Overview
Understanding biomarker testing results, such as PD-L1, in medical records can be challenging due to their unstructured nature and evolving documentation practices over time. This study aims to assess the effectiveness of advanced language models in promptly identifying key PD-L1 details within a comprehensive medical record database, spanning various cancer types. The objective is to determine whether the models can accurately extract essential information despite potential differences in how physicians document these details.
Why this matters
This research underscores the potential of cutting-edge technology, such as large language models, in expediting and enhancing the retrieval of critical information from medical records. By effectively utilizing these models to extract essential details concerning PD-L1 testing across various cancer types, the study demonstrates their utility in facilitating prompt access to vital information for informed treatment decisions by healthcare professionals and researchers. This advancement holds promise for optimizing patient care and outcomes in clinical settings.
View the abstract on the ISPOR website
More from Flatiron Health at ISPOR US 2024
Read about Flatiron's accepted research and presence at ISPOR US.
