How US Physicians are Using AI?

|

,

Reporter Jared Perlo from NBC News reports in a recent article that nearly two-thirds of U.S. physicians, roughly 650,000 doctors, are increasingly relying on OpenEvidence, an AI-powered medical search tool, to guide their clinical decisions, with another 1.2 million users abroad. According to NBC, the tool was used across almost 27 million clinical encounters in April alone, offering quick, tailored responses while also able to write patient discharge notes and provide custom study aids for medical exams. Perlo spoke with more than two dozen doctors, administrators, students, and researchers from Hawaii to Maine, and every one of them either used the tool or knew someone who did.

Below are key highlights from the article:

  • What OpenEvidence does. It is a free, AI-powered medical search engine that combs through peer-reviewed research and clinical guidelines to answer doctors’ questions and link back to the source papers. Sign-up requires a government-issued healthcare ID, and the service is currently funded by ads.
  • Why doctors adopt it so fast. Dr. Anupam Jena of Massachusetts General Hospital, who is studying 90 million OpenEvidence queries, told Perlo that about 60 percent of searches are about clinical decisions. Dr. Jeremy Cauwels of Sanford Health called it remarkably easy to adopt. Dr. Paul Sax of Brigham and Women’s Hospital said it often “borders on miraculous” and outperforms the long-standing reference UpToDate, which is now rushing to launch its own AI tool, Expert AI.
  • What makes it different. OpenEvidence has licensing deals with top journals such as the New England Journal of Medicine and JAMA, plus groups like the National Comprehensive Cancer Network. CEO Daniel Nadler told Perlo the company treats AI as “search glue,” using full-text access rather than letting the model invent answers.
  • The privacy question. OpenEvidence says it is HIPAA-compliant, but some systems, such as MaineHealth, ask doctors not to enter protected health information.
  • Real concerns. A December academic study found OpenEvidence answered complex medical questions accurately less than 45 percent of the time. Doctors told Perlo it sometimes flubs rare cases or overreads small studies. A Missouri physician worries that medical students leaning on the tool are losing the skill of sorting signal from noise.

Perlo notes that doctors have always had to adopt new tools, and Mount Sinai’s head of AI, Dr. Girish Nadkarni, argues it is better to bring this kind of “shadow AI” into the open than to ignore it. Mount Sinai recently linked OpenEvidence into its electronic health record system across its 47,000 employees.

Acknowledgement: This summary was generated using AI writing tools as part of my AI Explorations process. For more information, please refer to the original article listed below.

Related Resource Link