Related Experiment Videos
Using NLP to extract concepts from chief complaints.
Michael I Lieberman1, Thomas N Ricciardi
1GE Healthcare Technologies, Waukesha Wisconsin, and Department of Medical Informatics and Clinical Epidemiology, Oregon Health and Science University, Portland, USA.
AMIA ... Annual Symposium Proceedings. AMIA Symposium
|June 17, 2006
Summary
The National Library of Medicine's (NLM) MMTx engine processed over 600,000 chief complaints. This natural language processing (NLP) tool successfully extracted concepts from 25% of entries, demonstrating its utility in electronic medical records.
Area of Science:
- Medical Informatics
- Natural Language Processing
Background:
- Electronic medical records (EMRs) contain vast amounts of unstructured clinical data.
- Extracting meaningful information from clinical notes is crucial for research and improving patient care.
Purpose of the Study:
- To evaluate the performance of the National Library of Medicine's (NLM) MMTx natural language processing (NLP) engine.
- To assess the concept extraction rate and accuracy from chief complaints in an ambulatory EMR setting.
Main Methods:
- Utilized the NLM's MMTx NLP engine.
- Processed over 600,000 chief complaint strings from an ambulatory EMR.
- Analyzed the percentage of strings assigned at least one concept and the rate of incorrect assignments.
Main Results:
- Approximately 25% of the processed chief complaint strings were assigned at least one concept by the MMTx engine.
- The rate of incorrect concept assignments was found to be 2%.
Conclusions:
- The MMTx NLP engine demonstrates a viable capability for extracting concepts from clinical chief complaints.
- Further refinement may improve the concept assignment rate and maintain high accuracy in EMR data.
- This technology holds potential for enhancing clinical data analysis and information retrieval.