A new study published in the Journal of the American Medical Association (JAMA) has uncovered an unexpected trend in how doctors are using AI. The research, led by Harvard Medical School and Brigham and Women’s Hospital, found that physicians are asking AI more than twice as many questions about treatment decisions as they are about diagnosis.
The study analyzed nearly 39 million queries submitted by over 900,000 US clinicians to OpenEvidence, a medical AI tool, over two and a half years. Treatment and management questions accounted for 33.1% of physician inquiries, while diagnosis-related questions made up only 13.9%. Doctors frequently sought AI’s help in areas outside their expertise, with emergency department physicians leading at 88%, compared to 35% among oncologists and hematologists.
Experts highlighted the distinction between accessing specialized information and having the expertise to apply it. Aili McConnon, a Tech News Writer at IBM, noted that AI is shifting focus from diagnosing conditions to guiding treatment decisions. Sandhya Iyer, Associate Partner for AI and Analytics at IBM, pointed out that AI could help emergency physicians make faster decisions when consulting specialists isn’t immediately possible.
The study did not explore how doctors use AI’s answers, leaving questions about whether it replaces specialist consultations or serves as a preliminary step. Vishal Patel, the study’s lead author, emphasized the need for further research on the tools’ accuracy and impact on patient care.