OpenEvidence grows as more doctors use AI to answer clinical questions

In short: More clinicians are turning to OpenEvidence, an AI tool that summarizes medical research and guidelines into referenced answers for diagnosis and treatment questions.

What's going on

OpenEvidence is a clinical decision-support platform, which means it is a tool made to help doctors decide what to do, not to replace them. Doctors can type a question in plain language, and the system returns a short, structured answer with citations. Think of it like a research assistant that reads a large stack of medical papers and then shows its sources.

The company behind it, OpenEvidence Inc., is based in Cambridge, Massachusetts, and is led by founder and CEO Dr. Daniel Nadler. The product launched in 2023 and was developed with support from the Mayo Clinic Platform Accelerate program. OpenEvidence says it pulls from peer-reviewed sources like PubMed, major medical journals, and well-known guideline groups.

Evidence about how well it works is still limited. A 2025 study in a primary care clinic tested OpenEvidence on five real patient cases involving common chronic conditions. The authors reported that the tool’s recommendations matched the decisions physicians made, and that doctors rated the outputs as clear and well supported by cited research, but it did not change the original decisions.

Some adoption numbers, such as claims that over 40 percent of US physicians use it daily and that it is embedded in more than 10,000 hospitals, come from industry write-ups and are not independently verified in peer-reviewed research.

What to watch

AI systems can make mistakes, including “hallucinations” (confident statements that turn out to be wrong), so the key question is how often clinicians verify the cited sources before acting. It also matters how tools like this get built into paid products, such as Elsevier’s ClinicalKey AI, and how hospitals set rules for safe use.

Source: NYTimes