Profile and Data Quality of DataVerse Japan: A Multi‐Institutional Clinical Data Platform for Emergency and Critical Care Research

Le résumé fourni par la source

Aim: DataVerse Japan is a multi-institutional clinical data platform that retains all exportable clinical information from each participating center's information systems and generates analysis-ready datasets for each research question. This study aimed to describe its design, participating centers, data composition, and patient profile, and to characterize its data quality. Methods: The platform captures clinical information from each center's information systems, currently the electronic health record (EHR), which incorporates patient data management system (PDMS) data and claims data; reshapes the records into a common five-column long-format (Code, Day, Time, Variable, Value); and stores them as relationally linked. rds objects. We summarized admissions to the participating centers between January 2016 and December 2025. Results: Four emergency and critical care centers contributed 41,173 admissions from 38,287 individuals and 644,838 patient-days. The platform contains approximately 45 million vital sign records, 30 million laboratory records, and 32 million claims records. Median patient age was 70 years (IQR 49-81), and in-hospital mortality was 16.3%. Admission-level presence coverage, defined as the proportion of admissions with at least one record in a category, varied substantially by category, center, and year and improved after EHR and PDMS upgrades. Conclusion: DataVerse Japan provides a continuously updated multi-institutional clinical data platform that expands its participating centers over time. The coverage and validation metrics reported here allow investigators to judge whether the platform is suitable for a specific research question in emergency and critical care medicine in Japan.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.