Mass General Brigham Study Finds Synergy Between AI Models and Traditional Diagnosis Tools

Researchers at Mass General Brigham (MGB) evaluated the integration of large language models (LLMs) like OpenAI's GPT-4 and Google's Gemini 1.5 with traditional diagnostic decision support systems (DDSS) to improve patient diagnoses. In comparing these LLMs to MGB's long-established diagnostic tool, DXplain, which uses a comprehensive database for generating potential diagnoses, DXplain demonstrated superior performance with a 72% accuracy rate compared to 64% for ChatGPT and 58% for Gemini. The findings suggest that while DXplain was more effective at providing differential diagnoses, particularly with laboratory data, incorporating LLMs could enhance clinical decision-making by offering additional insights.

Opens on Healthcare IT News