본문 바로가기
  • Home

Identifying Missing Concepts in SNOMED CT by Utilizing Attribute Relationships of Sibling Concepts

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2024, 29(9), pp.197-205
  • DOI : 10.9708/jksci.2024.29.09.197
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : June 25, 2024
  • Accepted : September 3, 2024
  • Published : September 30, 2024

Wooseok Ryu 1

1부산가톨릭대학교

Accredited

ABSTRACT

SNOMED CT is the most widely used comprehensive clinical terminology system worldwide. However, due to the vastness of the terminology and the continuous growth of medical knowledge, the system involve quality issues such as structural inaccuracies and inconsistencies, including missing concepts or relationships and hierarchical errors. In this paper, we propose a method to enhance the consistency of the system by detecting potentially missing concepts by utilizing attributes linked to concepts. The proposed method analyzes the characteristics of the attribute relationships of concepts, extracts sibling concepts that share the same characteristics, and then evaluates whether the parent concepts reflect these characteristics to detect potentially missing concepts. By applying this method to the March 2023 SNOMED CT international release, we identified 564 instances where parent concepts did not reflect the common attributes of their sibling concepts, and a total of 384 potentially missing concepts were detected, including cases involving multiple parent concepts.

Citation status

* References for papers published after 2023 are currently being built.

This paper was written with support from the National Research Foundation of Korea.