Webinar: Healthcare Use Cases from Linguamatics - Simon Beaulah
Simon Beulah discussed three focus areas for Linguamatics as they work with clients in the healthcare payer and provider groups: operationalizing NLP, real world evidence, and precision medicine. His talk covered work done by Linguamatics and IQVIA, with illustrations from client projects described in this Virtual NLP Summit and elsewhere.
A key concept in operationalizing NLP is the data factory, where NLP is used in mining and supporting multiple business lines. Clients are building out workflows that are fully automated where possible, receiving multiple disparate data sets, processing them with NLP, and sending the results onto a data warehouse, where they can be used, for example, to support predictive models, and to look at high risk individuals. Simon cited an example where this approach identified a significant care gap. Kaiser Permanente analyzed echo reports to look for insights into aortic stenosis. They trained the model on three gold standards, and when they applied it retrospectively to a million echo reports from 10 years of EHR data, they found that ca. 66% of those patients were not being properly coded for aortic stenosis.
Simon closed by stressing that AI and NLP are absolutely key to process the vast amount of data that we see in unstructured form in medical records, PDF charts, the scientific literature, or clinical trials.gov, etc., as they are crucial pieces of information that life sciences and healthcare need to make much better use of.