Related Experiment Videos
The content coverage of clinical classifications. For The Computer-Based Patient Record Institute's Work Group on
C G Chute1, S P Cohn, K E Campbell
1Section of Medical Information Resources, Mayo Foundation, Rochester, MN 55905, USA. chute@mayo.edu
Summary
Major clinical coding systems like ICD-9-CM and ICD-10 inadequately represent patient record data. SNOMED International demonstrated superior coverage, highlighting limitations in current systems for clinical decision support and research.
Area of Science:
- Medical Informatics
- Health Information Management
- Clinical Documentation
Background:
- Patient record content relies on conditions and events.
- Standard vocabularies are crucial for computer-based records.
- Consistent data representation supports clinical decision support, research, and care delivery.
Purpose of the Study:
- To evaluate the content coverage of major clinical coding systems.
- To determine if existing systems can adequately represent clinical data.
- To assess the suitability of classifications for computer-based patient records.
Main Methods:
- A corpus of 14,247 words from clinical text across four medical centers was analyzed.
- 3,061 distinct concepts were identified and categorized (Diagnoses, Modifiers, Findings, Treatments/Procedures, Other).
- Concepts were coded using ICD-9-CM, ICD-10, CPT, SNOMED III, Read V2, UMLS 1.3, and NANDA, with a scoring system (0-2) for match completeness.
Main Results:
- SNOMED International achieved the highest overall score (1.74) and in Diagnoses (1.90).
- ICD-9-CM scored 0.77 overall, with 1.61 for Diagnoses.
- ICD-10 scored 0.62 overall, and CPT did not significantly improve ICD-9-CM's score.
Conclusions:
- No single classification system captured all clinical concepts.
- SNOMED International provided the most comprehensive coverage.
- Current major systems (ICD-9-CM, CPT, ICD-10) incompletely cover clinical content, potentially impacting analytic conclusions.