficult to use because of fragmentation, sensitivity, and overlapping rights and responsibilities. A virtual digital patient links a longitudinal representation of an individual patient with modeling capabilities for specified clinical or research questions. This article proposes a conceptual socio-technical framework for governing how virtual digital patients are created, accessed, used, and updated. A narrative synthesis of selected policy, modeling, interoperability, and health artificial intelligence literature informs four interdependent layers: data and representation; artificial intelligence modeling and simulation; governance and control; and clinical and external use. The framework connects standardized longitudinal representations, knowledge-supported reasoning, disease-specific simulation, authorization, output review, lifecycle management, and accountability. It distinguishes data resources, model services, operational capabilities, and potentially recognizable assets; technical functionality alone does not establish asset status. An illustrative oncology trial-design scenario shows how purpose-limited requests could produce bounded, reviewed outputs without transferring raw patient records. A staged validation agenda addresses data quality, model performance, workflow safety, clinical utility, and governance controls. The article reports neither a new algorithm nor a deployed or clinically validated system. Its contribution is an integrated reference architecture for evaluating how patient-specific modeling, institutional stewardship, and value return may be aligned. Disease-specific implementation and empirical evaluation are required before clinical or commercial claims can be established.

Abstract as supplied by the indexing source. © the authors / publisher.