- AI
- Molecular Imaging
- CT
- X-Ray
- Ultrasound
- MRI
- Facility Management
- Mammography
Reassessing Recommendations in Radiology Reports
Is there a happy medium with “recommendations” for referring clinicians?
Last week, I wrote about the bloating effect of excess verbiage crammed into our reports by third parties. However, the world isn’t quite so neatly divided into guilty and innocent. Radiologists have inflicted some of this on themselves.
Looking 20-plus years back on my residency training, I recall a typical report in our rad department having three sections: a patient history or “Reason for Exam,” the report’s body (“Comment,” or “Findings”), and a summation (“Impression,” “Conclusion,” etc.).
Your experience may differ, but in the ensuing couple of decades, I saw other sections gradually being added. I couldn’t tell you in what order they showed up, who required them, what reasons were given, etc. Now my typical report has at least six sections (double the original): exam type, history, reason for study, comparisons, technique, findings, and impression.
Along the way, it all started looking clunky and excessive to me. No matter how short you make them, six separate paragraphs fill up a lot more space than three, and I have a stubborn instinct that my report is less than tidy if it doesn’t all fit on one page. Furthermore, they aren’t all of great importance to most readers. Does anybody other than a regulator really care, for instance, what MRI sequences were utilized, or how much contrast got injected? I don’t think many docs would mourn the “technique” section if it quietly vanished.
Once one incorporates things into his or her routine, they tend to stay put. I can’t recall any time in my career that I was able to cut back on the number of sections in my report despite yearning for instance, to fold “comparisons” into the “technique” paragraph. With this in mind, I have been increasingly wary of adding any more sections.
At some point, I saw folks using a seventh section for “recommendations.” Placed after the “impression” section, they would put any suggestions regarding further workup, management, or future surveillance studies within “recommendations.” I admit I was briefly tempted to adopt this, but my aversion to adding more sections was strong enough to hold the line. Besides, I keep my impressions short enough so they can contain my recommendations without anything getting buried.
Additionally, I have found that this section is often empty (“recommendations: none”). To my eye, that fell flat. A practicing physician churns out a page or more of diagnostic information and has nothing to recommend? It might give a sense that he or she hasn’t contributed anything to the situation. I also had a foreboding sense that it could look bad in malpractice cases, even in frivolous cases. (“Isn’t it true, Doctor, that if you had made some recommendation, patient X might be alive/well today?”)
I could just imagine some rads overcompensating for that, and filling a recommendations section with meaningless fluff: “Clinical correlation suggested,” “Further action might be warranted by findings of history, physical exam, or other diagnostic investigations,” etc. Not only does that fluff add worthless length to the report, it understandably annoys some referrers. What, you think I didn’t already “clinically correlate” my own patient? Are you implying I’m some sort of charlatan?
Some clinicians get annoyed even when rads offer more substantial recommendations, like appropriate modalities and intervals for follow-up imaging. They don’t like having their hands tied by our reports, or they just resent what they perceive as our intrusion on their decision-making turf.
Meanwhile, some referrers sit at the other end of the spectrum, wanting us to tell them what to do on a routine basis. A couple of jobs ago, I encountered a doc who requested such addenda on virtually every one of his patients. More than a few times, I stared at my reports (which were often completely normal), trying to figure what he could possibly be asking about. At some point, I asked one of the ancillary staff if she had any idea. Hearing the doc’s name, she gave a little laugh and said he was notorious for this.
I hadn’t yet gotten salty enough to simply say “I have no recommendations. There is no need for an addendum” as some of my colleagues did. Instead, I wound up crafting a dictation macro that went something like “No recommendation for further action based purely on images of this exam and available clinical history.”
A thread about this recently turned up on one of the radiology social media I follow. The poster lamented how some referrers complain about us recommending too much while others complain that we don’t do it enough. I shared my experience. While I don’t hear from the vast majority of referrers either way, I feel like the small fraction which asks me for more has been roughly equal to the proportion that demands less. I have no reason to conclude that my “silent majority” subdivides any differently.
Thus, until and unless I find myself reading for a small enough referring audience that lets me keep track of who wants what, it feels reasonable to consider what I have been doing a “happy medium” that requires no adjustment.
New Study Examines Short-Term Consistency of Large Language Models in Radiology
November 22nd 2024While GPT-4 demonstrated higher overall accuracy than other large language models in answering ACR Diagnostic in Training Exam multiple-choice questions, researchers noted an eight percent decrease in GPT-4’s accuracy rate from the first month to the third month of the study.
The Reading Room Podcast: Emerging Trends in the Radiology Workforce
February 11th 2022Richard Duszak, MD, and Mina Makary, MD, discuss a number of issues, ranging from demographic trends and NPRPs to physician burnout and medical student recruitment, that figure to impact the radiology workforce now and in the near future.
Can Innovations with AI Help Address the Impact of Staffing Shortages on Radiology Workflow?
October 7th 2024While staffing shortages in radiology continue to persist after the COVID-19 pandemic, current and emerging innovations powered by artificial intelligence (AI) may help facilities navigate these challenges and mitigate rising costs of health care.
