Definition and validation of SNOMED CT subsets using the expression constraint language

Area of Science:

• Medical Informatics
• Computational Linguistics
• Graph Databases

Background:

• SNOMED CT Expression Constraint Language (ECL) defines intensional subsets of clinical meanings.
• Expression Constraints (ECs) are crucial for terminology binding in clinical information models.
• Efficient execution and validation of ECs are necessary for managing clinical data.

Purpose of the Study:

• To develop methods for simplifying, semantically validating, and executing ECL over graph databases.
• To create a platform (SNQuery) for EC execution and subset visualization.
• To enhance the exploration, understanding, and validation of SNOMED CT subsets.

Main Methods:

• Representing SNOMED CT content in a graph-oriented database.
• Translating ECL expressions into Cypher Query Language queries for execution.
• Implementing EC simplification methods to reduce complexity and execution time.
• Developing a circle packing-based visualization for subset validation.

Main Results:

• Developed SNQuery, a platform for defining intensional subsets using ECL execution on a graph database.
• Incorporated methods for EC simplification, semantic validation, and subset visualization.
• Evaluated SNQuery based on EC execution times, demonstrating platform efficiency.

Conclusions:

• Methods for simplifying, validating, and executing ECL over graph databases have been established.
• A visualization method aids in exploring and validating subsets, refsets, and terminology.
• The definition of intensional subsets is vital for semantic interoperability between EHR systems.