The Myth of Healthcare's Resistance to Change

I keep hearing the same narrative: AI pilots are failing in healthcare because its an old-fashioned, slow-to-adapt, highly regulated industry resistant to change.

Thats not just wrong. Its backwards. No industry changes faster than healthcare when the change actually works. The problem isnt that healthcare wont change. Its that most AI pilots haven't earned the right to change it.

Let me tell you what real change looks like.

October 1994

I was a medical student in the 1990s, drawn to the controlled chaos of the trauma room. Theres something about that environment, the practiced choreography of saving lives, that felt like the purest expression of medicine.

I spent hours trailing trauma surgeons in the ED, riding with EMS teams in the field. My role was simple, almost ritualistic: when a trauma patient came through the doors, Id open two 14-gauge IV lines, thick as pencils, and start pumping fluids into them as fast as gravity would allow.

Everyone did this. Everywhere. It was gospel.

The logic was intuitive: the patient is bleeding, losing volume. Replace the volume. Keep the pressure up. Buy time for the surgeons to stop the bleeding. It had been the standard for decades. No one questioned it.

Then, in October 1994, Bickell and Mattox published a study in the New England Journal of Medicine that turned everything upside down.

They showed that our gospel was killing people.

The aggressive fluids wed been pumping in were raising blood pressure just enough to blow out the bodys fragile, forming clots. They were diluting the bloods ability to clot at all. We were, in essence, helping patients bleed to death faster while congratulating ourselves for supporting them.

The paper was published in October.

Almost overnight, the standard of care that had stood for decades was defenseless, and trauma teams around the world started to change their protocols.

Not after committees. Not after pilot programs. Not after years of debate.

Within weeks, a practice that had defined trauma care for generations was abandoned because the evidence showed harm.

The Speed of Belief

This wasnt an isolated case. Its how healthcare has always worked when the evidence is real and the benefits to the patients are clear.

In 2002, the Womens Health Initiative published findings that hormone replacement therapy, prescribed to millions of women for decades, actually increased the risk of heart disease, stroke, and breast cancer.

Prescriptions dropped dramatically within months.

Millions of women stopped taking their pills because the data showed harm. Thousands of doctors rewrote their practice patterns because continuing the old approach became indefensible. An entire pharmaceutical market evaporated almost overnight.

Or consider CPR.

For generations, everyone learned the same sequence: A-B-C. Airway, Breathing, Circulation. You tilted the head back, pinched the nose, gave two rescue breaths, then started chest compressions. It was drilled into every medical student, every EMT, every workplace first aid course.

Then researchers noticed something troubling: bystanders werent doing CPR. Theyd see someone collapse and freeze. When asked why, the answer was consistent: they were terrified of the breathing part. The intimacy of it. The hygiene concerns. The complexity of getting it right.

So they did nothing. And people died because of friction in the workflow.

Meanwhile, other research was showing that those interruptions, stopping compressions to give breaths, were themselves deadly. Every time you stopped pressing, the pressure in the heart dropped to zero. It took several compressions to build it back up. You were essentially restarting the engine over and over.

In 2008, the American Heart Association did something radical based on this evidence: they told bystanders to skip the breathing entirely. Just push on the chest. Hard and fast. Dont stop.

By 2010, theyd rewritten the entire professional algorithm. A-B-C became C-A-B. Compressions first.

Survival rates improved because the new approach removed friction and matched the physiology. Bystander intervention skyrocketed because the workflow became simpler.

The lesson: by removing the most complex, most unpleasant part of the workflow, they unlocked action and saved lives.

The Gallbladder Revolution

Sometimes change doesnt come from a study. Sometimes it comes from patients who recognize undeniable value.

For most of surgical history, removing a gallbladder meant a large incision across the abdomen, a week in the hospital, and weeks more of recovery. It was major surgery, with all the pain and risk that implied.

Then, in the late 1980s, surgeons figured out how to do it through tiny incisions using a camera. Laparoscopic cholecystectomy. Patients could go home the next day with minimal scars.

This technique went from experimental to standard in roughly three years.

Not because administrators mandated it. Not because of billing incentives. Because patients learned about it and demanded it. They asked their doctors why they couldnt have the better option. Surgeons who didnt learn the technique fast enough lost their practices.

The change happened because the value proposition was undeniable: less pain, faster recovery, better outcomes. Once patients knew it existed, the old way became indefensible.

The Pattern Repeats

These arent isolated examples. Theyre the pattern.

Telehealth went from niche to commodity within months during the pandemic. Robotic surgery now occupies prominent space in operating rooms that didnt exist a decade ago. Interventional radiology transformed how we treat strokes and heart attacks. Personalized cancer therapy rewrote oncology protocols. Monoclonal antibodies moved from experimental to standard care. MI and stroke treatment protocols were fundamentally rewritten based on new evidence about timing and intervention.

And AI? Its already embedded in radiology and pathology departments, expanding continuously as the evidence accumulates.

The common thread isnt luck. Its evidence. Clear, rigorous, clinically validated evidence that shows measurable improvement in patient outcomes or workflow efficiency. When that standard is met, healthcare doesnt resist change. It accelerates it.

The question was never whether healthcare can change quickly. Its whether the change has earned the right to be adopted.

So Why Are AI Pilots Failing?

If healthcare can rewrite decades of dogma in weeks when the evidence is clear, if it can transform entire specialties in years when the value is undeniable, why do AI pilots struggle to get traction?

The answer isnt that healthcare resists change. Its that most AI pilots fail the same tests that these medical revolutions passed.

They dont make things measurably better. Bickell and Mattox showed survival improvement with rigorous data. What does your AI pilot show? Potential efficiency gains isnt the same language. Healthcare changes when harm is reduced or outcomes improve, not when dashboards promise optimization.

They add friction instead of removing it. Hands-Only CPR succeeded because it made lifesaving simpler. Most AI tools add steps, require new logins, create integration puzzles. They make the day harder, not easier. Healthcare abandoned rescue breaths to increase adoption; AI vendors add complexity and wonder why doctors wont engage.

They lack clinical validation. Healthcare changed HRT practice based on a rigorous randomized controlled trial with thousands of participants. AI vendors show up with 90-day pilots and customer testimonials. The evidentiary bar isnt arbitrary, it exists because good enough can kill people.

They dont come through trusted channels. When the NEJM publishes something, it flows through grand rounds, medical societies, clinical hierarchies. The information arrives with institutional credibility. AI vendors trying to sell around clinicians to administrators wonder why doctors wont use their tools. The channel matters as much as the message.

They solve the wrong problems. Laparoscopic surgery solved a patient problem: pain and recovery time. The value was immediate and undeniable. Many AI tools solve a billing problem, or an administrative problem, while making the clinicians workflow more complicated. When the person using the tool isnt the person benefiting from it, adoption fails.

The Real Story

Heres what I learned in that trauma room, and in the decades since:

Healthcare isnt slow to change. Its appropriately skeptical of changes that havent proven themselves in the way that matters.

Show me survival improvement. Show me workflow simplification with evidence. Show me better patient outcomes with rigorous data. Healthcare will move faster than any industry on earth because lives depend on getting it right.

Show me a dashboard and promise me AI-powered insights? Ask me to change how I practice based on a vendor pilot and projected ROI?

Thats not resistance to change. Thats pattern recognition.

Weve seen this before. The Swan-Ganz catheter was the gold standard in critical care for decades, until we learned that measuring everything didnt mean we could improve anything. Usage plummeted once the evidence was clear.

Weve watched health IT systems promise transformation and deliver workflow nightmares. Weve learned, sometimes painfully, that in medicine, good enough can kill people.

So yes, were skeptical. We should be. That skepticism has saved more lives than its cost.

The AI and HealthTech companies that understand this, that earn trust through evidence, that reduce friction instead of creating it, that solve clinical problems instead of administrative ones, will transform healthcare.

The ones that dont will keep writing think pieces about how healthcare is resistant to change while their pilots gather dust in IT closets.

The industry isnt resistant to change. Its resistant to mediocrity. Do the work to prove your value, and you wont have to ask for change, youll be the one struggling to keep up with it.

Dr. Yoram Friedman is a physician turned product manager with over 20 years of experience in enterprise software and digital transformation. He believes healthcare AI must rightfully earn its place through workflow, trust, and real-world outcomes.