New AI models have shown promise in the healthcare industry. But they still need some fine-tuning, doctors say.
Atropos Health co-founder and Chief Medical Officer Dr. Saurabh Gombar sat down with Alexis Kayser of Newsweek to discuss ChatRWD and Generative AI. Read to full article here
Over the last year, asking LLMs medical questions hasn’t always returned trustworthy or complete answers. That’s why Atropos Health built ChatRWD™.
After years of delivering just-in-time, publication grade, real-world evidence studies in under two days, Atropos Health developed a product that can do it in minutes by combining the latest LLMs with GENEVA OS™ (Generative Evidence Acceleration Operating System), the underlying platform containing the phenotype library.
Atropos Health recently released a paper on the results of ChatRWD beta. ChatRWD beta outperformed ChatGPT 3.5, Google Gemini 1.5 Pro. and Anthropic Claude 3 on answers to clinical questions when reviewed by independent physicians. OpenEvidence also outperformed the “Big Tech” LLMs by a wide margin. Most importantly, ChatRWD performed highest on physician trust and providing the best answer and was able to answer questions asked 94% of the time.
As the Gen AI cycle continues to move at a rapid pace, trust and accuracy are going to be critical, and publishing performance on both is table stakes. It is becoming clear that the abliity to answer 100% of questions with clinical grade accuracy by training LLMs solely on existing literature will not be enough. Producing new evidence in the form of high-quality transparent studies using appropriate methodology is needed to move healthcare forward.