Researchers have developed a way to quantitatively measure physicians' clinical reasoning abilities using artificial intelligence and machine learning, according to a study published in JAMA Network. The research team engaged 1,738 physicians, residents and medical students for the study, during which the participants used software from the Human Diagnosis Project to solve case simulations. The researchers used automated techniques to measure the participants' performance,based on three categories: accuracy, efficiency and a combined score known as Diagnostic Acumen Precision Performance. Their performance also was weighed against individual level of medical training. Attending physicians earned the highest scores in accuracy using the AI tool, followed by residents, interns and medical students, respectively. Attending physicians also earned the highest scores in efficiency, followed by residents, interns and medical students. Moreover, attending physicians from a U.S. News and World Report-ranked institution scored higher than nonaffiliated attending physicians. "The data suggest that diagnostic performance is higher in those with more training and that [Diagnostic Acumen Precision Performance] scores may be a valid measure to appraise diagnostic performance," the study authors write. "This diagnostic assessment tool allows individuals to receive immediate feedback on performance through an openly accessible online platform." To access the complete study in JAMA Network, click here.
