A recent study conducted by researchers from Cedars-Sinai, Tel Aviv University, and the digital health company K Health has suggested that artificial intelligence (AI) can compete with doctors in clinical decision-making, particularly in urgent care settings. The study, published in the Annals of Internal Medicine, compared the recommendations made by K Health’s AI chatbot with those from physicians during virtual appointments for various acute symptoms.
The findings indicate that the AI matched the clinical decisions of doctors in approximately 68% of cases. Furthermore, in instances where the AI's recommendations differed from those of physicians, the AI was reportedly more effective in adhering to established medical guidelines. However, experts noted that the AI's performance diminished in situations requiring nuanced clinical judgment.
The study analyzed 461 patient visits between June and July 2024 and involved external physician reviewers who assessed the quality of recommendations made by both the AI and the doctors. The AI produced potentially harmful recommendations 2.8% of the time, whereas the physicians had a higher rate of 4.6%.
Dr. Caroline Goldzweig from Cedars-Sinai emphasized that while the AI can be a valuable tool in straightforward cases, it is not expected to replace human doctors, particularly in complex scenarios involving multiple health issues. K Health's co-founder, Allon Bloch, expressed optimism about the AI's potential to enhance patient care and improve efficiency for healthcare providers.
As AI technology continues to evolve, there is an increasing interest in integrating it into healthcare systems, especially given the ongoing shortage of primary care physicians. K Health has already garnered significant funding and established partnerships with major health institutions, aiming to explore further applications of AI in clinical settings.