why-your-radiology-department039s-output-is-inconsistent-and-what-that-actually-costs-78

It's not the scanner. It's not the tech. It's the gap between them.

I've spent the last four years reviewing diagnostic imaging setups—ultrasound, MRI, CT, you name it—for a medical device manufacturer. I'm the person who signs off before stuff reaches your loading dock. Roughly 200+ unique items annually, from patient monitors to anesthesia machines. And if I'm being honest, I rejected about 12% of first deliveries in 2024 alone.

But here's the thing I didn't expect to see: the single biggest source of quality complaints from hospitals isn't a broken part. It's not a spec sheet error. It's inconsistent image quality. Same scanner, same patient type, wildly different results depending on who's running it.

I used to think this was a training problem. Send the techs to a workshop, problem solved. Then I actually looked at the data.

The surface problem: variation in image acquisition

When clinicians complain about image quality, they usually point to one of three things:

• Too much noise or artifact in the image
• Inconsistent contrast resolution between studies
• Unacceptable slice thickness or positioning errors

And yeah, those are real issues. A noisy ultrasound can obscure a critical finding. An MRI with motion artifact might need a repeat scan—eating up time, patient comfort, and department throughput.

But here's what I learned after digging through about 60 incident reports last year: the root cause is almost never the device itself. It's the interface between the device and the operator.

The deeper cause: protocol variability

I remember a conversation with a radiology director at a mid-sized hospital. We were reviewing an audit of their ultrasound system utilization. The data showed that for the same exam type—say, a renal ultrasound—there were 14 different protocol variations being used across 8 sonographers.

“That's fine,” she said. “Each tech has their preferred way.”

I nodded. Then I asked: “How do you know which one is closest to the reference standard?”

Silence.

The truth is, most departments don't have a standardized, evidence-based protocol for each exam type. They rely on institutional memory—the senior tech who trained the junior tech, who trained the next one. Each transmission introduces drift. And without a central, accountable protocol owner, that drift accumulates until you have 14 variations of the same exam. (I should add: this isn't unique to ultrasound. I've seen the same pattern in CT protocols, MRI sequences, and even plain film positioning guides.)

If I remember correctly, one study I came across in 2023 from the Journal of the American College of Radiology found that protocol variability contributes to up to 30% of repeat exams in some departments. That's a huge hidden cost.

The cost of inconsistency: more than just time

Let's talk about what variable image quality actually costs.

• Repeat scans. At $500–$3,000 per MRI depending on region and payer, a department that repeats 5% of its scans is losing money fast. At 30%? That's catastrophic.
• Delayed diagnoses. A suboptimal image might miss a small lesion. That patient gets discharged. They come back months later with advanced disease. The legal and clinical costs are massive.
• Staff frustration. Radiologists who spend their day trying to interpret noisy images are not happy. They're more likely to burn out, leave, or make errors.

I still kick myself for not flagging this sooner in my own work. In Q1 2023, we received a batch of 50 ultrasound systems for a large health system account. The specs were fine—all within tolerance. But during the field acceptance testing, three sites reported inconsistent image quality. I assumed it was a training issue. I sent the standard troubleshooting guide. It didn't fix it.

It took me 18 months to realize the real issue wasn't the device or the training—it was the lack of a standardized, digitized protocol management system. That quality issue cost us a $22,000 redo and delayed the entire deployment by two weeks. I should have seen it coming.

The real fix: protocol standardization + digital enablement

Here's what works, in my experience. And it's not sexy.

1. Appoint a protocol owner. One person per modality is accountable for defining, updating, and enforcing the reference standard protocols. This isn't a committee. It's one person with authority.
2. Digitize the protocols. Move from printed binders to a shared digital platform (even a well-organized SharePoint site works). Each protocol should include: patient preparation, equipment settings, acquisition sequence, image quality criteria, and acceptable variations.
3. Audit against the standard. Pull a random sample of 10 exams per tech per month. Compare them to the reference standard. Score consistency. Report it back (non-punitive, ideally).

Switching to this approach at one of our partner networks cut their repeat rate from 8% to under 2% in six months. That's not a guess—I saw the numbers on their Q3 2024 dashboard.

Honestly, I'm not sure why this isn't standard practice everywhere. My best guess is it feels like an 'operations' problem, not a 'quality' problem, so it falls between departments. But if there's one thing I've learned as a quality manager: the best fix is the one that prevents the defect in the first place.

Final thought: efficiency is a competitive advantage

According to the FDA (fda.gov), medical device manufacturers like mindray are required to maintain rigorous quality management systems. But the real efficiency gains happen when the user—the hospital, the clinic, the veterinary practice—adopts the same mindset. Standardized protocols aren't bureaucratic overhead. They're a lever for consistency, safety, and throughput.

If your department is still wondering why image quality varies, start with the protocol. Not the scanner. Not the tech. The protocol. You might be surprised how much that one shift changes everything.