Automated Mapping of Observation Archetypes to SNOMED CT Concepts - Foundations on Natural and Artificial Computation - page 558

In order to increase the recall of our approach, in the future we plan to com-
bine the lexical techniques used in the Algorithm 1 with structure-based tech-
niques developed in the Algorithm 2 and 3. We will also explore more significant
SNOMED CT relationships for validation and disambiguation.

6 Conclusions

In this study, we propose an automated approach to mapping terms identified
in archetypes to SNOMED CT concepts. Our approach applies a combination
of basic matching methods classically used in ontology matching[15]. First, seve-
ral lexical techniques identify similar strings between the archetypes terms and
SNOMED CT concepts. In parallel, a semantic technique processes mappings
by extracting the relevant SNOMED CT subontology to the archetypes. The
extraction of the subontology is carried out by covering the SNOMED CT re-
lationships with the meaning probably more similar to the intended meaning
of the archetype data. The methods were applied to the mapping of the twenty
different Observation archetype models, with a total of four hundred and ninety-
four selected archetype terms, to the complete SNOMED CT. In total, 94%
precision and 69% recall was reached. Our method has revealed some degree of
semantics in the structure of the information defined in the available archetypes
with some SNOMED CT relationships, such as interprets and IS A . Future
work will require to study more types of SNOMED CT relationships in order to
guarantee our work.

Acknowledgements. This work has been funded by the Ministerio de Edu-
cacion y Ciencia, through the national research project Gestion de Terminologias
Medicas para Arquetipos TIN2009-14159-C05-05.

References

1. SNOMED-CT: Systematized Nomenclature of Medicine-Clinical Terms,

http://www.ihtsdo.org/snomed-ct/
2. LOINC: Logical Observation Identifiers Names and Codes, http://loinc.org/
3. NHS: Connecting for Health project (2010),
http://www.connectingforhealth.nhs.uk/
4. OpenEHR. archetypes., http://www.openehr.org/
5. Sundvall, E., Qamar, R., Nystrom, M., Forss, M., Petersson, H., Karlsson, D.,
Ahlfeldt, H., Rector, A.: Integrations of tools for binding archetypes to snomed ct.
BMC Medical Informatics and Decision Making 8(Suppl. I), S7 (2008)
6. CEN. Europen Committee for Standarization, http://www.cen.eu/
7. HL7. Health level Seven International, http://www.hl7.org/
8. OMG. Omg Healthcare Domain Task Force, http://healthcare.omg.org/
9. UMLS. Unified Medical Language System, http://www.nlm.nih.gov/research/umls/
10. Metathesaurus, http://www.ncbi.nlm.nih.gov/books/NBK9684/pdf/ch02.pdf
11. MetaMap, http://mmtx.nlm.nih.gov/