An EMOI-Based Object-Level Event Governance Framework for Multimodal Healthcare Data Integration and Clinical Research

Abstract Background To address critical bottlenecks in hospital-wide multimodal medical data management, specifically, inadequate data association, semantic heterogeneity, and low utilization efficiency—this study proposes a multimodal medical data governance framework tailored for clinical applications. Methods Utilizing 72 TB of real-world medical data from a tertiary hospital, we designed a layered, lakehouse-based governance architecture. Central to this framework is the Enterprise Multimodal Object Identifier (EMOI), a mechanism built upon the Enterprise Master Patient Index (EMPI) that establishes an object-level event association chain linking the “encounter–examination–report–diagnosis–treatment” workflow. By integrating a nine-step closed-loop governance process, the framework facilitates data cleansing, standard mapping, identity unification, object association, privacy protection, and normalized data asset management. Results The implementation of this platform generated 336 million standardized data records, achieving an EMPI consistency of 86.7% and an EMOI association rate of nearly 100%. Furthermore, data consistency, accuracy, and timeliness reached 99.63%, 97.78%, and 95.70%, respectively. Conclusions These validated outcomes have been successfully applied to support clinical research cohorts, 360-degree patient profiling, intelligent clinical decision support, and operational analytics. Ultimately, the EMOI-based object-level event governance mechanism effectively resolves the fragmentation of multimodal medical data, enabling full-lifecycle management from data integration and quality governance to downstream research utilization. Scoping review registration: As a scoping review drawing on retrospective institutional operational and clinical data, this study does not report any findings from medical intervention trials with human participants.