Related Experiment Videos
Application of a Natural Language Processing Algorithm to Asthma Ascertainment. An Automated Chart Review
Chung-Il Wi1,2, Sunghwan Sohn3, Mary C Rolfes2,4
11 Department of Pediatric and Adolescent Medicine.
Summary
Natural language processing (NLP) accurately identifies asthma in electronic medical records (EMRs), improving childhood asthma research and care. This automated method enhances efficiency and recognition of pediatric asthma.
Area of Science:
- Medical Informatics
- Pediatric Pulmonology
- Computational Linguistics
Background:
- Asthma diagnosis presents challenges due to heterogeneity, hindering research and care.
- Electronic medical records (EMRs) offer vast data but require efficient analysis methods.
Purpose of the Study:
- To validate a natural language processing (NLP) algorithm for automated asthma ascertainment in EMRs.
- To assess the accuracy of NLP in identifying asthma criteria compared to manual chart review.
Main Methods:
- Retrospective birth cohort study of 497 subjects from the Mayo Birth Cohort (1997-2007).
- Evaluated NLP algorithm's criterion validity against abstractor gold standard and construct validity with asthma risk factors.
- Assessed sensitivity, specificity, and predictive values of the NLP algorithm.
Main Results:
- Asthma prevalence was 31% in the cohort.
- NLP algorithm demonstrated high performance: 97% sensitivity, 95% specificity, 90% positive predictive value, and 98% negative predictive value.
- Identified asthma risk factors (e.g., allergic rhinitis) consistently with manual review.
Conclusions:
- NLP-based asthma ascertainment is a viable and efficient method in the EMR era.
- Automated chart review using NLP can facilitate large-scale clinical studies.
- Improved asthma recognition and care for childhood asthma are potential benefits.