Feelings Aren’t Data. That’s the Problem.

Radiologists spend their days in the dark.

Not just physically, in dim reading rooms and behind glowing screens, but in a kind of professional silence. It’s a space defined by focus, pressure and constant decision-making. It’s also where something builders that we rarely acknowledge: how they feel in that moment.

Not because it isn’t real. Because the measurement  doesn't show up in a report

Over the past year, I’ve noticed a pattern in conversations with radiologists. It doesn’t come up in formal presentations or structured case studies. It shows up in passing comments, usually unprompted when asking about Rad AI.

They’ll say things like:

“I feel more confident.”
“I feel less buried.”
“I feel like I can think again.”
“I’m faster, but I’m also better.”
“I feel human again.”

And almost immediately, they’ll walk it back. They’ll add a qualifier or shift the conversation, as if what they just said doesn’t quite belong.

There’s an ingrained belief in medicine that unless something is formally measured, it doesn’t count. Feelings, by that definition, are easy to dismiss.

But if you actually listen, these statements are anything but vague. They’re precise descriptions of what’s changing in the way radiologists work.

Confidence means less second-guessing and clearer reads.
Feeling less buried reflects regained control over workload and pace.
Being able to think again signals a reduction in cognitive load.
Being faster and better points to systems that improve efficiency without sacrificing quality.
And feeling human again speaks to whether this work is sustainable over time.

These aren’t abstract sentiments. They are indicators of performance.

What’s striking is that in almost every other high-stakes field, this is already understood.

In aviation, pilot fatigue and cognitive load are actively monitored because they correlate directly with error risk. In the military, operator state, stress, focus, fatigue, is treated as a critical variable in mission performance. In elite sports, mental state is measured alongside physical metrics because it determines consistency and decision-making under pressure.

These industries don’t treat human experience as separate from performance. They treat it as part of the system.

Healthcare is the exception.

That distinction matters more than ever because as we all know radiology is under real strain.

Imaging demand has surged over the past decade, with CT volumes more than doubling and MRI continuing to climb. Nearly 70% of hospital patients rely on imaging as part of their care. At the same time, the workforce has not kept pace. Burnout rates are approaching 50%, and turnover has increased significantly.

This is no longer about optimizing workflows at the margins. Radiology is facing a fundamental capacity challenge with implications for patient care, financial stability, and system-wide risk, yet performance is still measured the same way. We track throughput, RVUs, turnaround times, and error rates with precision, but rarely ask a more basic question: what condition is the clinician in when those outcomes are produced? Other industries recognize that human state — cognitive load, fatigue, perceived stress — directly impacts performance. Healthcare has been slower to catch up.

But the connection is becoming harder to ignore.

When the experience of doing the work improves, the outputs improve as well.

We see this clearly in the impression workflow, one of the most cognitively repetitive and consequential parts of a radiologist’s day. Translating findings into a clear, structured impression requires constant mental effort, and that effort compounds across dozens of cases.

When that burden is reduced, the impact is immediate.

With tools like Rad AI Impressions, radiologists are saving more than an hour per shift, while reducing dictated words by up to 80% per impression. Time spent generating impressions drops significantly, but what stands out is not just the efficiency gain. It’s the shift in how radiologists describe their experience.

The shift shows up in ways that are hard to capture in metrics alone. As one radiologist shared, their spouse noticed they were no longer going back to the basement at night to finish reports — they had enough time to get through their work during the day. That’s not just efficiency. It’s what it looks like when cognitive load lifts, when the work fits back into the day, and when someone can leave it behind and be present at home.

That change in state is what drives the outcomes that follow.

Clinical error rates in impression sections decrease substantially. Broader analyses show meaningful reductions in report errors, including missed details and inconsistencies. At the same time, productivity increases and systems see improvements in throughput.

These results are not independent of how radiologists feel. They are a direct consequence of it.

For a long time, healthcare has treated clinician experience as something separate from performance, as if it sits outside the core equation. But that separation doesn’t hold.

Efficiency, quality and well-being are not competing priorities. They are interconnected outputs of the same underlying condition: the cognitive and emotional state of the clinician.

If that state is compromised, performance follows.
If that state improves, performance follows.

Radiologists don’t just see in the dark. They carry the weight of what they see, along with the volume, the expectations and the responsibility that comes with being central to nearly every care decision.

Ignoring that reality doesn’t make it less important. It simply leaves one of the most influential variables in healthcare unexamined.

So when a radiologist says, “I feel better using this,” it’s worth paying attention.

Not because it’s anecdotal, but because it points to something measurable even if we haven’t fully defined how to measure it yet.

And ultimately, it’s a signal about the quality of care.