That “increasing role,” needless to say, took decades to materialize, but only in part because of the capabilities of computers. As early as the late 1960s, informatics pioneers Warner Slack famously began reassuring physicians fearful of change, “Any doctor who can be replaced by a machine should be replaced by a machine.” Of course, in today’s era of cloud computing those fears no longer apply. These days, doctors are afraid of being replaced by apps and algorithms.
In a study published earlier this year, researchers gave chest X-rays and diagnostic advice from human experts to a group of radiologists. They labeled that advice as coming either from humans or AI. While some of the advice was deliberately inaccurate, the overriding factor for radiologists in flagging inaccuracy was where they thought the advice originated.
“As a group, radiologists rated advice as lower quality when it appeared to come from an AI system,” the researchers concluded.
Sometimes, of course, the problem is not anti-algorithm bias, but justified hostility to the tone-deaf manner in which digital health technology is sometimes injected into the clinical workflow – what Barry delicately called “point-of-care integration.”
Consider what happened when Duke University Health System implemented a home-grown AI tool to predict which patients might develop sepsis, a potentially fatal complication. It took two years, and a study involving medical anthropologists, for the health system to discover that nurses were being forced to present high-priority alerts to doctors they’d never worked as a result of a tool whose workings neither group really understood. The result was a dysfunctional workflow. “Nurses [were] uncomfortable and doctors defensive,” noted a STAT News article.
To address these types of issues, a group of Stanford researchers recently called for development of “a delivery science for artificial intelligence in health care.” That science, they say, will require a “broader set of tools, such as design thinking, process improvement and implementation science” to enable human and technical components to work together.
Medical futurist Bertalan Meskó, a physician and researcher, predicts that doctors will eventually follow the path carved out by grand masters in chess. Meskó sums up the coming transformation – which he carefully notes could be a lengthy one – by citing computer chess developer Monroe Newborn: “Masters used to come to computer chess tournaments to laugh. Now they come to watch. Soon they will come to learn.”