That Day in the ICU
I still remember the morning in early 2023 when I first walked into a busy ICU unit. I was there to observe how our Smiths Medical infusion pumps were being used in a real clinical setting. The nurses moved fast—IV lines everywhere, monitors beeping, alarms going off. It was controlled chaos.
But one thing caught my eye almost immediately: a nurse had taped a small handwritten note to the side of a Medfusion 3500 syringe pump. It read: 'Check flow rate every 2 hours—drifts.'
I was a quality manager. That note was a red flag.
Here’s the thing: I had approved that pump model. I had signed off on its specifications. And now a nurse was manually compensating for something we had missed. Not a catastrophic failure—just an inconsistency. A drift. A small one. But in ICU meds, small drifts matter.
The Assumption That Cost Us a Month
I assumed our lab tests covered every real-world condition. Turned out they didn’t.
We had tested the Medfusion pump in controlled environments—temperature, humidity, all stable. The pump performed beautifully: flow rate accuracy within ±2% across all standard tests. But in that ICU, the pump sat near an air vent. Cold air hit it intermittently. The internal fluid changed temperature slightly. The flow rate drifted. None of our protocols caught it.
I learned never to assume the proof represents the final product, no matter how many test cycles you run.
That’s when I started pushing for real-environment testing—not just bench tests. We spent four weeks redesigning our validation protocol. Some engineers thought it was overkill. “The pump’s specs are fine,” they said. “It’s a use error, not a design flaw.” Maybe. But to me, if the user has to compensate with a sticky note, that’s our problem too.
The Spirometer Incident: Another Assumption
Around the same time, I was reviewing specs for a new spirometer model—the one our catalog listed as compatible with Smiths Medical airway management kits. The marketing team loved it. The sales team had already promised delivery.
But when I looked at the connector interface, something felt off. I asked the sourcing team: “Did we verify that this spirometer mates with our Portex tracheostomy tube adapters?”
Silence.
Turned out—no. The assumption was that “industry standard” connectors meant automatic compatibility. I reject that phrase now. Industry standard is mostly wishful thinking when you’re dealing with devices from different production batches, different QC tolerances, and different engineers.
We tested it. The connection was loose under torque. Not unusable—but loose. In a code blue scenario, that loose connection could fail. We flagged it. The launch was delayed two weeks. Yes, there was internal pushback. But I’d rather delay a launch than receive a complaint from a surgical team in the middle of a procedure.
“5 minutes of verification beats 5 days of correction.” — That is not just a slogan. In Q2 2023, that mindset saved us an estimated $18,000 in potential rework and replacement related to connector mismatches across two product lines.
The Numbers Said Go. My Gut Said Wait.
I’ll be honest: there’s a tension in quality work between speed and thoroughness. Every day, you’re asked to balance launch deadlines against verification checklists. And sometimes the data says one thing, but your experience says another.
Take the Acapella Smiths Medical device (the positive expiratory pressure therapy device). We were considering a packaging change: new outer carton, same inner tray. The cost analysis said go—save $0.14 per unit. On a 50,000-unit annual order, that’s $7,000 in savings. The numbers were clear.
But my gut said: “Check the sterilization compatibility.” The new carton material was slightly different. Would it hold up under ethylene oxide sterilization? The data sheet said yes. The supplier said yes. But I had been burned before by “data sheets.” So I asked for a small trial batch: 100 units. It cost $300.
Turns out the new material had a slightly lower gas permeability. Sterilization took 12 minutes longer per cycle. On a large run, that adds up. More importantly, it changed the scheduling of our sterilization partner. Had we just flipped the switch, we would have missed delivery dates by three weeks—and paid rush fees.
Calculated the worst case: a $4,500 cost for re-sterilization and delays. Best case: we save $7,000. The expected value said go for the new carton. But the downside—a delayed delivery to a hospital—felt catastrophic. I stuck with my gut. That $300 trial batch was the cheapest insurance we ever bought.
Building a Checklist (Finally)
After these incidents, I did what any frustrated quality manager would do: I built a checklist. The third time we stumbled over an assumption, I finally created a formal verification protocol. Should have done it after the first one.
Here’s what it covers for every new or modified Smiths Medical product before it goes to production:
- Environmental range testing (temp, humidity, air pressure—simulating ICU, OR, ambulance transport)
- Connector compatibility with at least 3 different adapters from different production lots
- Sterilization cycle impact on packaging (beyond just material data sheets)
- User workflow test: 5 real-world clinicians using it without instructions
- Backup vendor spec: if supplier changes material, what triggers a re-review?
Is it perfect? No. But it has caught three issues in the past 12 months that would have reached customers. That’s three incidents of preventable rework. Three letters I didn’t have to write.
The Real Lesson
I’ve been a quality manager for over four years now. I’ve reviewed more than 200 unique deliverables in that time—specs, packaging, sterilization protocols, labeling, instructional inserts. I’ve rejected maybe 15% of first submissions in 2024 due to incomplete or unverified claims. And every time I push back, I hear the same thing: “But it’s within industry standard.”
Industry standard is the lowest bar. If your bar is industry standard, your product will be average. And in medical devices—especially for devices used in ICUs, ORs, and emergency departments—average is not good enough.
Prevention isn’t glamorous. It’s not a breakthrough discovery. It’s saying “let me check that one more time” when everyone else is ready to ship. It’s running a $300 trial to avoid a $4,500 problem. It’s listening to that nurse’s sticky note.
The pump we fixed? We updated the firmware to compensate for temperature drift and added a note in the instruction manual about vent placement. The spirometer connector? We added a small foam gasket to stabilize the interface. Simple fixes. But they came from the hard part: checking before we believed.
That’s prevention. And it’s always cheaper than the alternative.
