A Boston-based Harvard Medical School study found oncologists unsure how to handle artificial intelligence if it presents a treatment plan they would not recommend.
The study, published in March in JAMA Network Open, surveyed 204 oncologists between Nov. 15, 2022, and July 31, 2023. Of participants, 84.8% said AI-based clinical decision models needed to be explainable by oncologists but did not need to be explained to patients; and 81.4% said patients should consent to the use of AI to support treatment decisions.
However, the study also found that if AI selected a treatment regimen different from what an oncologist would recommend, 36.8% said they would present both options and let the patient decide. This was the most common response among oncologists.
"Interestingly, while oncologists did not think patients needed to be able to explain AI models, when we presented them with a scenario in which AI disagreed with their treatment recommendation, the most common response was to present both options to the patient and let them decide," study author Andrew Hantel, MD, an instructor in medicine at Harvard Medical School and a faculty member in the Divisions of Leukemia and Population Sciences at Dana-Farber Cancer Institute in Boston, said in a July 4 analysis. "This finding highlights that many physicians are unsure about how to act in relation to AI and counsel patients about such situations."
The study highlights the differences in opinion among oncologists and underscores the need for broader discussion on how the responsibility of decision-making will be shared between oncologists and AI-assisted tools, Shiraj Sen, MD, PhD, a medical oncologist at Texas Oncology, said in the analysis.