I Review Medical Devices for a Living. The Suppliers Who Say "No" Are the Ones I Trust

I'm the person who signs off on medical devices before they reach hospital floors. Specifically, I'm a quality/compliance manager at icu-medical — I review every infusion pump, patient monitor, and ICU bed that goes out our door. Roughly 200 unique products a year cross my desk. I've rejected about 7% of first deliveries in 2024 alone.

After four years of this, I've reached an unpopular conclusion: the medical device supplier who tells you "this isn't our specialty" is demonstrating quality control, not admitting weakness.

It took me a while to get here. I used to assume the ideal supplier was the one who could provide everything — a full catalog, every product category, one purchase order covering it all. I thought that was efficiency. (I also thought that was what "one-stop-shop" was supposed to mean.)

Turns out, after inspecting thousands of units, the pattern isn't about manufacturing skill. It's about focus.

The one-stop-shop assumption that didn't survive contact with reality

Early in my career, we evaluated a vendor who claimed to make ICU equipment and dental chairs. Not a problem, right? Both are "medical."

Wrong. The quality requirements for a dental chair — patient positioning, saliva drainage, handpiece connectivity — are completely different from what an ICU bed needs. Fluid resistance is different. Weight-bearing is different. Cleaning protocols are different. And the regulatory paperwork doesn't overlap nearly as much as you'd think.

What I learned the hard way: a company that says "we do it all" often has its core competency stretched thin across categories. The dental chair turned out fine. The ICU bed — the product we actually needed — had spec deviations on the siderail tolerance. Not catastrophic. Just enough that I wouldn't sign off on it.

We rejected the batch, and the vendor redid it at their cost. (Now every contract includes siderail spec requirements.) That was my initial misjudgment: I believed capability was additive. It isn't. Expertise is concentrated.

How does a pulse oximeter work? And why that question tests expertise boundaries

The pulse oximeter is the perfect example of why boundaries matter. It looks like a simple clip. Two LEDs, a photodetector, a screen. What most people don't realize is how much engineering goes into making that clip accurate.

Here's how it works: red light around 660 nm and infrared light around 940 nm shine through the fingertip. Oxygenated hemoglobin absorbs infrared light differently than deoxygenated hemoglobin. A photodetector on the other side measures the difference. The device isolates the pulsatile arterial signal — the "pulse" in pulse oximeter — from the background tissue signal, then estimates SpO2 from that ratio.

Sounds straightforward. In practice? Motion artifacts, low perfusion, skin thickness, ambient light, even nail polish can skew the reading. The algorithms that separate real signal from noise are where the specialization actually happens.

A generic electronics manufacturer could build a pulse-oximeter-shaped object. But one that stays accurate during patient movement in an ICU? That takes a team who has spent years working on photoplethysmography.

Now extrapolate that to a surgical robot. Haptic feedback, latency in milliseconds, sterility requirements that make a pulse oximeter look simple by comparison. The engineering depth is on another planet. So when a supplier claims equal expertise in ICU infusion pumps, pulse oximeters, and surgical robotics, I ask for their validation data. Not to be difficult — but because I've seen what happens when companies stretch beyond their validation boundaries. (It's never a small fix.)

The data on stretching beyond your lane

Everything I'd read about procurement says to reduce vendor count. Consolidate. Streamline. And that's true to a degree. But here's the experience override: in our Q1 2024 quality audit, we measured deviation rates across supplier types. Suppliers with a clearly defined product specialty had a first-pass yield of 98.7%. Suppliers who took on products outside their stated core competency? 89.4% on those non-core items.

Nine percentage points might not sound dramatic. In practical terms, it's a specialist being asked to perform outside their field, and failing nine times out of every hundred tries. (I can only speak to our audit data and our supplier pool. Different industries may have different numbers.)

The conventional wisdom is that more capability is always better. My inspection records say the opposite: know-how is tied to a specific context. And context is expensive to scale.

Sure, but what about one purchase order?

The obvious pushback: doesn't a broader portfolio help procurement teams simplify? Fair question.

Some breadth is genuinely useful when it's organized around a core competency. An ICU ecosystem — infusion pumps, patient monitors, IV solutions, hospital beds — represents related products serving a single care environment. Expertise transfers naturally across these categories because the clinical context is stable: the ICU workflow, the ICU staff, the ICU environment. That's not a stretch. That's depth.

The line we've drawn at icu-medical is that we serve the ICU ecosystem and its immediate surroundings. We know the Minnesota healthcare market well, and we understand how regional hospitals standardize their ICU equipment. That's our context. We don't claim to be the definitive authority on dental chairs, and we're not going to pretend we compete with surgical robotics specialists on their home turf. Different products. Different failure modes. (And different regulatory burdens, which deserves its own article.)

Do we lose some sales by being honest about this? Probably. But every order we take is one we can stand behind in a quality review. I'd rather ship ninety products I can defend than two hundred I'm unsure about. (Not that the sales team loves hearing it. Quality doesn't negotiate.)

Boundaries are a feature, not a problem

So here's my advice if you're evaluating a medical device supplier: listen not just to what they say they can do — but to what they tell you they can't.

The "we don't specialize in that" conversation is a signal. It means the company has a realistic view of its own strengths, and it's willing to risk losing a sale to protect its quality record. That's exactly the mindset you want in a partner supplying ICU equipment.

Because in medical devices, there are no low-stakes categories. A pulse oximeter that reads 98% when the real SpO2 is 93% isn't a product defect. It's a patient-safety event. Specialization is a quality control measure.

And that vendor who redid the ICU bed batch? After that, they stopped claiming dental-chair-level expertise in ICU products. Their first-pass yield went from 89% to 97% within the year. Turns out, telling customers what you're not good at is good business too.