Radiology Residency Spots Should Be Drastically and Immediately Reduced

This article was originally published in the Journal of the American College of Radiology - Volume 12, Issue 9, Pages 1131-1133, November 2015

by Colin M. Segovis, MD, PhD, Saurabh Jha, MD, Neil Lall, MD

At AMCLC 2014, the Resident and Fellow Section debated the motion, “Radiology residency spots should be drastically and immediately reduced.” A synopsis of the debate follows.

Point: Radiology Residency Spots Should Be Drastically and Immediately Reduced

By Neil Lall, MD

The market has changed. It is now not uncommon for radiologists to extend their training because of the difficulty in finding jobs at the ends of fellowships.

Imaging volumes have not increased to the extent we believed they would. The utilization of imaging in Medicare recipients plateaued from 2005 to 2009 and has decreased since 2010 [1]. High-deductible plans have grown in popularity. Greater cost sharing means that some patients forfeit elective imaging. Medicare is reimbursing less for imaging to reduce utilization [2].

Although volume and reimbursements have decreased, the number of practicing radiologists increased from 25,000 in 1999 to 31,000 in 2013 and continues to rise. To compound the supply, those nearing the ends of their career are delaying their retirement [3].

This surplus of radiologists not only leads to difficulty in finding jobs and a decrease in wages but hinders our undertaking new responsibilities. As experts in imaging, we are responsible for determining the appropriateness of studies, so that we can maximize the efficacy of our resources. However, because the number of new jobs is dependent on the volume of imaging, there is a disincentive to manage utilization, as that may decrease volume. A surplus of radiologists does not align incentives with the Imaging 3.0™ paradigm or with health care’s goal of decreasing costs.

It is true that there was a shortage of radiologists in the 2000s. This was because of fear that the health care reform of the 1990s would negatively affect radiology. This led to a contraction in the number of residency positions. Some feel that history will repeat itself if the number of residency positions is reduced again.

The comparison is inexact. The improvements in technology that made imaging central to the practice of medicine resulted in an unforeseen increase in demand, which was followed by a resurgence of interest among medical students [4]. Though such an advance could occur again, is it wise to stake the stability of our field on its possibility?

The shortage affected academic radiology most severely. Private practices offered higher starting salaries and short times to partnership, which academic groups were unable to match [4]. However, this situation was mutually beneficial, with some private groups funding fellowships for future hires [4]. Trainees were able to focus on their education, rather than hunting for jobs, and academic departments had help. With an excess of graduating trainees, there is little incentive for such collaboration.

We are seeing decreased interest in radiology among medical students. There were 6 unfilled residency positions in 2010 and 81 in 2014 [5]. The fear of overcorrection of surplus to shortage has convinced many that inaction is the best strategy to deal with labor imbalance.

The market for radiologists is not governed by the usual forces of supply and demand. Fewer jobs do not lead to a reciprocal decrease in the number of radiologists. Although, our field becomes less appealing to medical students, the residency positions will likely fill eventually, as there are 9,000 more medical students annually than positions in the match. However, as the desirability of radiology decreases, the quality of medical students will decline. We no longer will be training the brightest and the best. This is problematic, because the knowledge required of a radiologist is increasing, and many clinicians believe that they do not need radiologists to read imaging studies.

The data support oversupply. The annual workforce survey projected a 1:1 ratio of jobs to graduating trainees. In 2012, there were 1,069 jobs but 1,200 graduating trainees, yielding a surplus of >10% [6]. Making accurate predictions in a complex system is difficult. We will err to either a shortage or an excess of radiologists. A shortage of radiologists is easier to solve than a surplus. A shortage is mitigated by increasing productivity. A shortage can be solved by temporarily inviting non-US-trained radiologists.

A surplus makes it difficult for trainees to find employment. However, radiologists will not waste their training. Rather, they will accept salaries lower than customary. This risks a race to the bottom. Declining compensation is a problem for young radiologists, until a group loses its contract and seasoned radiologists enter the market. Oversupply affects all radiologists.

Counterpoint: Radiology Residency Spots Should Not Be Drastically and Immediately Reduced

By Colin M. Segovis, MD, PhD

“Uncertainty” dominates our field. It is unclear how reimbursement will change, if graduating residents and fellows will have jobs, and how the number of radiologists in the marketplace will affect reimbursement. Regardless, drastically reducing residency slots is reactionary and a potentially dangerous course with the potential to permanently change radiologists’ scope of practice as others fill the vacuum left by a future scarcity of radiologists.

Radiologists no longer set prices. Fee-for-service is the current model, but it is unclear how long fee-for-service will continue. Also, it is unknown how variables such as relative value unit valuation and the conversion factor will change. Regardless of future payment models, however, someone must read the imaging studies.

Despite uncertainty regarding future hiring, the 2014 workforce survey suggests that practices will hire radiologists [6]. The precise number of jobs is unclear; some sources state 1,100 jobs for 1,400 applicants, whereas others in the ACR leadership suggest a 1:1 applicant-to-job ratio [6]. The data suggest that jobs will continue to be available.

The ACR leadership is optimistic, and those who weathered the previous radiology job shortage have stated that new residency graduates will have jobs. It is unlikely that available jobs will satisfy all of an applicant’s desires, such as practice location, doing exactly what an applicant wants, and with exactly who the applicant wishes to have as colleagues. Still, it is likely that two out of three will be met.

The projection from data yields mixed results. From 2003 to 2011, there was decreased imaging utilization per Medicare enrollee on the basis of annual health spending on imaging per enrollee, but imaging spending per enrollee misleads our estimation of demand [7]. More important, the population is aging, with predictions that there will be 54 million individuals older than 65 years by 2020 and 72 million by 2030, compared with 40 million in 2010 [8]. Imaging utilization increases with age. Even if spending per enrollee on imaging has decreased, the total number of enrollees is increasing and could increase net imaging.

The utilization of imaging in emergency departments increased from 2002 to 2012; the use of CT increased by 159% (from 57.2 studies per 1,000 patients to 147.9 per 1,000 patients) [9]. The Patient Protection and Affordable Care Act will almost certainly add to the volume; net imaging demand will likely increase as individuals who did not have health insurance now obtain coverage. It is unwise to drastically reduce the number of radiologists when imaging volume will likely increase.

Some argue that there are enough radiologists if they are willing to work harder and faster to meet the demand. This is misguided. This philosophy suggests that volume does not interfere with quality. Specific metrics quantifying “quality” are still under development, but referring physicians expect a report that is correct and timely [10]. Additionally, failure to diagnose is responsible for a large number of lawsuits, and error rates have been reported as high as 30% [11, 12, 13]. Error increases with fatigue, yet the adage “work harder, work faster” is common [14]. Clinicians want more communication with radiologists, yet interruptions have a negative impact on report accuracy [15, 16]. Report timeliness, accuracy, and radiologist-referrer interactions are demanded by our patients and referring physicians. We are setting ourselves up for failure as a field if we believe that we can meet these demands by working faster.

If we decrease the number of residents, then we must answer the question of what happens if there are not enough radiologists to meet demand. The United Kingdom is currently facing this challenge [17]. Some have suggested expanding teleradiology [18]. Currently we are protected from overseas radiologists because of licensure requirements. A shortage of radiologists may lead to this protective barrier being lifted.

A drastic decrease in the number of residents is an alarmist’s prescription, which is not supported by data. Health care is changing, and we must adapt to the new system, but the utilization of imaging is unlikely to decrease as coverage expands and the population ages. Quality and value, not volume, are the new metrics. How quality and value are measured is still under debate, but decreasing the number of graduating radiologists will not increase quality. The ACR does not endorse decreasing the number of residency spots—neither should we [19].

The Moderator’s Perspective

By Saurabh Jha, MD

In considering whether to reduce the number of residency positions, there are three questions: What is the degree of oversupply? What are the consequences of reducing spots? How will residency positions be reduced?

All three questions are difficult to answer. The proponent of the motion hints at strong signals that a supply problem is brewing. However, the opponent counters that this cannot be asserted with partial information, notably because there is uncertainty regarding the actual demand for imaging, which may increase because of retiring baby boomers and more patients insured because of the Affordable Care Act.

Both caution that quality can be affected because of labor mismatch. Oversupply could reduce the incentives to manage utilization. Undersupply could discourage the adoption of noninterpretive work, as radiologists will be too busy reading films. Both agree that any mismatch is detrimental to the field: oversupply could reduce the quality of medical students, whereas undersupply might create a vacuum that could be filled by midlevel providers.

Neither side tackles the most difficult question: how will residency spots actually be reduced? Will this be voluntary? Who will give up the spots, community programs or large academic programs?

Labor mismatch is a wicked problem. No solution is palatable to all. Undersupply is easier to fix, but oversupply is easier to sustain. Inaction may not be a strategy but the path of least resistance. The most pertinent question is, what is the lesser of the two evils—over or undersupply of radiologists?


  1. Levin, D.C. and Rao, V.M. The declining radiology job market: how should radiologists respond?. J Am Coll Radiol. 2013; 10: 231–233
  2. Meltzer, C.C., Mullins, M.E., Chertoff, J.D., Bluth, E.I., and Canon, C.L. Are we training the right number of radiologists?. J Am Coll Radiol. 2013; 10: 757–759
  3. Moriarity, A.K., Brown, M.L., and Schultz, L.R. Work and retirement preferences of practicing radiologists as a predictor of workforce needs. Acad Radiol. 2014; 21: 1067–1071
  4. Dodd, G.D., Fletcher, T.B., and Thorwarth, W.T. The crisis in academic radiology: will we help ourselves?. J Am Coll Radiol. 2006; 3: 243–247
  5. Chen, J.Y. and Heller, M.T. How competitive is the match for radiology residency? Present view and historical perspective. J Am Coll Radiol. 2014; 11: 501–506
  6. Bluth, E.I., Truong, H., and Bansal, S. The 2014 ACR Commission on Human Resources workforce survey. J Am Coll Radiol. 2014; 11: 948–952
  7. Dodoo, M.S., Duszak, R., and Hughes, D.R. Trends in the utilization of medical imaging from 2003 to 2011: clinical encounters offer a complementary patient-centered focus. J Am Coll Radiol. 2013; 10: 507–512
  8. Administration on Aging. Projected future growth of the older population. Available at: Accessed April 6, 2015.
  9. Levin, D.C., Rao, V.M., Parker, L., and Frangos, A.J. Continued growth in emergency department imaging is bucking the overall trends. J Am Coll Radiol. 2014; 11: 1044–1047
  10. American College of Radiology. ACR practice parameter for communication of diagnostic imaging findings (revised 2014). Available at:∼/media/C5D1443C9EA4424AA12477D1AD1D927D.pdf. Accessed May 5, 2015.
  11. Berlin, L. Reporting the “missed” radiologic diagnosis: medicolegal and ethical considerations. Radiology. 1994; 192: 183–187
  12. Pinto, A. and Brunese, L. Spectrum of diagnostic errors in radiology. World J Radiol. 2010; 2: 377–383
  13. Renfrew, D.L., Franken, E.A. Jr., Berbaum, K.S., Weigelt, F.H., and Abu-Yousef, M.M. Error in radiology: classification and lessons in 182 cases presented at a problem case conference. Radiology. 1992; 183: 145–150
  14. Krupinski, E.A., Berbaum, K.S., Caldwell, R.T., Schartz, K.M., and Kim, J. Long radiology workdays reduce detection and accommodation accuracy. J Am Coll Radiol. 2010; 7: 698–704
  15. Jones J. Case study: survey says…. Available at:∼/media/ACR/Documents/PDF/Economics/Imaging3/CaseStudies/Imaging3_SurveySays_Final_102213.pdf. Accessed May 5, 2015.
  16. Balint, B.J., Steenburg, S.D., Lin, H., Shen, C., Steele, J.L., and Gunderman, R.B. Do telephone call interruptions have an impact on radiology resident diagnostic accuracy?. Acad Radiol. 2014; 21: 1623–1628
  17. Campbell D. NHS patients facing “unacceptable” wait for scan results. The Guardian. Available at: Accessed May 5, 2015.
  18. Strickland N. Regional teleradiology addresses U.K. radiologist shortage. AuntMinnieEurope. Available at: Accessed April 7, 2015.
  19. Bluth EI, Brink JA, Canon CL, et al. Rethinking residency. ACR Bulletin. Available at: Accessed May 5, 2015.

About the Author

This article was originally published in the Journal of the American College of Radiology - Volume 12, Issue 9, Pages 1131-1133, November 2015