June 5, 2025
ARSF Burnout Blog 2: Radiologist Well Being
Workloads and Staff Shortages: A Conversation with Dr. Elizabeth Rula
This is the second of an 8-part series on radiologist well-being. The author is Emory University Center for Ethics professor John Banja, and the blog is supported by an unrestricted grant from the Advanced Radiology Services Foundation.
Comments and responses are more than welcome.
With apologies, I’ll begin this blog by stating the obvious: Radiologist well-being—indeed, the well-being of any health professional—is significantly affected by the nature and volume of their workload. But as we all know, the radiologic workload has become excessively high while staff shortages have become acutely worrisome—so worrisome that staffing shortages in radiology were the biggest concern reported at the 2024 ACR meeting (Fornell 2024).
The radiologic workload has dramatically increased from 2011 to 2021 with CT and MRI usage increasing 3 to 5 percent annually (Fornell 2024). During that period, however, CMS’s budget reconciliation act resulted in a reimbursement decline of about 5.5 percent a year for the 50 most performed imaging studies among Medicare beneficiaries (Schartz 2022). The size of the radiologic workforce, on the other hand, has not kept up with the increasing work volume, but it has created a tremendous demand for radiologists. An anonymous commentator alleged that compensation for radiologists increased 16 percent from 2022 to 2023 (Anonymous 2024), with three job openings currently available for every radiologist graduating from a residency program (Fornell Jan. 31, 2024). In 2024, radiology technologists had a vacancy rate of 18.1 percent, while the number of medical residents in radiology fell to 1,006 in 2023 compared to 1,084 in 2010 (Stempniak February 9, 2025). A recent study by Eric Christensen and his colleagues predicted that without dramatic increases in residency training slots or other supportive measures that would reduce the workload, the shortage of radiologists will not change over the next three decades (Stempniak February 12, 2025).
There are numerous factors to explain the increase in demand for radiologic services, and it’s tempting to speculate on the most likely cause. For instance, some commentators might point to the aging population in the US, and the way that population’s increasing longevity and health care needs drive an increase in imaging referrals. Furthermore, an increasing number of persons in this population will have good health insurance, somewhat enabled by the Affordable Care Act, so reimbursement won’t be an obstacle (Fornell August 22, 2024).
I tend to be more cynical, though, and look to other, less savory explanations. First of all, what about declining reimbursement? I was recently asked by a non-radiology, clinical leadership group to talk about the ethics of unnecessary testing. The group leader was frank in telling me, “We’ve watched our reimbursement decline every year for at least the last decade. So what do we do? We make it up in volume, either for ourselves or for the hospitals where we work.” And there doesn’t seem to be much the radiologist can do to diminish the number of imaging referrals, very possibly coming from physicians who are pressured to meet their clinic’s or hospital’s productivity targets. Yes, there certainly are appropriate use criteria for radiologists, but as I’ll write in my next blog, referring physicians don’t pay much attention to them.
Maybe most of all is the possibility that physicians are afraid of being sued if they don’t order that imaging study. This is one of the most interesting categories of radiologic overutilization leading to excessive workloads. As I read the related literature, two things stand out: The first is that the “standard of care” seems very unpersuasive in discouraging physicians from ordering imaging studies that have low value or yield, i.e., those tests that won’t improve the patient’s outcome and will just drive up health care spending. Physicians are known to be risk averse in situations like “If I don’t order this imaging study when it might detect the patient’s problem (no matter how improbable), then it’s simply not worth taking a malpractice risk by omitting it. In hindsight, failure to order the test that then leads to a seriously bad outcome would look very bad in court.” And who can blame them?
Furthermore, it’s interesting that nurse practitioners and physician assistants order about 5 percent more imaging studies per emergency room visit than physicians (Rula 2024). Do they interpret the standard of care differently from physicians? Are they simply less certain of themselves and are committed to leaving no stone unturned? Are they pressured to meet productivity targets that are less important to their physician supervisors? Whatever the answer, one commentator opined that their ordering imaging studies amounts to an additional one million imaging studies per year (Rula 2024).
A second, very likely factor driving overutilization is human nature and our dislike of uncertainty. For both clinicians and patients, those feelings of uncertainty are obviously unpleasant, and they show up abundantly in the infamous instances of “incidentalomas.” Although follow-up care on incidentalomas can be costly and usually constitutes a low-yield effort, you never know when an incidentaloma represents something very serious. And of course there’s a lot of self-interest at play: If I were the patient and an incidentaloma was reported on my imaging study, I’d certainly want it checked out no matter how improbable the value of a follow-up might be. That’s why I have insurance, right? Again, the standard of care—which might discourage further testing on scientific grounds—is remarkably poor at putting that uncertainty to rest, either for me or the clinician. Why? Because the standard is based on probability estimates gleaned from thousands of patients. BUT NONE OF THEM ARE ME! So I will be naturally inclined towards self-protection and look to my physician as an ally.
As far as solutions to the work shortage problem go, it doesn’t appear that we can ask radiologists to take on more volume. First of all, radiologists seem to be at the outer limits of their performance capacity right now. But the peer reviewed and anecdotal literature also remark about widely differing characteristics among radiologists in terms of their reading speed, level of expertise, practice setting, differing workloads, differing imaging studies, and time of day (Alexander et al 2022). The consensus of that literature seems to be that we simply don’t know enough about how these variables play out across the landscape of diverse radiologic personnel to confidently make generalizations about an “average” radiologist’s effectiveness, accuracy, and speed. Consequently, we aren’t on good scientific grounds to even begin arguing that radiologists could or should assume more work.
The most popular recommendation seems to be increasing residency positions, which will require the co-operation of Congress. The American College of Radiology recently appealed to Congress to increase the number of federally supported residency positions by 2,000 annually for the next seven years (Stempniak February 9, 2025). We’ll see.
How about artificial intelligence? I think it’s fair to say the jury is still out. I recently talked to an eminent radiologic researcher, who told me that it would be another 10 years at least before AI starts fulfilling its hype and relieves radiologists of at least the easy, uncomplicated reads. On the other hand, hardly a day passes when I don’t read a news announcement of a new AI model that at least equals if not surpasses a human radiologist’s performance. My guess is that the first early benefit of AI will be its assisting clinicians with their documentation and communications. That by itself could be a tremendous relief as doctors and nurses are thought to spend as much as 30 percent of their time doing documentation. Alternatively, that relief can come with a significant downside if leadership then demands more productivity, given the time AI is saving clinicians from having to enter all that information on the medical record.
There’s an old Danish proverb that goes: “Making predictions is hard. Especially about the future.” We not only don’t know how long it will take AI to deliver on its promises, we also can’t predict how some of its other, currently unforeseeable impacts might play out, like its effect on the work force, on clinician-patient relationships, on legal liability attribution, and on patients.
One thing I’d very much like to see—although I doubt it will happen—is for state legislatures to enact liability protections for physicians who can show they complied with the standard of care in omitting to order some diagnostic imaging study. The idea would be that if a physician can reasonably show that he or she complied with a standard of care, as articulated for example in ACR’s appropriate use criteria (Ivanidze et al 2024), then any negligence suit against that physician would fail. This makes all the sense in the world—indeed, it’s a tautology—because following the standard of care means you didn’t commit a negligence. Thus, the plaintiff’s case, which requires they show the defendant-clinician committed an unreasonable or unjustifiable departure from the standard of care by not ordering the test, would be dismissed because the decision to withhold ordering that study would be considered non-negligent. Legislation that would protect physicians in that instance could go a long way to reducing the overutilization phenomenon, which accounts for a considerable percentage of the radiologist’s excessive workload.
The chances of my proposal coming into being, however, are slim. Plaintiff lawyers, who contribute lots of money to legislators’ campaigns and who control powerful lobbying interests, would robustly resist such a move, while legislators typically find tort reform difficult, especially when it comes to the practice of medicine (which they rightfully don’t want to legislate). Furthermore, the hindsight bias, which plaintiff lawyers don’t want to lose, plays a very powerful role in suggesting to jurors that negligence has occurred: The failure to order a test that arguably would have revealed an important finding affecting the patient’s subsequent care can appear damning. But all of that’s in hindsight. The reality of medicine, though, exists in a world of probabilities and human limitations. To expect its practitioners to perform at an errorless level of performance is simply unreasonable and unfair. To expect them to perform at the standard of care level is not.
Ultimately, I don’t expect the work overload situation to change, which is all the worse for radiologic well-being. We’ll now turn to what my interviewee, Dr. Elizabeth Rula has to say.
References:
Alexander R, Waite S, Bruno M, et al. 2022. Mandating limits on workload, duty, and speed in radiology. Radiology, 304(2):274-282.
Anonymous. 2024. Radiologists need to be realistic about the job market. KevinMD.com. Available at https://kevinmd.com/2024/10/radiologists-need-to-be-realistic-about-the-job-market.html.
Fornell D. August 22, 2024. Radiology workforce shortage a major concern for the American College of Radiology. Radiology Business. Available at https://radiologybusiness.com/topics/healthcare-management/healthcare-staffing/radiology-workforce-shortage-major-concern-american-college-radiology.
Fornell D. January 31, 2024. Radiology at tipping point with limitations of RVUs and the growing shortage of radiologists. Radiology Business. Available at https://radiologybusiness.com/topics/healthcare-management/healthcare-economics/radiology-tipping-point-limitations-rvus-and-growing-shortage-radiologists.
Fornell D. December 11, 2023. Putting US radiology reimbursement cuts in context and what comes next. Radiology Business. Available at https://radiologybusiness.com/topics/healthcare-management/healthcare-economics/putting-us-radiology-reimbursement-cuts-context-and-what-comes-next.
Ivanidze J, Shih RY, Utukuri PS, et al. 2024. ACR Appropriateness Criteria – Brain Tumors. American College of Radiology. Available at https://acsearch.acr.org/docs/3195155/Narrative.
Rula E. 2024. Radiology workforce shortage and growing demand: something has to give.
Available at https://www.acr.org/clinical-resources/publications-and-research/acr-bulletin/Radiology-Workforce-Shortage-and-Growing-Demand-Something-Has-to-Give.
Schartz E, Manganaro M, Schartz D. 2022. Declining Medicare reimbursement for diagnostic radiology: A 10-year analysis across 50 imaging studies. Current Problems in Diagnostic Radiology 51(2022):693-698.
Stempniak M. Feb. 12, 2025. Radiologist shortage will persist into 2025 without counteraction. Radiology Business. Available at https://radiologybusiness.com/topics/healthcare-management/healthcare-staffing/radiologist-shortage-will-persist-2055-without-counteraction.
Stempniak M. Feb. 9, 2025. ACR wants radiology prioritized in Senate bill to boost number of residency slots. Radiology Business. Available at https://radiologybusiness.com/topics/healthcare-management/healthcare-policy/acr-wants-radiology-prioritized-senate-bill-boost-number-residency-slots.
Elizabeth Rula, PhD
Executive Director, Harvey L. Neiman Health Policy Institute