Country / Region
EMEA
Tags
Data analytics, Mapping, Research
Semantic interoperability of laboratory data is essential for enabling secondary use in research networks like the German Medical Informatics Initiative (MII). While the MII's Lab Module provides standardized HL7 FHIR-based structures for laboratory parameters using LOINC codes, harmonizing these data across institutions for analytics remains a significant challenge. This project investigates the use of precoordinated LOINC-SNOMED expressions as a scalable, computable alternative to manual aggregation mappings, aiming to support a wide range of research questions with varying data granularity needs and to lay the groundwork for future integration of increasingly precise laboratory coding, such as the growing number of LOINC codes for established domains like immunohistochemistry in oncology.
Description
The scope of this project centers on evaluating approaches for aggregating laboratory data across multiple university hospitals participating in the MII. Specifically, we compared the traditional, manually curated LOINC-to-parent mappings from the INTERPOLAR project with the automated, ontology-driven relationships provided by the LOINC-SNOMED Ontology. The challenge is the harmonization of different laboratory parameters across a multitude of university hospitals. As an example: In addition to "1751-7 Albumin [Mass/volume] in Serum or Plasma", there exist at least 10 different LOINCs describing similar concepts without being identical or interchangeable at the analytical level.
Our analysis focused on the 300 most frequently used laboratory parameters, assessing how each approach supports dynamic value set generation for analytics platforms. This enables researchers to quickly determine the number of patients or analyte-related measurements available for study, without resorting to overly broad aggregation that could compromise clinical relevance. The project also explores how these methods can be extended to support the integration of new, more precise LOINC codes, which facilitate structured data collection and sharing in established laboratory domains.
Scope
SNOMED CT was selected for its robust ontology, hierarchical relationships, and capacity for formal logical definitions. These features make it uniquely suited for supporting computable aggregation and semantic interoperability across diverse and evolving laboratory data sets. The ability to dynamically generate and expand value sets, particularly as new and more precise laboratory parameters and codes become available, positions SNOMED CT as a future-proof solution for large-scale data harmonization and analytics. The LOINC-SNOMED-Ontology is an important publically available ressource usable to aggregate different LOINCs under similar concepts. This integration with LOINC further enhances its utility for laboratory data, supporting both broad and granular research needs.
How SNOMED CT will be used
SNOMED CT is utilized in this project through precoordinated expressions that integrate LOINC terms as formal, computable SNOMED CT concepts in the LOINC SNOMED Ontology. This integration allows for the logical representation of laboratory parameters within SNOMED CT's hierarchical structure, enabling automated reasoning and dynamic generation of value sets. The LOINC-SNOMED Browser and downloadable distribution files are employed as a key tool for validating and visualizing both manual and ontology-based mappings, facilitating real-time exploration and identification of gaps or inconsistencies.
Why SNOMED CT will be used
Contact
