Medical Vocabulary, Terminological Resources and Information Coding in the Health Domain
Duclos C., Burgun A., Lamy J. B., Landais P., Rodrigues J. M., Soualmia L., Zweigenbaum P.
This chapter explains why it is hard to use medical language in computer applications and why the computer must adopt the human interpretation of medical words to avoid misunderstandings linked to ambiguity, homonymy and synonymy. Terminological resources are specific representations of medical language for dedicated use in particular health domains. We describe here the components of terminology (terms, concepts, relationships between concepts, definitions, constraints). The various artefacts of terminological resources (e.g. thesaurus, classification, nomenclature) are defined. We also provide examples of the dedicated use of terminological resources, such as disease coding, the indexing of biomedical publications, reasoning in decision support systems and data entry into electronic medical records. ICD 10, SNOMED CT, and MeSH are among the terminologies used in the examples. Alignment methods are described, making it possible to identify equivalent terms in different terminologies and to bridge different domains in health. We also present plans for multi-terminological servers, such as the UMLS (Unified Medical Language Systems), which provide a key vocabulary linking heterogeneous health terminologies in different languages.