AI Adoption in Healthcare Through the Eyes of a 24-Year Orthopedic Innovator

Every orthopedic conference now has an AI track. Every exhibit hall has a booth promising to transform the clinic. AI adoption in healthcare is accelerating, and so is the noise around it.

So how does a surgeon tell real clinical value from a great demo?

Dr. Michael Durkin has been answering that question for more than two decades. A board-certified orthopedic surgeon at the Illinois Bone and Joint Institute (IBJI), he studied bioengineering before medical school, joined Hinsdale Orthopedics in 2002 (which merged with IBJI in 2020), and was one of the early surgeons using Mako robotic technology. On EP 10 of Truly Integrated, he walked iScribe CEO Pat Williams through how he evaluates new technology, from the OR to the exam room.

Prefer to listen? Hear Dr. Durkin’s full conversation on Truly Integrated, EP 10: episode page · Apple Podcasts · Spotify

Key Takeaways

• The clearest early signal of real value in robotic surgery was a narrower bell curve, with fewer outliers and post-op X-rays that matched the surgical plan.
• Value has to show up quickly and visibly for the patient.
• First impressions from training aren’t the final word. Learning from an experienced peer often reveals the real benefit.
• In documentation, the bar is whether AI preserves the physician’s own clinical reasoning and protects the integrity of the record.
• Honest evaluation includes knowing where a technology still falls short.
• Physicians nationally set a similar bar, putting safety and efficacy validation at the top of their list for broader AI adoption.

The “Second Wave” Adopter’s Advantage

Dr. Durkin doesn’t describe himself as a pioneer. “I’m a little bit of an early adopter. I’m kind of that second wave,” he said. The innovators build and test. The second wave looks hard at what the innovators produced and decides what’s worth bringing into everyday practice.

We think that’s a smart posture for AI adoption in healthcare right now. Specialty practices win by being right, and being first matters far less.

The national picture shows why that posture matters. In the AMA’s 2026 Physician Survey on Augmented Intelligence, 81% of physicians said they use AI in their practices, more than double the 2023 rate of 38%. Yet 40% said they are equally excited and concerned about it. Adoption is moving fast. Conviction is still being earned.

What Robotics Taught Him About Real Clinical Value

Dr. Durkin’s robotics story starts with skepticism. Around 2009 and 2010, his local hospital bought a Mako robot, and he was one of only two surgeons doing partial knees. Neither had asked for it. “I was skeptical,” he said. “At that time, I was doing a pretty nice uni, and I liked my uni.”

His first training session was, in his words, “very unimpressive.” What changed his mind was spending a morning with an experienced early user in Fort Lauderdale. That visit delivered the “aha moments” where the clinical benefit to patients became clear.

What did real value look like? “What’s been the difference for me with Mako was just how much faster you could see the benefit to the patient,” he explained. “I noticed early on that the bell curve was narrower. You didn’t have outliers. Your post-op X-rays looked like they were supposed to.”

That’s a useful test for any technology. Does it make outcomes more consistent? Does it remove the outliers? Does it produce what the surgeon planned?

Published research points the same way. In a randomized controlled trial of partial knee replacement published in The Journal of Bone and Joint Surgery, which compared a Mako robotic arm with conventional instruments, 80% of robotic-assisted patients had the tibial component placed within 2° of the target position in the sagittal plane, compared with 22% of conventional patients. The study measured implant position on CT scans at three months rather than long-term outcomes, but it is the kind of consistency Dr. Durkin describes.

He went on to teach and consult on robotic technology, and his use expanded from partial knees to total hips, total knees, and, this year, robotic total shoulders.

A Practical Framework for AI Adoption in Healthcare

Dr. Durkin’s experience translates into questions every specialty practice should ask before adopting AI.

1. Does it narrow the bell curve? Real value shows up as consistency. In documentation, that means every note meets the same standard, on the first patient of the day and the fiftieth.

2. How fast does the benefit show up? The best technology proves itself quickly. iScribe delivers a processed note in roughly 30 seconds, and providers leave the room with a completed note.

3. Who will teach you to use it well? Dr. Durkin’s breakthrough came from an experienced peer. Onboarding quality decides adoption. iScribe’s white-glove onboarding has providers fully set up before go-live, with zero schedule reduction. Our trial-to-customer conversion rate is 92%.

4. Does it preserve your thinking? This is the bar Dr. Durkin set for ambient AI.

5. Where does it still fall short? Every technology has limits. Know them before you commit.

Vendor Hype vs. Real Clinical Value

Dr. Durkin’s questions come down to a simple contrast. Here is how the two tend to look when a new tool shows up at your practice.

• What you see first → Vendor hype: A polished demo → Real clinical value: Consistent results in everyday cases.
• How value is measured → Vendor hype: Feature lists and promises → Real clinical value: Fewer outliers and results that match the plan.
• When the benefit shows up → Vendor hype: Someday, after a long rollout → Real clinical value: Quickly, and visibly for the patient.
• How you learn it → Vendor hype: A one-time training session → Real clinical value: Time with an experienced peer.
• How limits are handled → Vendor hype: Rarely mentioned → Real clinical value: Named upfront.

If everything you know about a tool fits the left column, treat its claims as marketing until your own results say otherwise.

Documentation That Sounds Like the Surgeon

Dr. Durkin is candid that the EMR has never been his strength. “I could never be the EMR doctor who was sitting, typing, you know, maybe occasionally glancing over my shoulder at the patient. I was never going to be that guy.”

Today he uses an ambient AI scribe on his phone, alongside his care team. What he values most is the summary at the end of the visit, in his own words. “I know what I was thinking, and I know what the plan is,” he said. “That is, I think, what hopefully protects me and also preserves, you know, for the patient’s benefit, what I was thinking.”

In our view, that’s the standard for AI adoption in healthcare documentation. The note has to reflect the physician’s clinical reasoning. That’s a meaningful step up from older dictation software, where, as he put it, “the errors and the funny things that it says when you go back and read your notes” were “pretty comical.”

Being Honest About What Still Needs Work

Dr. Durkin was equally candid about where ambient AI still asks something of the physician. Describing the physical exam out loud doesn’t come naturally. “I’m not good yet at describing the physical exam to the AI scribe,” he said, so his scribe or athletic trainer still helps capture it.

That’s an important lesson for any practice. The physical exam is often where documentation quality, and coding defensibility, is won or lost. It’s also why iScribe builds provider-specific macros, specialty-tuned to each surgeon’s exam, so the details that matter are captured the way that surgeon documents them, every time.

The Real Barriers to AI Adoption in Healthcare

Dr. Durkin’s standards match what physicians nationally say they need before using AI more broadly. In the AMA’s 2026 survey, 88% of physicians called robust safety and efficacy validation critical for broader AI adoption, and 86% said the same about data privacy. Clear liability frameworks ranked highest among the regulatory actions physicians said would build their trust.

Most also want a say: 85% want to be consulted or directly involved in decisions about AI adoption. Strong clinical evidence and practical implementation guides were the support resources physicians cited most.

None of that is a request for more features. It is a request for proof, safeguards, and a voice in the decision, which is much how Dr. Durkin approaches any new technology.

Why Independence Raises the Stakes

Dr. Durkin is a strong advocate for large independent orthopedic groups. “I would support the independent orthopedic practice,” he said, “because you don’t want hospitals telling you what to do.”

For independent practices, every technology decision carries real financial weight. Documentation-related denials, coding inaccuracy, and lost clinic hours hit margins directly. That’s why iScribe focuses on precision, speed, and protecting the value of the work for partner-owned specialty groups, with E&M code recommendations at 95% audited accuracy through iScribe Code Intelligence for orthopedic and urology practices on athenahealth or NextGen.

The Bottom Line

AI adoption in healthcare shouldn’t be driven by conference buzz. It should be driven by the same evidence Dr. Durkin looked for in the OR, meaning consistent outcomes, visible value, strong partnership, and honesty about limits. If you evaluate new technology differently, we’d like to hear how.

Frequently Asked Questions

How should practices evaluate AI adoption in healthcare?

Look for consistent, measurable results, a fast path to visible value, high-quality onboarding and support, documentation that reflects the physician’s reasoning, and a clear understanding of current limitations.

What makes ambient AI documentation trustworthy?

It should accurately preserve the physician’s clinical reasoning and plan, integrate with the EHR, and be tuned to the individual provider’s documentation style.

Why does onboarding matter in AI adoption?

Adoption depends on how quickly providers see value. iScribe’s onboarding has providers fully set up before go-live, and 92% of trials convert to customers.

Why is AI adoption in healthcare accelerating?

Physicians see real potential to lighten their workload. In the AMA’s 2026 survey, 70% said they see AI as a tool to automate tasks that contribute to work-related burnout.

How can a practice tell real clinical value from vendor hype?

Ask for evidence from everyday use, not a demo. Look for consistent results, a benefit you and your patients can see quickly, and a vendor that will tell you plainly where the tool still falls short.

Should physicians help choose the AI tools their practice adopts?

Yes. Physicians nationally have said they want to be consulted or directly involved in AI decisions, and the providers who will use a tool every day are best placed to judge whether it fits their work.

Is it better to be an early adopter or to wait?

Dr. Durkin describes himself as a second-wave adopter. Second-wave adopters learn from the innovators, look hard at the results, and adopt once the clinical value is clear, so the goal is to be right rather than first.

About Dr. Michael Durkin

Dr. Michael C. Durkin is a board-certified orthopedic surgeon at Illinois Bone and Joint Institute (IBJI) in Westmont, Illinois. His practice covers joint replacement, including robotic partial knee replacement, along with sports medicine, arthroscopy, and fracture care. He earned a bioengineering degree at the University of Illinois at Urbana-Champaign and his medical degree from the University of Illinois College of Medicine, then completed his internship and orthopedic residency at the University of Illinois Hospitals and Clinics in Chicago. He has been recognized as a Castle Connolly Top Doctor since 2012, previously served as team physician for York, Immaculate Conception, and Willowbrook High Schools, and is a member of the American Academy of Orthopaedic Surgeons.

Hear the Full Conversation

Listen to Dr. Michael Durkin on Truly Integrated, EP 10, “Separating Real Clinical Value from Vendor Hype.”

Hold your documentation to the same standard as your surgery. Start a complimentary 14-day trial of iScribe AI, fully set up before go-live with zero schedule reduction.