Implementing a quality assurance and maintenance framework for standardized clinical terminology data in a multi-institutional setting.

Abstract

Ensuring high-quality mapping between local terms and standard clinical terminologies is crucial for maximizing the value of real-world data (RWD). This study aimed to apply and evaluate a scalable quality assurance (QA) and maintenance framework for local-to-standard vocabulary mappings across multiple institutions. The framework was operationalized through six data quality dimensions: validity, uniqueness, completeness, consistency, timeliness, and accuracy. Using a centralized terminology master within Kakao Healthcare (KHC)'s Universal Data Model (UDM), data from three hospitals were mapped and assessed based on standard frameworks. Validity checks identified format and reference errors within the reviewed code; uniqueness audits removed duplicates based on various composite key constraints; completeness analyses identified unmapped codes; and consistency was improved through intra- and inter-institutional harmonization. Timeliness was managed via regular updates aligned with current Observational Medical Outcomes Partnership Common Data Model (OMOP CDM) and Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) releases, and accuracy was ensured with expert review and adjudication. Issues in all six dimensions were resolved. The framework corrected format errors, removed redundancies, and bridged mapping gaps. Inter-institutional harmonization minimized semantic discrepancies, ensuring high-fidelity standardization across three hospitals. This scalable framework ensures terminology mapping quality and data integrity. It provides a robust foundation for reliable research and operational use of multi-institutional RWD.