Expanding research on clinical agent data to a virtual space of location, time, and context: exploration of data structures and limitations using a pragmatic digital twin.

INTRODUCTION: Fragmented healthcare data and the inflexible structure of current hospital information systems (HIS) hinder the routine use of data for research and process optimization. Conventional HISs often fail to represent physical resources, spatial relationships, and temporal dependencies among hospital entities. In this feasibility study, we demonstrate the practical application of OMNI-SYS, a pragmatic object-centric hospital information system (oHIS) framework that represents objects within a 3D space of Time, Location, and Context. Using pre-operative patient tracking as an example, we show how OMNI-SYS generates structured datasets suitable for process mining and clinical workflow optimization. MATERIAL AND METHODS: At TUM University Hospital in Munich, 18 healthy volunteers followed predefined pre-operative pathways for seven common gastrointestinal surgery procedures. Participants moved through real hospital departments, including administrative registration, premedication, ECG, lung function, endoscopy, CT, and MRI, while spatiotemporal events were recorded using OMNI-SYS. OMNI-SYS is an object-centric digital twin that integrates this data, capturing all relevant agents and their interactions throughout the study. The resulting dataset, with 88 timestamped events, was analyzed for its structure and suitability for advanced data analysis. RESULTS: OMNI-SYS successfully captured interactions among patients, simulated staff, and devices at all pre-operative stations. The resulting object-centric data structure was directly compatible with contemporary process mining and workflow optimization methods. Patient co-occurrence at different stations could be efficiently assessed, and OMNI-SYS enabled all recorded events to be replayed for comprehensive post-hoc analysis of inter-object relationships. CONCLUSION: oHIS-like frameworks, such as OMNI-SYS, can facilitate routine clinical data and process mining by providing structured datasets for modern evaluation. In our study, OMNI-SYS generated object-centric data that models healthcare entities and their interactions, addressing key limitations of current data infrastructures. These results provide a foundation for designing real-world studies to analyze routine clinical workflows and align data structures to bridge clinical routine and research.