Overutilization of Radiology Services

Blog #3 of Radiologist Well Being

This is the third 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.

Watch the interview with Michael A. Bruno, MD, MS, FACR below.

This blog is dedicated to radiologist well-being. As previous blogs have shown, a major factor affecting radiologists’ well-being is their work volume, which has risen dramatically every year for the last decade but without a sufficient number of radiologists entering the work force to absorb the increasing load.

One factor driving that ever-increasing workload is the overutilization of radiologic services, meaning the excessive ordering of “low value” imaging whose outcomes impact is usually, although not always, clinically negligible and usually, although not always, a waste of healthcare dollars (Brandsaeter et al 2023). Some estimates place that waste at $12 billion annually with as much as 4 to 30 percent of all American radiologic imaging being of low-value (Stempniak 2025). A study published in late March 2025 in the American Journal of Radiology added another consideration: That the environmental impact of overutilized imaging may be extremely high. The authors calculated that the carbon footprint of that overutilization appears equivalent to the yearly amount of electricity needed to service a town of 70,000 persons (Cavanagh et al 2025). And that only applies to imaging overutilization among Medicare Part B recipients. Total carbon emissions for radiology overutilization in the US may be 4 to 5 times higher.

Of course, a major driver of the overutilization phenomenon is defensive medicine, where a referring physician fears a malpractice suit for failing to order an imaging study even if no clinical evidence justifies it. But the matter isn’t just avoiding a lawsuit. In a marvelous 2024 article, Carly Stewart and her colleagues noted other overutilization factors such as:

· Radiologic detection of an X is more valued than the outcome of the detection;

· Reimbursement for the study is a revenue generator;

· Imaging studies are thought to be harmless (Stewart et al 2024).

Stewart’s observation that radiologic detection is valued more than outcome is important because it underlines—but I don’t believe entirely appreciates—a fundamental feature of radiologic performance and value: That the radiologist’s role is precisely to refine differential diagnoses or reduce diagnostic uncertainty (Kassirer 1989), not to order or determine the course of a patient’s ongoing care. Assuming the radiologist’s report is accurate and the patient’s eventual outcome will be due to variables and future events that are well beyond the radiologist’s control, we can understand how radiology might affect or inform a patient’s ongoing care but not directly determine it.

Furthermore, we not only have the referring physician’s fear of being sued for not ordering an imaging study, we also have the fundamental clinical insistence on getting the diagnosis right. Uncertainty and the therapeutic discomfort it provokes, experienced both by patients as well as their clinicians, will encourage image ordering until some decisive finding occurs—and, surprisingly, maybe not even then.

Consider the much-discussed case of “incidentalomas.” Incidentalomas are just that: Findings that are noted during a diagnostic or treatment intervention that are unforeseen as well as apparently unrelated to the study’s primary interest. Consider a very common one: the radiologist’s detecting an unexpected lesion in the adrenal glands while performing a CT or MRI study for abdominal or back pain (Chatzellis and Kaltsas 2024). In radiology, incidentalomas are thought to occur between 15 percent to 30 percent of the time for all diagnostic imaging tests and between 20 percent to 40 percent of all CT and MRI studies (Davenport 2023). Because it’s usually impossible to determine the importance of the incidentaloma at its initial detection, the clinical challenge then becomes whether the incidentaloma should be worked up or not.

And here’s where things get interesting. It’s unusual for incidentalomas to reveal anything clinically useful or worrisome unless they are associated with some underlying morbidity (Gaillard et al 2022). On the other hand, the risks of additional imaging are very real. Stewart et al (2024) lists the carcinogenic risk of ionizing radiation, pain and complications from IV lines, complications from the contrast materials, or additional false positive findings that lead to more workups.

Moreover and ironically, workups for incidentalomas often fail to relieve patient anxiety. The patient knows about the initial finding of the incidentaloma and wants the matter clinically or therapeutically resolved. But as the patient learns that further diagnostic testing reveals nothing, some will nevertheless harbor lingering anxiety that “something still isn’t right” (Rolfe and Burton 2013). That raises a number of interesting problems, with which I’ll conclude.

The first involves transparency. We assume that the radiologist will reveal the incidentaloma finding to the referring physician. Things get ethically interesting, though, if the referring physician believes that the incidentaloma is indeed insignificant or irrelevant and decides the patient need not know about it. I had such an incident occur about 15 years ago when I was experiencing chest pain and was referred for a stress test. The test showed a depressed ST segment, which prompted the cardiologist to have me undergo an angiogram that was negative. When I met with my primary care physician weeks later and told her about it, she looked at my EKGs over the years and said “You’ve had that depressed ST segment for the last 20 years I’ve been treating you. It’s nothing.” Now maybe she did tell me about that depressed ST segment years before, or maybe she didn’t because she thought it unimportant. The question this raises, though, is how much a patient needs to know or how much a physician should reveal from testing results, even if the physician thinks it insignificant.

We’ve already mentioned a second problem: The fear of a lawsuit if a test isn’t ordered, even if the referring physician believes the test will be of little if any treatment value. I can certainly understand this. Ten years ago, Anupam Jena and his colleagues (2015) found that physicians who robustly practice defensive medicine are sued considerably less often than ones who aren’t. Furthermore, if the patient is told about the incidentaloma but the physician shrugs off its importance and doesn’t pursue it, he or she will be a sitting duck for a medical malpractice suit if that testing or treatment omission directly results in the patient’s experiencing a serious adversity.

Which leads to a third, really interesting question: Suppose I was a very conscientious physician who wanted to do right both by my patient but especially not burden our already too expensive healthcare system, or our environment, or our overworked radiologists with unreasonable or unnecessary demands. What yardstick should I or my specialty use to determine when it is “reasonable” to withhold diagnostic testing for an incidentaloma or some other occurrence that I (and my colleagues) believe does not merit additional testing? Another way of saying this, as Stewart and her colleagues do, is “What is a reasonable miss rate?” that clinicians can use to justify their refusal to order a follow-up test for an incidentaloma? Suppose 99 out of 100 follow-up imaging studies would fail to make a clinically relevant finding for a given incidentaloma. Is that 1 percent miss ethically and legally acceptable?

I use a 1 percent miss rate because Stewart et al like it, and I’ve seen it before, especially in scenarios where there’s a 1 percent or less risk of X associated with the treatment under consideration (Than et al. 2012). Since we can’t tell the patient all the risks associated with a diagnostic test or treatment under consideration, might the 1 percent or less threshold serve as a reasonable cutoff that distinguishes risks that legally and ethically should be disclosed or followed up from ones that need not?

In any event, there’s good reason to believe that even these low-risk rates don’t deter referring clinicians from ordering follow-up radiologic testing. And that raises the final question of what can yet be done to relieve the radiology workload and our financially strained health system.

Because 1) the incentives for overutilization often appear to outweigh arguments for reducing image ordering and 2) we have no reliable legal standards to protect physicians who want to reduce their imaging ordering but fear a lawsuit if they do so, I believe it highly unlikely that the overutilization phenomenon will improve. Certainly, a huge change agent would be legislatively authorized immunization of physicians who do not order “unnecessary” testing because that testing is known to be, well, unnecessary. But then we’d need “reasonableness standards” that define precisely what those parameters of reasonableness or unnecessariness are. Interestingly, many are already out there like the American College of Radiology’s Appropriateness Criteria (https://www.acr.org/Clinical-Resources/Clinical-Tools-and-Reference/Appropriateness-Criteria) that classifies screening exams into categories like “usually appropriate,” “may be appropriate,” “usually not appropriate.” Should the radiologist in his or her report to the referring physician then indicate that “there is no evidence to support further radiology imaging for the present unexpected finding”? There doesn’t seem anything wrong with that. But how much weight, if any, would that note have in a medical malpractice proceeding alleging a negligent failure to order another imaging study? Would most referring physicians find it compelling and not pursue the incidentaloma? I wonder. But until such diagnostic assertions or recommendations from the radiologist are supported by some legislatively sanctioned immunizing or protective effect on the referring physician’s decision to withhold further diagnostic testing, the threat of a lawsuit would continue to be powerful.

Of course, insurers could and do simply refuse to pay for unnecessary testing on the grounds that it is neither “reasonable nor necessary.” But that doesn’t automatically relieve the treating clinician from liability for failing to order a test or intervention, especially in an emergency. Physicians know this, of course, so they will typically perform the intervention and then deal with the reimbursement dilemma later.

Ultimately and at least for right now, it seems to me that the phenomenon of overutilization and the extent it contributes to radiology burnout will not change. But let’s see what my guest Dr, Michael A. Bruno thinks.

Michael A. Bruno, MD, MS, FACR is a Professor of Radiology and Medicine and Director of Quality Management Services for the Department of Radiology at the Penn State Milton S. Hershey Medical Center, in Hershey Pennsylvania. He is a 1982 graduate of The Johns Hopkins University and the University of California, Irvine School of Medicine, where he earned his M.D. degree, along with a Masters’ of Science Degree in Biophysics, in 1987. He completed his residency in Diagnostic Radiology at UC Irvine in 1992 and a Fellowship in Nuclear Radiology at Vanderbilt University in 1996.

Dr. Bruno is Certified in Diagnostic Radiology (1992) with Special Competence in Nuclear Radiology (1997). He became a Fellow of the American College of Radiology in 2012. Dr. Bruno is the co-author of two textbooks, including, most recently Quality and Safety In Radiology, published in 2012 by Oxford University Press. He has also authored or co-authored several book chapters, numerous journal articles, and has created several enduring online educational materials. He is a frequent speaker on Quality and Safety and Musculoskeletal Imaging Topics throughout the United States and internationally.

Dr. Bruno’s current research focuses on mechanistic causes of radiologists' perceptual errors and enhancement of clinical decision support to improve the appropriateness of imaging utilization. He is a Fellow of the American College of Radiology, past chair of the ACR Emergency Radiology Committee and the ABR Core Committee on Radiology Quality & Safety. He is also a member of the ACR Intersociety Committee, ACR Advisory Group Taskforce on Artificial Intelligence, the SIDM Research Committee and the American roentgen Ray Society e-Learning Committee.

References:

Brandsaeter I, Andersen ER, Hofmann BM, et al. 2023. Drivers for low-value imaging: a qualitative study of stakeholders' perspectives in Norway. BMC Health Services Research 23(1):295.

Cavanagh G, Schoen JH, Hanneman K, Rula EY. 2025. Excess greenhouse gas emissions associated with inappropriate medical imaging in the US Medicare Part B Population from 2017 to 2021. Available at the Excess Greenhouse Gas Emissions Associated With Inappropriate Medical Imaging in the US Medicare Part B Population From 2017 to 2021 - Journal of the American College of Radiology.

Chatzellis E, Kaltsas G. 2024. Adrenal Incidentaloma. In: Feingold KR, Ahmed SF, Anawalt B, et al., editors. Endotext [Internet]. South Dartmouth (MA): MDText.com, Inc.; Available at https://www.ncbi.nlm.nih.gov/books/NBK279021/.

Davenport MS. Incidental findings and low-value care. 2023. American Journal of Roentgenology 2023;221(1):117-123.

Gaillard F, Knipe H, Botz B, et al. Incidentaloma. 2022. Reference article, Radiopaedia.org. https://doi.org/10.53347/rID-10832. Available at https://doi.org/10.53347/rID-10832

Jena AB, Schoemaker L, Bhattacharya J, Seabury S A. 2015. Physician spending and subsequent risk of malpractice claims: observational study. BMJ 351:h5516 doi:10.1136/bmj.h5516.

Kassirer JP. Our Stubborn Quest for Diagnostic Certainty. 1989. New England Journal of Medicine 320(22):1489-1491).

Rolfe A, Burton C. 2013. Reassurance after diagnostic testing with a low pretest probability of serious disease: systematic review and meta-analysis. JAMA Internal Medicine 173(6):407-416.

Stempniak M. 2025. Unnecessary imaging wastes $12B a year and uses enough electricity to power a small town. Available at Unnecessary imaging wastes $12B a year and uses enough electricity to power a small town.

Stewart C, Davenport MS, Miglioretti DL, Smith-Bindman R. 2024. Types of evidence needed to assess the clinical value of diagnostic imaging. NEJM Evidence 3(7):1-15. DOI: 10.1056/EVIDra2300252.

Than M, Herbert M, Flaws D et al. 2013. What is an acceptable risk of major adverse cardiac event in chest pain patients soon after discharge from the Emergency Department?: a clinical survey. International Journal of Cardiology 166(3):752-4. DOI: 10.1016/j.ijcard.2012.09.171.