BACKGROUND
Aspirin exacerbated respiratory disease (AERD) is an acquired inflammatory condition characterized by the presence of asthma, chronic rhinosinusitis with nasal polyposis, and respiratory hypersensitivity reactions on ingestion of aspirin or other nonsteroidal anti-inflammatory drugs (NSAIDs). Despite AERD having a classic constellation of symptoms, the diagnosis is often overlooked, with an average of greater than ten years between the onset of symptoms and diagnosis of AERD. Without a diagnosis, individuals will lack opportunities to receive effective treatments such as aspirin desensitization or biologic medications.
OBJECTIVE
To develop a natural language processing (NLP) algorithm to identify patients with AERD from an electronic health record (EHR).
METHODS
A rule-based NLP algorithm was developed using clinical documents from the EHR at Mayo Clinic. From clinical notes, seven features were extracted that included the following: AERD, asthma, NSAID allergy, nasal polyps, chronic sinusitis, elevated urine leukotriene E4 level, and documented no-NSAID allergy by a health care provider. MedTagger was used to extract these seven features from the unstructured clinical text given a set of keywords and patterns based on the chart review of two allergy/immunology experts for AERD. The status of each extracted feature was represented as either present or absent per subject. To determine the representative combination of features to discriminate the different AERD features, we utilized the entropy approach of the decision tree classifier.
RESULTS
The AERD-NLP algorithm achieved an accuracy of 88.00% (95% CI=[82.10, 91.74]) and 84.50% (95% CI=[78.73, 89.22]) for the training and test set, respectively.
CONCLUSIONS
We developed a promising AERD-NLP algorithm that needs further refinement to improve AERD diagnosis accuracy. Continued development of NLP and other artificial intelligence technologies has the potential to reduce diagnostic delays for AERD and improve the health of our patients.