Comparative analysis of the VA/Kaiser and NLM CORE problem subsets: an empirical study based on problem frequency

Area of Science:

• Health Informatics
• Medical Terminology
• Electronic Health Records

Background:

• The problem list in electronic medical records (EMR) is vital for clinical care and research.
• Its full benefits rely on standardized coded terminologies.
• SNOMED-CT and ICD-9 are standard terminologies, with SNOMED-CT offering VA/KP and CORE subsets.

Purpose of the Study:

• To evaluate and compare the coverage of SNOMED-CT subsets (VA/KP and CORE) for problem list documentation.
• To characterize the overlap and measure the clinical coverage of these subsets.

Main Methods:

• Applied SNOMED-CT VA/KP and CORE subsets to 100,000 random electronic medical records from Brigham and Women's Hospital.
• Quantified the proportion of coded problem entries covered by each subset.

Main Results:

• The CORE subset (5,814 terms) covered 94.8% of coded problem entries.
• The VA/KP subset (17,761 terms) covered 84.0% of coded problem entries (p<0.001).
• CORE demonstrated superior coverage and fewer clinically significant gaps in the sampled records.

Conclusions:

• The CORE subset of SNOMED-CT provides more comprehensive coverage for electronic health record problem lists than the VA/KP subset.
• CORE is a more effective terminology for documenting patient problems within EMR systems.
• Further analysis of clinically significant gaps in both subsets is warranted.