Artificial intelligence-enabled tools allow clinicians to ask questions like, "How should I explain cholesterol levels to a patient?" or "What were the outcomes of the ATLANTIS trial?" and receive synthesized responses in seconds. However, data on which clinicians use these tools, how often they use them and for what purposes has remained limited.
A new study led by Vishal Patel, MD, MPH, and overseen by Anupam Jena, MD, PhD, both of Mass General Brigham and Harvard Medical School, helps close these knowledge gaps. The researchers analyzed clinicians' use of OpenEvidence, the most widely used AI-powered clinical platform in the United States.
Between January 2024 and June 2026, active users, query volume and user retention increased dramatically. Overall, nearly 952,000 clinicians submitted approximately 271 million queries through the platform.
OpenEvidence was used widely across the clinical workforce but was especially popular among nurse practitioners, physician assistants, early-career physicians, residents and fellows. Use was also higher among physicians in oncology/hematology, internal medicine and family medicine, and lower among those in ophthalmology, radiology, orthopedics and anesthesiology.
Researchers also analyzed patterns in query purpose and medical topic area. Most queries related to treatment and management decisions rather than diagnosis, and clinicians frequently sought information outside their primary specialty.
By providing one of the first large-scale views of how clinicians use a clinical AI platform in practice, the study may help inform future efforts to evaluate and integrate AI tools in clinical medicine.
Published in JAMA on October 15, 2026 | Read the paper: " Clinician Use of an Artificial Intelligence–Enabled Clinical Information Platform "
Summary reviewed by: Vishal Patel, MD, MPH, lead author
