Desiderata for controlled medical vocabularies in the twenty-first century

Abstract

Builders of medical informatics applications need controlled medical vocabularies to support their applications and it is to their advantage to use available standards. In order to do so, however, these standards need to address the requirements of their intended users. Over the past decade, medical informatics researchers have begun to articulate some of these requirements. This paper brings together some of the common themes which have been described, including: vocabulary content, concept orientation, concept permanence, nonsemantic concept identifiers, polyhierarchy, formal definitions, rejection of "not elsewhere classified" terms, multiple granularities, multiple consistent views, context representation, graceful evolution, and recognized redundancy. Standards developers are beginning to recognize and address these desiderata and adapt their offerings to meet them.

Fig. 1

Multiple views of a polyhierarchy. a) Internal arrangements of nine concepts in a polyhierarchy, where E has two parents; b) Hierarchy has been collapsed so that specific concepts serve as synonyms of their more general parents; c) Intermediate levels in the hierarchy have been hidden; d) Conversion to a strict hierarchy; e) Strict hierarchy with multiple contexts for term E; f) Multiple contexts for E are shown, but are inconsistent (different children).

Fig. 2

Definitional, assertional, and contextual information in the vocabulary showing how concepts can be combined and where they will appear in a clinical record.

Fig. 3

Interchangability of redundant data representations. The structure on the left depicts the post coordination of a disease concept (Pneumonia) and a body location (Left Lower Lobe) to create a finding in an electronic medical record. The structure on the right shows a precoordinated term for the same finding (Left Lower Lobe Pneumonia). Because this latter term includes formal, structured definitional information (depicted by the is-a, has-site, and participates-in attributes), it is possible to recognize, in an automated way, that data coded in these two different ways are equivalent.