HealthMore in Health→
AI Language Models Turn Unstructured EHR Notes Into Computable Patient Journeys
Key Takeaways
- Most EHR data sits in unstructured encounter text, yet most real-world evidence studies still rely on structured fields such as diagnosis and procedure codes.
- The paper describes a framework using large pre-trained language models to extract computable clinical data from unstructured EHR text.
- Extracted entities are integrated with structured records, embedded with medical ontologies and organized into a knowledge graph that links all variable relationships.
- Physician adjudication confirmed high accuracy against a blinded expert reference standard, with high inter-reviewer agreement among expert physicians.
DE
DT Editorial Team··via nature.com