When the Clinical History is the Equivalent of a Low Signal-to-Noise Ratio

If you have read this blog for the long haul, you have seen me sound off about worthless histories of “R/O pathology” and the like. I have gotten studies in which the “reason for exam” field has been completely blank, something that most software can prevent. Some referrers retort by making their RfE a single “.” so the software won’t object. I think they invest more effort into not providing a history than they would if they just said why they are ordering the exam.

They still provide the majority of misbehavior, so I wouldn’t ask most of them to “say less.” Even if I did, those saying nothing (or “.”) leave themselves no room to comply.

“Say less” is for a more recent wave of health-care personnel with deficits in their learning, training, and/or caring. The histories they give are wordier to be sure, but the words don’t add up to much. To put this in rad speak, their signal-to-noise ratio is low.

One rad on social media recently posted an example: “Abdominal pain. Cholelithiasis vs. cholecystitis vs. choledocholithiasis vs. pancreatitis vs. appendicitis vs. SBO vs. gastritis vs. GERD vs. PUD.”

Compared to what I often see, that is somewhat tidy. The kitchen sinks I receive commonly have one or two extra lines on top of that.

When these things started crossing my path, I had momentary joy that referrers finally got the message: Try to act like doctors or at least worthwhile physician extenders. Don’t just fling your patients at radiologists or other consultants and expect them to do all your work for you. At least show that you gathered some relevant information and knocked a couple of your brain cells together, thinking about what might be causing the patient’s problem.

Then I saw that many of the items in their diagnostic laundry lists didn’t really belong together. The pathologies they referenced were all over the place and unlikely to present in such a way that they would be confused with one another.

Indeed, their verbiage didn’t give any indication that they had actually seen the patients. A third-year medical student in an ER could pick up a triage sheet that says “abdominal pain” and immediately tell you a dozen different potential sources for pain.

Better than that and presumably within the capabilities of folks referring patients for imaging would be to find out via H&P where the pain is, how long it has been there, if it moves around, etc. Is there a fever, relevant background of surgery, cancer, or other known diagnoses?

If, on the other hand, your “reason for imaging” is just the first dozen pathological entities you can pull off the top of your head (and not why any of them strikes you as probable), you are not contributing to the situation. All you are doing is wasting your time by spewing bits of medical vocabulary, and the radiologist’s time by making him or her read them. Saying “cholecystitis” and “testicular torsion” in a word salad isn’t going to make the rad look for those things any more diligently than he or she would have if you had just said “Pain.”

Maybe, on occasion, you can’t do more. You could explain why: “Patient is in pain but is demented and won’t indicate where or allow physical examination.” At least that implies you would have given more details if you could out of some consideration for the rad (or, go figure, the patient).

If, on the other hand, every single history you provide has the same deficits, we can see that you are the common denominator and are not fooling anyone.

The other “say less” histories I see are massive blocks of text that someone copy/pastes from the patient’s intake or admission files. Relevant information might be buried in there somewhere but so is a novella of irrelevant babble. Devoid of editing or even formatting, it might make the reader dream of having a pickax to go mining for useful nuggets in the boulder you have left at his or her door.

More likely, the busy individual, having been through this exercise before with little to show for it, just glances at your wall of text and ignores it. Alternately, the rad skims and fails to identify the couple of words that might have helped him or her (and the referrer).

Since many folks have limited ability (or interest) in understanding such things from another person’s point of view, allow me to showcase how it would be for our referring clinicians if we responded in kind.

Suppose that every time you gave a laundry list “history,” our reports proceeded to give you a laundry list of everything we “cannot rule out,” instead of just telling you whether we think any particular abnormality is actually present. How much more time do you suppose you would spend reading our stuff and trying to figure out if we are telling you anything useful? Might you feel compelled to request consults or additional imaging for each item we specified could not be ruled out?

Alternately, suppose you are one of the wall-of-text referrers. What if we abandoned all efforts to keep our reports formatted and tidy? Maybe we stop using structured templates and never separate our findings into paragraphs. Now it is an impenetrable rectangle of characters that takes up entire pages or screens at a time. Would that be helpful? Would it convey any sense at all that we were trying to be of service, or might it seem uncaring or even more of a passive-aggressive response?

Yes, I know some rads actually do these things. Just like a conscientious referring clinician might say regarding their bad history providing colleagues, it is an unfortunate minority that is not representing the rest of us. We do what we can to encourage better behavior from them. I like to imagine clinicians do the same.

For the rest, if it is a given that you are not going to provide better histories any time soon, please at least keep your “garbage in” brief. Useless words don’t become useful just because you put more of them together.