Overlapping Complex Concepts Have More Commission Errors, Especially in Intensive Terminology Auditing

Area of Science:

• Medical Informatics
• Terminology Management
• Health Data Standards

Background:

• SNOMED CT is a comprehensive clinical terminology vital for health data interoperability.
• Auditing is crucial for maintaining terminology quality, focusing on commission (modeling) and omission (completeness) errors.
• Previous remodeling efforts targeted specific SNOMED CT subhierarchies, highlighting the need for efficient auditing methods.

Purpose of the Study:

• To evaluate the effectiveness of 'overlapping concepts' for identifying commission errors in SNOMED CT.
• To assess the utility of this auditing strategy during both intensive and routine maintenance periods.
• To determine if focusing on overlapping concepts enhances auditing yield, particularly under resource constraints.

Main Methods:

• A longitudinal study analyzing SNOMED CT releases over several periods.
• Investigated the efficacy of auditing complex, 'overlapping concepts' to detect commission errors.
• Compared auditing yield during intensive and maintenance phases using this focused methodology.

Main Results:

• Overlapping concepts proved effective in identifying commission errors within SNOMED CT.
• The methodology demonstrated increased auditing yield, especially when auditing resources were limited.
• This approach is beneficial during both intensive auditing and ongoing maintenance cycles.

Conclusions:

• Auditing overlapping concepts is an efficient strategy for detecting commission errors in SNOMED CT.
• This method maximizes auditing yield, offering a practical solution for resource-scarce environments.
• The findings support the targeted auditing of complex concepts for improved terminology quality assurance.