A Knowledge Intensive Approach to Mapping Clinical Narrative to LOINC

Area of Science:

• Clinical informatics
• Natural Language Processing (NLP)
• Medical Terminology

Background:

• Natural Language Processing (NLP) systems require context-specific tuning for clinical text.
• Accurate mapping of clinical information to standardized codes is essential for data interoperability and analysis.

Purpose of the Study:

• To develop a knowledge-intensive methodology for mapping clinical text to Logical Observation Identifiers Names and Codes (LOINC).
• To improve the precision and recall of extracting vital signs, body measurements, and diagnostic procedures from clinical reports.

Main Methods:

• A knowledge-intensive approach was employed, integrating textual, ontological, and pragmatic knowledge.
• The system processes published case reports to map clinical information to fully specified LOINC codes.
• Knowledge domains include physiology and clinical process understanding.

Main Results:

• The system achieved high performance in mapping clinical text to LOINC codes.
• Evaluation on 4809 sentences demonstrated a precision of 89% and a recall of 93%.
• An F-score of 0.91 indicates the overall effectiveness of the methodology.

Conclusions:

• The proposed knowledge-intensive methodology offers a robust approach for clinical text mapping to LOINC.
• This method can support the development of semi-automated tools for clinical coders.
• Enhanced NLP in clinical settings can improve data standardization and utilization.