Electronic health records can help identify patients with pulmonary embolisms (PEs) and deep vein thrombosis (DVTs) with "high accuracy" according to a new study in the Journal of the American Medical Informatics Association. The study used statistical natural language processing and "bag of words" representation to a random sample of 2,000 narrative radiology reports with suspected DVT and/or PE from McGill University Health Centre from 2008 to 2012.
The researchers found that this "novel" approach worked well to identify potential adverse events. They suggested that using EHR for this purpose could enable clinicians to avoid manual chart review for potential PEs and DVT, which is time consuming and expensive, and thus impractical for routine monitoring for these conditions, which have high hospitalization and mortality rates. Study