Improving the interoperability of drugs terminologies: Infusing local standardization with an international perspective
Jean Noël Nikiema,James Liang,Man Qing Liang,Davllyn Dos Anjos,Aude Motulsky,Davllyn dos Anjos
DOI: https://doi.org/10.1016/j.jbi.2024.104614
IF: 8
2024-03-01
Journal of Biomedical Informatics
Abstract:OBJECTIVES: The objective of this study is to describe how OCRx (Canadian Drug Ontology) has been built to address the dual need for local drug information integration in Canada and alignment with international standards requirements.METHODS: This paper delves into (i) the implementation efforts to meet the Identification of Medicinal Product (IDMP) requirements in OCRx, alongside the ontology update strategy, (ii) the structure of the ontology itself, (iii) the alignment approach with several reference Knowledge Organization Systems, including SNOMED CT, RxNorm, and the list of "Code Identifiant de Spécialité" (CIS-Code), and (iv) the look-up services developed to facilitate its access and utilization.RESULTS: Each OCRx release contains two distinct versions: the full and the up-to-date version. The full version encompasses all drugs with a DIN code sanctioned by Health Canada, while the up-to-date version is limited to drugs currently marketed in Canada. In the last release of OCRx, the full version comprises 162,400 classes; meanwhile, the up-to-date version consists of 36,909 classes. In terms of mappings with OCRx, substances in RxNorm and SNOMED CT fall below 40%, registering at 37% and 22% respectively. Meanwhile, mappings for CIS-Code achieve coverage of 61%. The strength mappings are notably low for RxNorm at 40% and for CIS-code at 28%. This affects the mapping of clinical drugs, which are predominantly alignable through post-coordinated expressions: 56% for RxNorm, 80% for SNOMED CT, and 35% for CIS-Code. The main support service of OCRx is a look-up service known as PaperRx that displays OCRx's entities based on description logic queries (DL-queries) performed through the classified structure of OCRx. The look-up services also contain a SPARQL endpoint, an OCRx OWL file downloader, and a RESTful API.DISCUSSION: The OCRx ontology demonstrates a significant effort towards integrating Canadian drug information with international standards. However, there are areas for improvement. In the future, our focus will be on refining the structure of OCRx for better classification capability and improvement of dosage conversion. Additionally, we aim to harness OCRx in constructing an ontology-based annotator, setting our sights on its deployment in real-world data integration scenarios.
medical informatics,computer science, interdisciplinary applications