389P A framework for evaluating performance of LLM-based extraction from the electronic health record across different healthcare systems

Abstract

Large language models (LLMs) are transforming real-world data curation in oncology, but their adoption introduces new quality challenges. The Flatiron Health Validation of Accuracy for LLM/ML-Extracted Information and Data (VALID) framework provides a comprehensive approach to evaluating data quality. Applying this framework across different healthcare systems requires accounting for differences in language, treatment patterns, and documentation that may make extracting certain clinical information inherently more complex in some countries than others.

Ähnliche Arbeiten