A novel method of studying the epidemiology of allergic drug reactions
Published online: December 17, 2019
An accurate understanding of the epidemiology of allergic drug reactions is important in healthcare, both for improving patient care as well as guiding public health and preventive medicine efforts. However, allergic drug reaction epidemiologic data are sparse because it remains difficult to identify true cases in large datasets using manual chart review. The rapidly increasing use of electronic health records in combination with automated methods that use computer-based algorithms, such as natural language processing, may prove useful in reducing the magnitude of chart review needed. Building on this group’s prior research experience using ICD-9-CM codes, Banerji et al., developed and validated a novel informatics method based on natural language processing to identify allergic drug reactions from unstructured clinical notes and discharge summaries in their electronic health records system.
In this month’s Journal of Allergy and Clinical Immunology: In Practice, Banerji et al., describe a novel informatics research tool, natural language processing in combination with diagnosis codes, that improved the ability to study the epidemiology of allergic drug reactions using the electronic health record. Previously studied and high-yield ICD-9-CM codes were used to screen for possible allergic drug reactions among all inpatients admitted in 2007 and 2008 at a large academic hospital. A random sample was then selected for manual chart review to identify true cases of allergic drug reactions. A rule-based natural language processing algorithm was developed to identify allergic drug reactions using free-text clinical notes and discharge summaries from the filtered cases. The performance of using manual chart review of ICD-9-CM codes alone was compared to ICD-9-CM codes in combination with natural language processing.
Banerji et al., developed and validated a novel natural language processing algorithm for extracting information from clinical free-text documents throughout the electronic health record, including full note search. They started with a lexicon of reaction terms and used encounters from the years 2005 and 2010 to train the algorithm. The algorithm was subsequently evaluated against the current gold standard, manual specialist chart review, and showed a high performance for identification of allergic drug reactions in cases stratified by ICD codes with an overall sensitivity of 77% and specificity of 89%. These results suggest that natural language processing has high utility in studying the epidemiology of allergic drug reactions, especially considering the expanding use of electronic health records.
The Journal of Allergy and Clinical Immunology: In Practice is an official journal of the AAAAI, focusing on practical information for the practicing clinician.
Full Article