A Large Language Model Application for Specialty-Specific Care Optimization

Large Language Models (LLMs) are a type of Artificial Intelligence (AI) technology with the potential to transform healthcare delivery. LLMs can generate answers to questions and create summaries of large amounts of text that are often indistinguishable from those a human would create. Leveraging these powerful capabilities, we developed and deployed an LLM-based clinical note summarization system that enables clinicians to rapidly extract concrete, specialty-specific information from the vast amount of notes in an electronic medical record (EMR) system. This presentation will: (1) provide an overview of the development, validation and deployment of our LLM tool; (2) reveal the innovative strategy we developed to create responses most applicable to different clinical specialties; (3) discuss the safeguards we have put into place to ensure the information returned by our LLM tool is accurate, precise and reliable; and (4) present data on efficiency improvements we have observed since deployment of our LLM tool.