Challenging what we think we know in radiology.
A few years ago, I became aware of routine, year-round, cheap availability of mangos in my local grocery stores. Living far from climes where such produce will grow, I recognized this as quite a luxury. Certainly, it hadn’t been the case for the majority of my prior few dozen years on the planet.
One reason I got into the habit of picking up a couple each week was that they seemed absurdly cheap for something so divinely tasty (and procured from afar): The store near me priced them at $1 each. When I first encountered this, I thought there must be some kind of oversupply, and assumed it was a fluke. Time went on, though, and they were still a buck a piece, more often than not…and not just at my usual store.
Many in our field (and other scientific ones, I would guess) have heard, if not said, that “the plural of anecdote is not data.” Yet it’s fascinating to see how many of us routinely take our handful of personal experiences and overly eagerly extrapolate them to expectations of how the world works.
Case in point: After just a few years of buying mangos under these circumstances, in my mind the proper price for one has been established as $1. If I see them being sold for less (even 99¢), my brain proclaims it a bargain. If they’re $1.10, depending on my mood, I might consider it unreasonable, even forego my fruit in protest.
But really, what factual basis does my mindset have? For all I know, the past few years have seen a worldwide mango glut-or shortage. Maybe “normal” should actually be $1.50, or 80¢.
Our brains like it when the world conforms to our expectations, and resist or resent it when it seems that we might have been wrong about something. Especially those of us who have come to believe that we’re good at puzzling things out, making theories, and drawing conclusions (like radiological diagnoses).
I cannot, for instance, tell you which of my residency attendings first taught me that the short-axis diameter diagnostic of aneurysm in the aortic arch was 5 cm in the ascending aspect, 4 in the transverse, and 3 in the descending. I certainly can’t tell you what body of evidence upon which that attending’s was based; for all I know, he heard it from an attending in his own residency, and never questioned the matter.
Neither did I…at least, not until a few years had gone by and the Internet became a little more robust than it was at the time. I started noticing that other docs didn’t always seem to be using the same reference values as me. At first, my anchoring bias (that’s what it’s called, by the way; I didn’t invent the concept with my mangos) was strong enough for me to summarily decide that these other docs were wrong. Enough instances, though, and I was motivated to do some research to see if my original information was, indeed, correct.
I’ve had multiple instances of this-that is, questioning what I (and others) have taken to be Established Fact. It’s most common with numerical stuff-measurements of size, attenuation, signal intensity, etc. It’s reassuring when I can confirm what I thought I already knew, since the alternative means that I (and others) might have been reading cases wrong for years.
It’s not just a matter of good information or bad. Sometimes, the actual facts change, or further research clarifies them. Suppose, for instance, that as our population gets progressively less-than-perfectly-fit over the decades, there are more and more people walking around with more ectatic aortas. If we continue to define “normal” as being within a certain number of standard deviations of the mean, the diagnostic cutoff for an aneurysm might drift upwards.
There’s also the matter of anchoring bias in other aspects of our profession. Compensation, for instance: Someone who started working as an attending radiologist in, say, 1990 is going to have a very different idea of what is a reasonable salary or daily RVU tally than someone who emerged from residency a couple of years ago.
Not a bad habit, then, to occasionally recheck what you think you know…it might just turn out that you really don’t.
The Reading Room: Artificial Intelligence: What RSNA 2020 Offered, and What 2021 Could Bring
December 5th 2020Nina Kottler, M.D., chief medical officer of AI at Radiology Partners, discusses, during RSNA 2020, what new developments the annual meeting provided about these technologies, sessions to access, and what to expect in the coming year.