Published in:
Open Access 01-12-2018 | Research
Evaluating lexical similarity and modeling discrepancies in the procedure hierarchy of SNOMED CT
Published in: BMC Medical Informatics and Decision Making | Special Issue 4/2018
Login to get accessAbstract
Background
SNOMED CT is a standardized and comprehensive clinical terminology that is used in Electronic Health Records to capture, store and access clinical data of patients. Studies have, however, shown that there are inconsistencies inherent in the modeling of concepts in SNOMED CT that can have an impact on its usage to record clinical data and in clinical decision-making tools.
Methods
An effective lexical approach to identifying inconsistencies with high likelihood in the structural modeling of the concepts of SNOMED CT is discussed and assessed. The approach uses the two or more concepts in the context of their lexical similarity to compare their modeling in order to identify inconsistencies. A sample of 50 sets is randomly picked from the Procedure hierarchy of SNOMED CT and evaluated for inconsistencies.
Results
Of the 50 randomly picked sets, 58% are found to exhibit one or more concepts with inconsistencies. In terms of concepts, 29% of the 146 concepts are found to exhibit one or more inconsistencies.
Conclusions
The assessment of the sample concepts shows that SNOMED CT is not free from inconsistencies which may affect its use in clinical care and decision support systems. The proposed methodology is found to be effective in identifying areas of SNOMED CT that may be in need of quality assessment.