If you second-guess a diagnosis more often than you like, this is for you

This is a story about the biggest trap in radiology – confidence.

Too much or too little confidence makes radiologists burn out, earn less, annoy clinicians, or let down patients. Yet, for some reason, it’s not taught. Radiologists are just expected to figure it out on the job.

That seems like a bad idea to me! So, I spent the last two years writing a book on confidence.
I read widely and spoke with radiologists around the world, who told me things like:

"I get stuck on findings that are probably not relevant, but I’m not sure. I never feel comfortable."
"I just don't have the confidence I'd hoped I have at this stage."
"I’m reading 40 MRI studies a day then going home not confident at what I did. I worry. It’s impacting my time with the kids."

Many of these radiologists know their stuff—but low confidence is ruining their quality of life.

In fact:

Less than 50% of radiologists feel “very confident” in their daily reporting.
In MSK and neuroradiology, that number drops to 15% and 16%.
83% experience imposter syndrome at some point in their careers.

It’s not surprising. As a radiologist, if you make a wrong call, it’s not just the patient who suffers. You may be sued! Increasingly, this leads to reports that are more “defensive”: longer, vaguer, and full of unnecessary differentials.

And while this can be painful, that’s not the biggest problem…

We all know that in life, the most confident people are not always the most competent.

It’s the same in radiology. Some write over-hedging, vague reports. Others overcall everything in absolute terms, whether they’re right or not.

Both fall into the same cognitive trap: the Dunning-Kruger effect.

​Doctors with low competence feel highly confident because they don’t know what they don’t know. Meanwhile, highly competent doctors doubt themselves because they’re painfully aware of everything they still don’t know.

But overconfidence is actually MORE common than underconfidence. Why?

There are two reasons.

1. The Better-Than-Average Effect.

Ask yourself, are you a better driver than most people?
If you said “yes”, it’s likely an example of the Better-Than-Average Effect.

In the 1970s, psychologists found that most people believe they are safer and more skilled drivers than the average person, even though this is statistically impossible. This bias appears in many fields– including medicine. It leads doctors to overestimate their accuracy and fail to double-check findings.

​

2. The Beginner’s Bubble.
Beginners often experience a confidence spike after their first phase of learning. They get familiar with patterns, gain some speed, and feel like they’ve mastered it.

This is the ‘Beginners Bubble’.

Then reality hits. They make an error. Confidence comes crashing down. I’m sure you remember. It’s all part of the process. It’s a predictable cycle that can be overcome with resilience, guidance, and time, but it sucks!

Both reasons have identical root causes:

We’re selective about what we pay attention to, forgetting our errors.

We have a self-serving bias when it comes to choosing the criteria we judge ourselves on.

Most importantly, we lack honest, consistent feedback from others (more on that shortly).


​For radiologists, the real question isn’t just about confidence—it’s about how clinicians respond to it.

I once heard an orthopedic surgeon in London say:

"When I get a radiology report, the first thing I check is the name at the bottom. If I don’t know or trust them, I ignore the report completely."

That stuck with me.

​Because here’s the thing: clinicians won’t tell you they don’t trust you. They won’t argue with your report. They won’t send you an email explaining. They just ignore you. The key to earning their trust isn’t more confidence or less confidence—it’s better calibration. And improving your calibration starts with understanding how you think.

Imagine this.

A first-year resident is reporting a CT abdomen. He carefully checks each structure, double-checking anatomy, unsure if something is normal. He looks up the differentials to be sure. 45 minutes later, he finishes the report.

A senior resident looks at the same case. He scrolls, quickly recognizing a normal upper abdomen. He focuses on the lower abdomen, spots inflammation around the colon, and calls diverticulitis in just a few minutes.

What’s the difference?

The first-year resident is using a slow, deliberate way of thinking—analysing every detail, questioning findings, and taking his time. The senior resident is using a fast, intuitive process to quickly rule things out.

We slip between these two ways of thinking all the time. You may know them from the famous book ‘Thinking, Fast and Slow’, by Daniel Kahneman.

System 1 is fast, intuitive, and driven by pattern recognition.

System 2 is slow, analytical, and deliberate. It requires effort, logic, and careful checking.

Experienced radiologists rely on System 1—and that’s a good thing. It’s what allows us to report faster and more efficiently.

But here’s the problem: System 1 is prone to bias. If you expect to see something (e.g., a meniscal tear), you’ll probably see it—even if it’s not there (confirmation bias). If the referral says "suspicious for sacroiliitis," you’re more likely to call it—even if the findings don’t fully match (anchoring bias).

Without System 2 Thinking, you’ll never catch those errors. If we over-rely on one system, we run into problems. And this is the key to real confidence:

If you work the case properly, your reports improve.

If your reports improve, you get better feedback.

If you get better feedback, your confidence increases.

It’s a cycle.

Overconfidence happens when you trust System 1 too much—you don’t slow down, don’t check, and miss things. Underconfidence happens when you stay in System 2 too much—you overthink, hesitate, and second-guess.

The goal is not to be more confident or less confident. The goal is to calibrate confidence correctly.


​And that comes from knowing when to trust System 1… and when to slow down and use System 2.

It’s not about being more confident or less confident. It’s about calibrating confidence—so you trust yourself when you should and slow down when you need to.

The best radiologists aren’t the ones who feel the most sure of themselves.

They’re the ones who think about how they think—who recognise their blind spots, seek feedback, and refine their approach over time.

A good radiologist doesn’t just sound confident. They make clinicians confident in them. That starts with knowing when to trust your instincts—and when to stop and think.

​Because confidence doesn’t earn trust, accuracy does.