Recent research published in New England Journal of Medicine AI shows that large language models (LLMs) can accurately process hospital quality measures, aligning with manual reporting in 90% of cases. The pilot study used an LLM system to analyze Severe Sepsis and Septic Shock Management Bundle (SEP-1) measures at UC San Diego Health, demonstrating significant potential for streamlining the traditionally labor-intensive reporting process.
By automating data abstraction from EHRs, LLMs can reduce manual workload, speed up processing times, and lower administrative costs, paving the way for more efficient quality reporting in healthcare. Researchers aim to further validate and refine this approach to improve patient care outcomes and data access.