Concept representation of decision logic for hypertension management using SNOMED CT

Objective: To obtain sharable and reusable knowledge among various hospital information systems, it is essential to represent each term with standard terminology. To support knowledge representation for interoperable clinical decision support system for hypertension management, the feasibility of SNOMED CT was evaluated. Methods: Concept matching was conducted using the method of direct matching, post-coordinated matching and general matching. For semantic matching, the SNOMED CT hierarchy was considered, and for raising the mapping rate, preferred terms and synonyms were used. Results: Excluding the recommendation concepts that were not used in clinical data, finally 182 concepts were evaluated in terms of concept matching. Seventy two percent of the concepts was directly matched to pre-coordinated concepts in SNOMED CT. For the post-coordinated matching and the general matching to broader meaning, 9.3% and 18.7% were covered respectively. Conclusion: The direct coverage of SNOMED CT was moderate to high level for representing guideline knowledge concepts without loss of semantics. To supplement the coverage, it is inevitable to consider defining local concepts for implementing hypertension management systems.