Mass General Brigham Researchers Explore Hybrid Approach to Diagnosing Patients
Mass General Brigham researchers have been delving into the world of artificial intelligence, specifically focusing on a hybrid approach that combines generative AI with a diagnostic decision support system to enhance patient diagnoses. In a recent study comparing two large language models (LLMs) – OpenAI’s GPT-4 and Google’s Gemini 1.5 – with their own diagnostic system, DXplain, researchers found that while the DDSS outperformed the LLMs in accuracy, combining the two AI technologies could lead to more informed treatment decisions.
Evolution of DXplain
DXplain, initially developed in Boston in 1984 as a standalone platform, has since evolved into a web-based application and cloud-based differential diagnosis engine. With over 2,680 disease profiles, 6,100 clinical findings, and a vast database of data points, DXplain generates and ranks potential diagnoses based on user input. The system has proven to be a valuable tool in the diagnostic process.
Study Methodology
For their study, researchers prepared a collection of 36 diverse clinical cases based on real patients from academic medical centers. Three physicians manually assessed these cases, inputting clinical findings into the DDSS, LLM1 (ChatGPT), and LLM2 (Gemini) for comparison. The study aimed to evaluate the performance of each system in accurately diagnosing patients.
Findings and Insights
The study revealed that the DDSS performed better when all laboratory test results were included in case reports, listing the correct diagnosis 72% of the time compared to LLMs. While LLMs like ChatGPT and Gemini showed promise in generating accurate diagnoses, they lacked the explanatory capabilities of the DDSS.
Researchers highlighted the potential benefits of combining the linguistic capabilities of LLMs with the deterministic and explanatory capabilities of traditional DDSSs. This hybrid approach could lead to synergistic benefits, improving the overall clinical efficacy of both systems.
Future Implications
With the ongoing advancements in AI technology, the healthcare industry is likely to see a shift towards utilizing hybrid AI systems to enhance patient care. By leveraging the strengths of both generative AI and diagnostic decision support systems, healthcare providers can make more informed treatment decisions and improve diagnostic accuracy.
As the field of AI in healthcare continues to evolve, researchers are optimistic about the potential of hybrid approaches to revolutionize patient care and diagnostic processes.
This article was written by a Senior Editor of Healthcare IT News and is a HIMSS Media publication.
