“The study documents how AI is already being used effectively in many health care processes, but the current so-called point solutions, where AI is incorporated into specific and established applications, offer only marginal improvements.
“Crucially, the study argues that substantial productivity gains in health care from AI adoption will happen only when the entire health care value chain—and not just specific individual activities—is redesigned with AI at its core.
“For example, AI “scribe” is a tool already in use to automatically generate clinical notes from physician–patient conversations. While this saves physicians time that would otherwise be spent typing information into electronic health records, a system-wide solution might, for example, see AI unobtrusively record conversations in the exam room, structure the clinical notes, update the medical records, arrange follow-up appointments, coordinate with the pharmacy, and flag relevant test results.
“Similarly, in medical research and discovery, machine learning tools are already used to search prior literature, perform statistical analysis, and sift through molecular databases. An example of a system-wide solution could be a “self-driving lab” where AI actively designs, runs, and adapts entire experiments in real time without human intervention, accelerating the search for new drugs and optimizing treatment protocols.
“Artificial intelligence—and in particular system solutions with AI at their core—holds great promise for health care,” Goldfarb said.
““Whether the health-care system can realize this potential will depend on whether health care leaders are willing to embrace system-level redesign, while keeping patient health as their core mission.”
“Media Contact:
Steven Globerman, Senior Fellow, Fraser Institute
To arrange media interviews or for more information, please contact:
Drue MacPherson, Fraser Institute
604-688-0221 ext. 721
drue.macpherson@fraserinstitute.org
