Dr. Anupam Jena at Massachusetts General Hospital watched his colleagues reach for their phones not to scroll social media, but to consult an AI tool that has become as embedded in American medicine as stethoscopes—only most patients have no idea it's happening. OpenEvidence, a clinical decision-making platform that functions as a kind of expert colleague on call, has grown from a curiosity into the quiet backbone of American medical practice. In April alone, roughly 650,000 U.S. doctors—about 65% of the nation's physician workforce—turned to OpenEvidence across nearly 27 million clinical encounters.
The tool arrived without fanfare, spreading through hospitals and private practices across every state from Hawaii to Maine with the efficiency of something genuinely useful. At Sanford Health, the country's largest rural healthcare system serving over 2,500 providers in the Upper Midwest, Dr. Jeremy Cauwels describes OpenEvidence as "remarkably easy to adopt." It's free. It works seamlessly on phones. And it answers clinical questions faster than traditional methods. A junior doctor in New Hampshire needed to confirm whether a patient's dangerous potassium drop was a normal medication side effect or a new emergency—OpenEvidence provided the answer and several treatment options within seconds. A physician at a rural Indian Health Service center in South Dakota consulted the tool to determine whether a spine fracture required a CT scan instead of just an X-ray. These are not dramatic moments, but they are the steady substance of good medicine: checking your thinking against current evidence when it matters most.
What's remarkable is the scale. Jena, a Harvard healthcare policy professor, is now analyzing 90 million OpenEvidence queries submitted since 2024 as part of an emerging research project. Sixty percent of those searches focus on clinical decision-making—doctors asking the system: Given this patient, this condition, these comorbidities, what's the right treatment? The growth has been "exponential," Jena says. Internationally, another 1.2 million doctors use OpenEvidence, making it one of the most widely adopted medical AI tools on Earth, though one that operates almost entirely below public consciousness.
Yet this quiet expansion unsettles some experts. They worry about hallucinations or incomplete answers, the absence of rigorous studies proving the tool improves patient outcomes, and the risk that doctors' critical thinking skills may atrophy through over-reliance. OpenEvidence's own terms of service emphasize that the platform should supplement, never replace, clinical judgment. The company also complies with HIPAA privacy regulations, though some health systems remain skeptical about overall safeguards.
But across specialties and geographies, physicians express something closer to gratitude. The tool helps them work faster, think clearer, and stay current with medical literature that would otherwise demand hours of searching. In an era when physicians face burnout and time pressure, OpenEvidence offers something medicine has always needed: a way to check your instincts against evidence when you need it most. The question now isn't whether AI will transform medicine—it already has—but whether we can build transparency and accountability around tools that are already reshaping how millions of patients receive care.