How MedVanta ensures real results from AI initiatives

By: Cameron Cortigiano

Fighting through the AI hype to find systems that actually provide tangible benefits is a point of emphasis for AI initiatives at this musculoskeletal technology firm. 

MedVanta, the largest physician-owned, fully integrated musculoskeletal platform in the U.S., which was founded by the physicians behind Bethesda, Md.-based The Centers for Advanced Orthopaedics, is thinking about the return on investment in a different light.

Determining the goal of adding AI ahead of time is crucial. Without a well-thought-out plan in place, any integration will more than likely flop.

“I think many organizations are chasing AI hype instead of focusing on clear outcomes and deliverables,” Tony Yi, chief technology and information officer of MedVanta told Becker’s. “You have to have a plan, and oftentimes they don’t. A lot of times they’re just jump-starting it without having a clear-cut roadmap in terms of what they want.”

Organizations can’t view AI as an overnight fix or a way to stop a leaky pipe. The top successes of AI augment something that is already established.

“When you adopt AI into a very fragmented system or an operationally inefficient organization, the initial phases of AI implementation typically replicate and automate those inefficiencies,” said Mr. Yi.

Instead of looking at the success of a technology as a straight-up return on investment, MedVanta looks at the value it provides to physicians and patients, which doesn’t always result in directly improving the bottom line.

Creating value can be something as simple as an agentic AI taking phone calls. That program allows the physician to spend more time treating patients.

“When we assess AI initiatives given to us by these physicians, we don’t necessarily just look at ROI, but we look at the value of the investment,” said Mr. Yi. “Some examples of the VOIs we look at are things like, ‘How much physician burden does this reduce?’ or ‘How much patient satisfaction would this potentially increase?’ We take a multistep approach, ensuring it solves a problem, assessing how feasible it is to implement and understanding what we’re trying to accomplish.”