SNOMED2HL7: A tool to normalize and bind SNOMED CT concepts to the HL7 Reference Information Model

Area of Science:

• Health Informatics
• Clinical Data Management
• Biomedical Terminologies

Background:

• Clinical research demands system interoperability, facing challenges with heterogeneous data across institutions.
• Existing efforts in common data models and terminologies have not fully resolved distributed clinical data integration issues.
• Manual data transformation processes hinder efficient exploitation of clinical data.

Purpose of the Study:

• To address the lack of tools supporting data experts in adopting clinical standards, particularly for linking data models and vocabularies.
• To present SNOMED2HL7, a novel tool for automatically linking biomedical concepts to the HL7 Reference Information Model (RIM).

Main Methods:

• Implementation of the SNOMED Normal Form to decompose concepts and reduce storage options, adhering to IHTSDO recommendations.
• Core functionality involves binding clinical terminologies to HL7 RIM components, annotating concepts with corresponding interoperability standard options.
• Development of a web-based tool to automate information retrieval from normalization and terminology binding processes.

Main Results:

• SNOMED2HL7 demonstrates broad coverage for concepts used in legacy systems.
• The tool adheres to HL7 recommendations for managing binding overlaps and provides normalized concept bindings.
• Validated in EU projects, the tool achieved 88.47% accuracy in integrating real-world clinical research data.

Conclusions:

• SNOMED2HL7 represents a pioneering initiative for automated retrieval of information needed to transform legacy data into interoperability standards.
• While further enhancements are planned for data transformation automation, the tool currently offers essential functionality for the clinical interoperability community.