Auditing of SNOMED CT’s Hierarchical
Structure using the National Drug File -
Reference Terminology
Aleksandr Zakharchenko, BS
1
, James Geller, PhD
1
1
New Jersey Institute of Technology, Newark, NJ
Abstract. With the ongoing development in the field of Medical Informatics, the
availability of cross-references and the consistency of coverage between
terminologies become critical requirements for clinical decision support. In this
paper, we examine the possibility of developing a framework that highlights and
exposes hierarchical incompatibilities between different medical terminologies in
order to facilitate the process of achieving a sufficient level of consistency
between terminologies. For the purpose of this research, we are working with the
Systematized Nomenclature of Medicine – Clinical Terms (SNOMED CT) and the
National Drug File - Reference Terminology (NDF-RT) – a clinical terminology
focused on drugs. For discovery of inconsistencies we built an automated tool.
Keywords. SNOMED CT, Terminology as Topic, Biomedical Ontologies, Health
Care Quality Assurance.
Introduction
SNOMED CT (formerly an acronym for Systematized Nomenclature of Medicine) –
(Clinical Terms) is a large controlled medical ontology governed by the International
Health Ontology Standards Development Organization (IHTSDO) [1, 2]. It is currently
considered to be the most comprehensive, multilingual clinical healthcare terminology
in the world. In the July 2014 US release, SNOMED CT contained over 300000 active
concepts divided into 19 is-a hierarchies, represented as Directed Acyclic Graphs
(DAGs). SNOMED CT’s attribute relationships are not considered in this paper.
In the past, we have conducted extensive research on auditing and quality
assurance of terminologies with a focus on SNOMED CT [3-6]. In this paper, we are
combining our SNOMED CT work with the National Drug File - Reference
Terminology (NDF-RT), produced by the U.S. Veterans Health Administration (VHA)
[7]. By comparing the is-a hierarchical structures of these two terminologies, it
becomes possible to further improve the quality of the coverage of both terminologies.
As different terminologies use various sources of information, differences in
coverage are inevitable. Thus, the is-a hierarchies of these terminologies are likely to
differ significantly in certain areas. An attempt to combine the hierarchies of two
terminologies could help with identifying missing concepts as well as suggest better
ways to classify concepts in the is-a hierarchies. The idea of performing a comparative
classification analysis between terminologies has been investigated [8-11], with some
approaches focusing on SNOMED CT and NDF-RT. The most detailed analysis of
Digital Healthcare Empowering Europeans
R. Cornet et al. (Eds.)
© 2015 European Federation for Medical Informatics (EFMI).
This article is published online with Open Access by IOS Press and distributed under the terms
of the Creative Commons Attribution Non-Commercial License.
doi:10.3233/978-1-61499-512-8-130
130