Secondary Logo

Journal Logo

Invited Commentaries

We Must Graduate Physicians, Not Doctors

Dewan, Mantosh J. MD; Norcini, John J. PhD

Author Information
doi: 10.1097/ACM.0000000000003055
  • Free


Names are carefully pronounced. Medical students are lovingly hooded by someone special. A third-generation doctor follows one who is the first in her family to attend college. Each graduate walks across the stage, shaking hands with the dean, and the auditorium erupts in celebration. Every year, thousands of medical students complete the long journey to become doctors of medicine as proud family members and faculty look on. Most of these new doctors will enter hard-earned residencies and can almost touch their dream of becoming a physician–healer. But too many U.S. citizens are doctors who will never become physicians because they are unable to obtain a residency, which is required to convert a doctor into a licensed physician.

Every year, many U.S. medical schools have a handful of students who do not match. We know this at commencement, so amid what is otherwise a joyous occasion, there are moments of sadness, disappointment in our system, and a challenge to improve it. It is infinitely more wrenching for the unmatched doctor who may never become a physician—as one said, graduation is a happy time, but one eye is crying.

The Gap Between Undergraduate and Graduate Positions

In 2018, 4,099 applicants (1,078 U.S. allopathic medical school seniors, 846 students/graduates of U.S. osteopathic medical schools, and 2,175 U.S. citizen students/graduates of international medical schools [USIMGs]) failed to obtain a postgraduate year 1 (PGY-1) position in the Main Residency Match; some subsequently obtained positions through the Supplemental Offer and Acceptance Program (SOAP).1 The total number and number of unmatched U.S. allopathic senior, U.S. osteopathic student/graduate, and USIMG applicants in the Main Match were similar in 2019: 4,170, 1,162, 925, and 2,083, respectively. In 2018, 1,171 unfilled positions were offered in the SOAP and 1,053 residents were accepted to these positions; of these, 581 were U.S. allopathic medical school seniors and 211 were osteopathic students/graduates. Every year, a few others get positions after the Match.1 Accurate numbers or characteristics of those who are unmatched after the SOAP are not available, but we can reasonably estimate this total to be 3,000: approximately 1,000 U.S. graduates and 2,000 USIMGs. Clearly, the gap between the number of graduating doctors and available residency positions in the United States is a yawning chasm.

Defining the Problem

First, it is important to accurately understand the magnitude of the problem. To avoid gaps in their CVs, some unmatched students delay graduation by a year and enhance their CVs with elective experiences; others graduate and join a master’s program (e.g., in public health or business administration) or do research. All of these activities incur additional personal expense. Anecdotally, we observe that many (but not all) unmatched doctors reapply through the Match, do not match, do not become physicians, and are lost to the clinical workforce. We urgently need to convert anecdotes into factual data. For instance, combining American Medical Association and National Resident Matching Program (NRMP) files will allow us to track the number and medical schools of graduates who are unmatched after the Match, after SOAP, and in each subsequent year. This will refine our current working estimate of 1,000 U.S. graduates and 2,000 USIMGs per year going unmatched. These are already significant numbers. More concerning, several factors indicate these numbers will continue to grow. Established medical schools are intentionally increasing class sizes, and new medical schools are opening in an attempt to counter the anticipated physician shortage.2 But, to achieve this goal, medical schools must produce physicians, not doctors.

The Plight of the Unmatched Doctor

A doctor who graduates without a residency spot is not a physician and cannot practice. This doctor cannot examine a patient even with supervision, which is less than they could do as a medical student. (The exception is the “assistant physician” role in a couple of states, which allows some doctors to enter supervised practice on a limited license.)3 An unusable medical education is expensive in terms of time (opportunity costs of 4 rigorous years) and money (the median cost of attendance for the class of 2018 was $243,902 in public schools and $322,767 in private schools).4 The median debt for the class of 2017 was $192,0004; many unmatched students take longer and carry greater debt. For the dedicated underrepresented student who has overcome great personal odds and incurred crushing debt to become a doctor, it is especially painful not to go on and practice as a physician. It is even worse to hear faculty testify in some cases that the student is excellent with patients but has trouble with examinations, which handicaps their chance of getting a residency.

The Need to Graduate Physicians, Not Just Doctors

This personal tragedy also has a societal cost. There is a desperate need for clinicians: the Association of American Medical Colleges predicts a shortfall of 112,000 physicians (49,300 primary care physicians and 72,700 specialists) in 2030.5 Producing approximately 3,000 additional physicians—our estimated number of unmatched doctors each year—would not only provide care for hundreds of thousands more patients but also prevent the waste of 12,000 cumulative years spent in medical school and $750 million invested in medical education each and every year. Of note, U.S. medical school graduates make up one-third (about 1,000) of these 3,000, representing a potential waste of 4,000 medical school years and $250 million. Only a negligible fraction of unmatched doctors do not consider their education a waste, since they planned not to practice medicine but to leverage their degree in business, consulting, administration, or other non-practice ventures instead.

Potential Solutions for Graduating More Physicians

We have an opportunity to develop thousands more doctors into physicians every year. The problem we described does not exist in many countries, where all experiences necessary for licensure are included in requirements for graduation as a general practitioner (GP). In the United States, we, too, can design medical schools to graduate physicians, not just doctors. We present 2 practical models here.

Guaranteed minimum training for licensure

Obtaining and completing one year of residency is the rate-limiting step toward getting a license to practice as a physician. One year of residency allows U.S. graduates to practice in most but not all states; about 15 states require 2 or more years of residency. For USIMGs, most states require 3 years of residency.6 U.S. medical schools could guarantee 1 year of a primary care or transitional residency that students would enter by choice or as a default if they do not match into another residency. We would need about 1,000 new residency slots to accommodate unmatched allopathic (94.3% of 18,818 U.S. seniors matched into a PGY-1 residency in the Main Match, plus some more in the SOAP in 20181) and osteopathic (81.7% of 4,617 U.S. seniors/graduates matched, plus some more in the SOAP1) students. The more challenging fact is that another 2,000 three-year residency slots would be needed for U.S. citizens who are international students, primarily in Caribbean schools (in 2018, only 57.1% of 5,075 USIMGs matched).1

Can we afford this? If medical schools agree to graduate physicians and guarantee at least 1 year of residency for all students, can we collectively afford 1,000 new positions for U.S. graduates? Or 3,000 positions to accommodate all unmatched U.S. graduates and USIMGs? It is often lamented that the 1997 cap on federally funded residency slots prevents training additional residents; however, PGY-1 residency slots continue to grow (20,602 in the Match in 2002; 32,194 in 2019, the 17th consecutive annual increase).1 This is partly driven by data showing that residents pay for themselves7 and by creative arrangements, such as regional hospitals supporting residency training after which the graduate “pays back” the cost by working in this shortage area.8 However, recent increases may reflect the migration of an unknown number of osteopathic positions, which used to be exclusively in a separate American Osteopathic Association (AOA) match, into the NRMP. All residency slots—allopathic and osteopathic—will be in one NRMP Match starting in 2020. Separately, a major boost could be provided by just-introduced legislation (H.R. 1763, the Resident Physician Shortage Reduction Act)9 to have the federal government pay for an additional 3,000 positions each year. This bill deserves strong support.

Abbreviated medical school curriculum plus required residency year

Another model could be to replace the traditional 4-year undergraduate medical education curriculum with a 3-year curriculum plus a required 1-year residency in primary care. Three-year medical schools already exist.10 The fourth year of the traditional curriculum has been described as unstructured and wasteful; replacing it with a residency year would provide trainees experience in independent practice, boost their confidence, and allow certification of their clinical competence.11 It would also enable doctors who graduate from U.S. medical schools to obtain a license and become physicians at graduation without changing current licensure requirements. Paying these residents a salary equal to current first-year residents’ salaries would be equitable and would decrease the cost and time of medical education for students. The institution would need to support the residency year if residents were to receive a salary. In many countries where the final “internship” year (as distinct from residency) is part of the medical school curriculum, the student–doctor (“intern”) is paid a token amount—less than a resident—decreasing the cost for both the student and the hospital. Irrespective of the fiscal model, this year would need to provide strong education and training that enables students to pass all requirements for licensure. Unfortunately, although this model would allow U.S. graduates to practice in many (but not all) states, it does not help USIMGs who would still need 3 years of residency.


Not everyone agrees that every medical school graduate should be a physician and that medical schools must make accommodations to ensure all their graduates become physicians. In informal discussions with chairs, program directors, and faculty, specific and separate concerns were raised about graduates of U.S. medical schools and USIMGs. We were told that some unmatched students from U.S. medical schools are weak, often graduate after multiple, barely remediated failures, and cannot be trusted to provide safe care to patients.

There were more global concerns regarding USIMGs. We heard repeatedly (sometimes in hushed tones) that large numbers of U.S. citizens are graduating from international, primarily Caribbean, schools, some of which are considered substandard.12 Although there are excellent graduates of Caribbean schools in our residencies, on our faculty, and in practice, graduates of Caribbean schools tend to have lower United States Medical Licensing Examination (USMLE) scores than U.S. graduates,13 and their clinical outcomes tend to be poorer than either non-U.S. IMGs or U.S. graduates (e.g., a higher mortality rate in patients with cardiac failure or myocardial infarction in their care),14 which leads to a pervasive impression that Caribbean schools have limited access to quality clinical training sites. Variability among Caribbean schools is well documented. For instance, the Educational Commission for Foreign Medical Graduates (ECFMG) certification rates at individual Caribbean schools range from 28% to 86%, and specialty board certification varies widely.12 It is not clear what percentage of U.S.-born students graduate out of those who start at Caribbean schools, but, of the graduates, 43%—about 2,000—do not match.1 Collectively, these 2,000 U.S. citizens have invested—and now cannot use—half a billion dollars and 8,000 education years. Since many have taken federal loans, which are now difficult to repay, the Department of Education was charged with verifying and disseminating participating foreign medical schools’ student debt, graduation rates, and USMLE pass rates.13

Program directors confided to us informally that they strongly prefer excellent non-USIMGs over weaker U.S. graduates or weaker USIMGs, a preference that is apparent in the increasing numbers of successful non-USIMGs in the NRMP Main Match (3,692 in 2018 and 4,028 in 2019).1 This increase comes despite the steadily decreasing number of non-USIMG applicants, down from 7,460 in 2016 to 6,869 in 2019.1

We heard that we must advocate for the public good and not allow weaker U.S.-citizen doctors to practice, especially if weaker U.S. graduates can enter primary care practice after just 1 year of residency in most states. First, the faculty told us they were opposed to the very idea of a GP, dismissively called “doctor-light.” These faculty firmly espoused ten Cate’s formulation of medical education: that the Flexner Report (1910)

reflected the prevailing practice of relying on undergraduate medical education as sufficient preparation for lifelong medical practice. . . . In the 21st century the medical degree, while still significant in its legal status, has become an intermediate station in a long educational trajectory, rather than an end point. . . .15(p966)

In fact, the purpose of medical training today has moved from readiness for independent medical practice (graduating physicians) to readiness for postgraduate training (graduating doctors).

Second, program directors argued that primary care is perhaps the broadest, most difficult field in which to provide safe care, and even 3 years of residency training may be insufficient to achieve competence. At least one program has extended their family medicine residency to 4 years. We reminded them that the minimum education and training, and supervised clinical experience of physician assistants (6 years, 45 weeks) and nurse practitioners (6 years, 27.5 weeks) required to practice in primary care is markedly less than even that of a physician from a 3-year medical school with a 1-year residency (8 years, 110 weeks).3 However, this did not soften program directors’ opposition, as they saw it as markedly decreasing standards for becoming a physician. Resolving this contradiction will require reflection and dealing with educational and regulatory scotomas.3 Combining their philosophical objection against producing GPs, the difficulty of being a good primary care clinician, and weaker students led to significant concerns and opposition.

The Path Forward

Under the current system, each year sees a terrible waste of thousands of unmatched doctors while we face an enormous shortage of physicians. It is therefore imperative that medical schools make it their goal to produce physicians, not doctors. We need to dismiss the idea that a GP is a “doctor-lite” and unworthy of being called a physician. A commitment to graduating only physicians will require several steps. First, we need to generate accurate data so as to define the extent of the problem. This will become easier once the AOA match merges with NRMP in 2020. Second, as strong advocates for the public good, if we judge a student cannot be trusted to be a physician, we must decide this as early as possible in training and redirect that student away from medicine, not graduate a doctor by default. Third, we must return to the purpose of graduating only physicians but, unlike the Flexnerian era, offer multiple paths simultaneously: the majority of physicians will continue on to specialty training and a smaller but significant cohort will complete 1 year of residency and practice as much needed GPs. This will require reflection and a change in our mindset; the GP is not a discount physician but an essential and valuable practitioner in most parts of the world. For instance, GPs make up 31% of all physicians in the United Kingdom.16 Fourth, it is not enough to present data on poorly functioning schools as a “buyer beware” strategy. We must instead insist on medical schools achieving and maintaining high standards. Recognizing the variability in the quality of international medical schools, the ECFMG 2023 initiative was launched in 2010. It aims for fully functioning international accreditation recognized through an entity such as the World Federation for Medical Education by 2023, and may be an important way to reduce the number of medical school graduates who should not be physicians by both raising standards and decertifying failing medical schools.17 More challenging is the notion that medical schools should be measured by meaningful outcomes that are both subjective (e.g., a range of clinical sites that are well staffed by highly qualified academic faculty with protected time to teach) and objective (e.g., a 90% rate for passing USMLEs on the first attempt, obtaining a residency in the year of graduation, and obtaining an unrestricted license). Eventually, these will be replaced with the most fundamental and meaningful measure18: “differences in clinical outcomes for patients cared for by these physicians.”14(p1462) The residency match rate will become superfluous if a year of residency is included in the curriculum or guaranteed by the medical school. However, the other 2 measures can remain and serve to protect our patients as well as the next generation of aspiring physicians.

Even a few fresh graduates with one eye crying compel us to repurpose medical education. The end point of medical school cannot be a piece of paper stating the graduate is a doctor of medicine if that graduate may never be able to earn the privilege of being a licensed physician with the joy of practicing clinical medicine. To protect both our patients and our students, we must ensure that we graduate all our students not as doctors but as trustworthy physicians from highly regarded medical schools. We have the models and means to avert this personal and societal tragedy. We cannot stand by and watch as up to 3,000 U.S.-citizen doctors, 12,000 medical education years, and three-quarters of a billion dollars drown in tears each year.


1. National Resident Matching Program. Results and Data: 2019 Main Residency Match. Accessed October 21, 2019.
2. Smith-Barrow D. Map: Where to find the newest medical schools. US News and World Report, July 11, 2016. Accessed September 27, 2019.
3. Dewan MJ, Norcini JJ. Pathways to independent primary care clinical practice: How tall is the shortest giant? Acad Med. 2019;94:950–954.
4. Association of American Medical Colleges. You can afford medical school. Accessed September 27, 2019.
5. AAMCNews. New research shows increasing physician shortage in both primary and specialty care. April 11, 2018. Accessed September 27, 2019.
6. Federation of State Medical Boards. State specific requirements for initial medical licensure. Accessed October 11, 2019.
7. Iannuzzi MC, Iannuzzi JC, Holtsbery A, Wright SM, Knohl SJ. Comparing hospitalist-resident to hospitalist-midlevel practitioner team performance on length of stay and direct patient care cost. J Grad Med Educ. 2015;7:65–69.
8. Dewan M, Meszaros ZS, Manring J. Addressing the shortage of psychiatrists in rural areas. Psychiatr Serv. 2014;65:1508.
9. Resident Physician Shortage Reduction Act of 2019, H.R. 1763, 116th Congress.
10. Hartocollis A. NYU and other medical schools offer shorter course in training, for less tuition. NY Times, December 23, 2012. Accessed September 27, 2019.
11. Dewan M, Norcini J. A purpose-driven fourth year of medical school. Acad Med. 2018;93:581–585.
12. Norcini JJ, McKinley DW, Boulet JR, Anderson MB. Educational commission for foreign medical graduates certification and specialty board certification among graduates of the Caribbean medical schools. Acad Med. 2006;81(10 suppl):S112–S115.
13. van Zanten M, Boulet JR. Medical education in the Caribbean: A longitudinal study of United States Medical Licensing Examination performance, 2000-2009. Acad Med. 2011;86:231–238.
14. Norcini JJ, Boulet JR, Dauphinee WD, Opalek A, Krantz ID, Anderson ST. Evaluating the quality of care provided by graduates of international medical schools. Health Aff (Millwood). 2010;29:1461–1468.
15. ten Cate O. What is a 21st-century doctor? Rethinking the significance of the medical degree. Acad Med. 2014;89:966–969.
16. NHS Confederation. NHS statistics, facts and figures. Published July 14, 2017. Accessed October 11, 2019.
17. Tackett S. Examining the Educational Commission for Foreign Medical Graduates Announcement Requiring Medical School Accreditation Beginning in 2023. Acad Med. 2019;94:943–949.
18. Dewan M, Walia K, Meszaros ZS, Manring J, Satish U. Using meaningful outcomes to differentiate change from innovation in medical education. Acad Psychiatry. 2017;41:100–105.
Copyright © 2019 by the Association of American Medical Colleges