Generating SNOMED CT subsets from clinical glossaries: An exploration using clinical guidelines

Abstract

The large SNOMED CT (SCT) terminology has gained adoption in the last years. However, its practical application for coding clinical information is hampered by its complexity and size. The mechanism of subsets allows for creating clusters of SNOMED CT terms that cover a particular application or clinical domain. These subsets are usually defined following some sort of consensual expert-driven process that is effort-intensive. The automated generation of subsets from clinical document corpora have been proposed elsewhere, but they still require a collection of documents that is representative for the targeted domain. This paper describes an experiment in using clinical guidelines' glossaries as a seed terminology for automatically generating subsets by traversing SNOMED relationships. Quantitative analysis reveals that traversing patterns need to be limited, and expert assessments point out that the approach may be viable at least for bootstrapping the process of elaborating the subsets. © 2011 Springer-Verlag.