Mayo CEO says AI could scale its care model nationwide

The latest:

Mayo Clinic is trying to show the federal government and other hospitals how to copy its AI-assisted care system, President and CEO Dr. Gianrico Farrugia told Axios. He said the model depends on organizing health data so both physicians and AI agents can use it — and that most hospitals could not deploy the technology even if handed it for free.

Details:

• The technology: According to Axios, Mayo doctors and patients draw on large medical databases sorted and analyzed by AI, then channeled into roughly 500 separate algorithms used to identify treatments and guide care. Farrugia described it as a working fusion of clinicians and machines applied to patients with comparable conditions.
• The Microsoft deal: Mayo recently signed an agreement with Microsoft to expand and speed up its own frontier AI model, Axios reported. Both are fine-tuning the system with the stated hope of sharing it broadly. Terms of the agreement and a timeline for wider release were not disclosed.
• The obstacle: Farrugia said the barrier is not the software but the recipients. “I could give this to every hospital — and they can’t use it,” he told Axios, citing most hospitals’ lack of technology, funding, systems and focus to exploit patient and population-level health data.
• His policy ask: Farrugia said the federal government needs to force hospitals to change using carrots and sticks, and that a new national data architecture must be built as soon as possible. He did not name a specific agency, legislative vehicle or funding source for that overhaul.
• The care model: Axios reported Mayo physicians work in teams and are paid to heal rather than by procedure or prescription, while nurses handle about four patients each. Appointments begin on time, imaging and blood tests are scheduled immediately, and results post instantly to patient portals.
• The cost question: A variety of studies show Mayo is cost-competitive, according to Axios, and the clinic works with private insurance as well as Medicare and Medicaid. The column did not identify the studies or quantify the comparison against other academic medical centers.
• The contrast drawn: VandeHei wrote that his wife Autumn spent nearly a quarter of the past four years in emergency rooms and hospitals in Washington, D.C., managing three chronic conditions, often waiting double-digit hours on hallway stretchers despite the city’s top-rated hospitals.
• The system critique: He described specialists unwilling or unable to consult across domains, overworked nursing staff, and medical records scattered across unconnected portals with no one assembling them. The U.S., he noted, spends twice as much on medical care as rival nations.
• The target group: Farrugia said that done properly, the new data architecture would be a life-changer for chronically ill patients, who VandeHei argued receive the least coordinated attention. Patients with multiple simultaneous conditions requiring several specialties are the clearest test case.

Between the lines:

VandeHei’s argument rests on a gap he reports himself: his family had blue-chip insurance, system knowledge and resources, and still hit walls. If that profile struggles, the scattered-records problem he describes is structural rather than a matter of patient effort. Farrugia’s own admission that hospitals cannot use the tool suggests the constraint is capital and data plumbing, not AI capability.

What’s next

Watch whether Mayo and Microsoft release their fine-tuned model beyond Mayo, and whether federal regulators take up Farrugia’s call for data-architecture rules backed by incentives and penalties.