In a study of 461 virtual primary care visits at Los Angeles-based Cedars Sinai, AI-powered clinical recommendations were often rated as superior to those made by physicians, according to research published April 4 in Annals of Internal Medicine.
Researchers from Tel Aviv University, Cedars-Sinai Medical Center and K Health evaluated the use of K Health’s clinical AI platform at Cedars-Sinai Connect, the health system’s virtual primary and urgent care clinic.
The AI system, trained on a large volume of real-world clinical data, assists with patient intake and generates diagnostic and treatment recommendations for common acute issues, including dental, eye, respiratory, urinary and vaginal complaints.
Physicians had access to the AI-generated suggestions, although researchers could not confirm whether those inputs were reviewed before final decisions were made.
In about two-thirds of cases, physicians made the same recommendation as the AI. The adjudicators rated the AI’s recommendations as superior twice as often as they rated them inferior. Adjudicators also found the AI’s decisions were less likely to be potentially harmful.
