Please. Just Tell Me Something

I remember reading (and also being told about this repeatedly in a million lectures on compliance and other unstated but critical physician tasks) that you need a reason for what you do. In imaging, that is often the clinical information provided in the request. Yes, a request. Someone is requesting that you perform the imaging study on their patient. Most of the time, it is a no-brainer: a simple match of the clinical question and of the study.

Stroke? Head MR. Check. Match. Pneumonia? Chest x-ray. Check. Match.

However, sometimes (as we all know), you may not have the best fit for the request.

Radiology is again under the microscope, and seems to be attracting a lot more interest than perhaps is warranted or deserved. AI is out there, lurking. It seems it is either going to replace us or generate the need for many more of us, depending on who you read.

There are several items that make me happy I am at the portion of the career arc I am now inhabiting—this is a big one.

I have thoughts, but I am not looking at another 30 years of doing this. Regardless of what side you may come down on in this debate, certain facts cannot be ignored. And, one of them, to me, is the clinical data generating the request.

Do you think AI machines will argue with one another about whether the study is indicated? This AI looks through the data and says, yes; this AI looks through the data and says no.

Will they duke it out in the hallway? Try to pull each other’s plugs? Figure out how to reboot each other? Resort to name-calling?

What made me think of this was a particular run of weird-odd-impossible-crazy-nonsensical items of clinical data to accompany exams today. A personal favorite was the provided clinical data for an IAC brain MR: “illness, unspecified.” Another best-of-the-day close contender was this clinical data for spine MR: “long history of.”

Yes.

I went to the EMR. You know how it is when not everyone the patient is seeing is on your own EMR? That big black hole of information that pertains to the current study? Yep. That happens much too often to be just chance.

I went to a patient data intake questionnaire for one of the studies. The patient circled the question when asked what they were being scanned for. Clearly, they were in on the “let’s withhold data from the radiologist” game. Were the studies indicated? Well, as best as I can say, yes.

Just trying to keep moving forward.

Keep doing that good work. Mahalo.