As I was preparing to write this editorial about the Match—which occurs every March and determines the residency programs and specialties for most graduating medical students across the country—I was reminded of the musical Fiddler on the Roof. It tells the story of a traditional Jewish couple in the early 1900s in Russia with three daughters, all of whom are considering future marriages. The dramatic tension of the story revolves around the questions of how each daughter will be matched with the appropriate husband, whether the parents will choose for them, and the role of an outside expert (the matchmaker), who will help negotiate the best match. Or will the daughters and the young men who have been secretly courting them be able to make their own decisions? There are many concerns about breaking tradition and the consequences of a bad match on the future for each of the daughters and their parents. But the mood of this musical is mostly optimistic and joyful.
One song addressed to the matchmaker goes, in part, “Find me a find, catch me a catch, … look through your book and make me a perfect match.” This might well be the plea of our fourth-year medical students, more than a hundred years later, who wait breathlessly on Match Day for the envelopes that will reveal their fates. Will it be the perfect match? Or will it be something less desirable or, even worse, a disaster, something unimaginable, like the paper with the black dot described in “The Lottery” by Shirley Jackson?1
That short story explores the darker side of group decision making. Most of the narrative describes details that seem ordinary and inconsequential—the village where the lottery takes place, the people who live in the village, which villagers will pick papers from a box of folded papers, and the moment when all the papers are unfolded. It is not until the end of the story that we find out the purpose of the lottery. One person will be chosen to die. It will be the unlucky one who picks the paper with the black dot. There is no explanation about why this might be necessary, but the townspeople seem not to question the tradition, the randomness, or the consequences. In the story the loser of the lottery is a woman who has not done anything that would merit such a punishment. “It isn’t fair,” she says as the townspeople set upon her and begin to throw stones at her. And the horror of the story is that we realize that this woman could be any of us, and that a random decision could determine our fate. It reminds us that terrible things happen even if we live a moral and upstanding life. As I reread the story, the contrast between the horror of the black dot and the joys of Fiddler on the Roof made me wonder which was the more appropriate image for the Match. How many students would be dancing in joy, and how many would trudge away sadly saying, “It isn’t fair”?
Of all the topics recently published in Academic Medicine, the selection process for residency has probably been the most controversial, inspiring critiques from medical students, program directors, student affairs deans, and graduate medical education (GME) deans. Almost everyone agrees that the process is expensive, inefficient, encourages excessive focus on the United States Medical Licensing Examination (USMLE) Step 1, does not adequately account for national workforce needs, and, because of all of those reasons and many others, does not seem to be fair. But is there a better way to select students for residencies? In this editorial I explore that question.
What Is the Problem?
Historically, the most serious problem with residency selection related to the chaotic system of applying for and offering internship positions, which reached a crescendo in the late 1940s when hospitals would offer internship positions and require an answer within 12 hours. Students did not know whether acceptance of an offer from one hospital might cut them off from a more desirable option, and because there was no agreed-upon day for making offers, there was pressure on hospitals to be the first to make an offer. This increasingly ungovernable system led to discussions about a better matching system for hospitals and students. Roth2 describes the problem that surfaced in 1949:
Hospitals were finding that if an offer was rejected after even a brief period of consideration, it was often too late to reach their next most-preferred candidates before they had accepted offers.
The solution that was developed was meant to eliminate telephone calls with time pressure for decisions and last-minute scrambles. The National Resident Matching Program was established in 1952 to match students into GME programs through a process that was “fair, efficient, transparent, and reliable.”3
Ironically, many of the problems that the Match was created to solve have reappeared today. Gruppuso and Adashi4 note that there has been an intensification of the residency selection process with increasing numbers of applications to programs, increasing numbers of interviews, increasing average scores on the USMLE Step 1 scores, and increasing participation in audition clerkships. They believe that the rise in numbers of applicants to U.S. GME programs has led to increased competition, including competition from graduates from non-U.S. medical schools, although graduates from U.S. medical schools have thus far maintained a consistent match rate between 92.1% and 95.1% from 1982 to 2015.4
The intensification of the selection process has also raised pressure on program directors to have a mechanism to screen candidates for interviews and develop a match list that will yield high-quality residents. The USMLE Step 1 examination has increasingly played the screening role for program directors, with specialties setting different minimum scores in the initial review of applications to help speed up the review process for granting interviews. Gauer and Jackson5 have demonstrated substantial differences in mean USMLE scores for students, depending on the specialty, ranging from 213 for family medicine residents to 244 for dermatology.
The USMLE Step 1 examination score has long been a highly regarded criterion for residency selection according to a survey of program directors by Green et al,6 who ranked the top five criteria as (1) grades in required clerkships, (2) USMLE Step 1 score, (3) grades in senior electives in specialty, (4) number of honors grades, and (5) USMLE Step 2 Clinical Knowledge score. The USMLE Step 1 score has been useful to program directors because the score can directly compare students from different medical schools,7 the score is available early in the application process, and there has been an association between scores on the Step 1 examination and performance on in-service and certifying examinations.8 Because of the increasing use of the Step 1 score in the selection process, rising average scores, and higher mean scores for certain specialties, medical students have increasingly focused their learning on educational materials that are aligned with performance on the Step 1 exam rather than attending to other important curricular activities.
Moynahan9 describes some of the consequences of focusing on the USMLE Step 1 exam. He raises concerns about the adverse effects of such a focus on holistic admissions to medical schools, which have been recommended to diversify the medical profession and are generally not based on performance on the MCAT past a certain threshold but, rather, on other criteria like life experience, resiliency, and desire to practice in an underserved community. He notes that “these same students who are carefully admitted to medical school through these holistic processes soon learn that their future will be heavily influenced by a single score on a standardized multiple-choice test.” Moynahan also states that the emphasis on Step 1 exam scores undermines the medical school curriculum.
Students largely dismiss as irrelevant content in the curriculum that they believe is not included on the Step 1 exam due to their focused attention on this exam…. At many medical schools, the curriculum presented in the months prior to the Step 1 exam might as well not exist; students put almost all of their effort into studying for the Step 1 exam. They routinely tell us that they feel there are two separate curricula they have to negotiate—their school’s curriculum and a separate one for the Step 1 exam. The never-ending stream of Step 1 study aids and programs that our students spend increasing amounts of money on are a testament to this second curriculum.
In this issue, Chen et al10 provide medical students’ perspectives on the impact of the USMLE Step 1 exam preparation industry and the emphasis on the preclinical learning environment, which they characterize as the “Step 1 climate.”
The Step 1 climate, arguably, cedes medical educational authority to the for-profit industry that creates Step 1 commercial resources. These resources now define the de facto national curriculum for preclinical medical education…. Ultimately, the emphasis on Step 1 raises profound questions about the purpose of medical school…. If commercial resources more effectively prepare students for a national preclinical competency examination, why should students spend time engaging with their institutional curriculum?
In this issue of the journal, Andolsek,11 in an Invited Commentary on Chen and colleagues’ essay, notes that
Step 1 was not designed to, nor does it, predict success as a resident. Its misuse has created a “Step 1 climate” inimical to learning, diversity, and well-being. We are collectively responsible for and must collaboratively solve this problem.
Previous research by McGaghie et al12 supports the assertions by Andolsek, demonstrating that USMLE Step 1 and Step 2 scores are not correlated with reliable measures of medical students’, residents’, and fellows’ clinical skills acquisition.
Andolsek provides actionable recommendations for the various stakeholders involved in creating the Step 1 climate and the current chaos in the Match process. She suggests that medical schools improve the information that they provide to residency programs about their medical students by encouraging more longitudinal relationships between students and faculty and making improvements in the Medical Student Performance Evaluation (MSPE). The assessments should be honest and transparent, and the handoff of the students should demonstrate that the student is ready for GME. She also suggests that more students be admitted to medical school with a commitment to GME at the same institution, allowing for a tailored curriculum and avoiding the anxieties and pressures of the Match and the USMLE test.
For the National Board of Medical Examiners, Andolsek suggests that changing the Step 1 test to pass–fail might be helpful but that program directors would likely find other ways of screening applicants. She wonders whether the exam could be redesigned to align with the competencies of interest to program directors or whether the exam might be replaced for licensing purposes with other tests developed by the American Board of Medical Specialties.
For program directors, she recommends increased collaboration with undergraduate medical education colleagues to share information about which attributes they are seeking in medical school graduates. She also notes that while programs rate the Step 1 scores highly in decisions about whom to interview, the Step 1 score is not considered as important a factor in ranking applicants, falling behind ethics, professionalism, specialty-specific letters, and grades. She suggests that resources that would provide the information that program directors most value should be developed. She notes that the Accreditation Council for Graduate Medical Education (ACGME) lowered the bar for the percentage of residents passing board exams, which should reduce the importance of Step 1 scores to affect future ACGME accreditation.
An Imperfect Solution
The problems with residency selection have been building for years. It has now spawned an industry dedicated to test performance for a test that is widely recognized as not being very important in the overall performance of future physicians. The focus on the USMLE Step 1 has adversely affected the learning climate at many medical schools. The good thing about this problem is that its solution is within the power of the various stakeholders involved in medical education. The federal government is not needed to solve this problem, unlike many other difficult academic medicine and health care issues. The bad thing about the problem is that it has persisted, which suggests that it will not be easy to solve. If it were easy, it would have already been solved. So what can we do?
I now return to the two works I introduced at the beginning of this editorial, Fiddler on the Roof and “The Lottery.” If we look at “The Lottery,” we see that all the attention of the townspeople was on process—who would pick a paper out of the box, when all the papers would be unfolded, and who would get the deadly one. There was no discussion about why it was happening, what the purpose of the lottery was, or whether it should be changed. I think we need to ask these questions about our current selection process. Are we trying to get the best test takers to become, say, dermatologists rather than family practitioners? Is that an important goal for our selection system? If so, we are succeeding. But if not, we have created a system with perverse outcomes.
Is our selection process helping us find those students who will best meet the needs of our population by going to the places where they are needed and practicing specialties that are needed, working collaboratively to meet the health goals of our communities? Do we have a selection system that encourages diversity as well as excellence? If not, perhaps a multiple-choice test score should not have a major influence in the process.
Moving to a pass–fail scoring system for all USMLE exams would diminish the influence of the test immediately. While it would create problems for program directors who have come to depend on it for rapid screening of candidates, they would quickly adjust and perhaps have an incentive to find other, better ways to screen applicants. Such a change would also force our various communities to consider which information is most important and how we might convey that across the continuum. In some cases, we might find more programs in which students maintained a relationship with an institution for both undergraduate medical education and GME. In others, we might find more emphasis on longitudinal relationships between teachers and students, perhaps involving a connection with the area of chosen interest by the student. The value of the MSPE might be enhanced so that the student’s experience in medical school could be fairly considered. The MSPE could be made more standardized if medical schools adhered to guidelines about what the information and words mean in their letters to program directors. The best predictor of future performance is past performance, and we need to be able to describe the various attributes of performance clearly and accurately.
Finally, we need to connect our selection system to our workforce needs. While payment differences between specialties may affect student decisions about career choice, it is also likely that the altruistic values that led many students to a career in medicine could continue to be utilized to drive specialty and location choices if linked to positive experiences in the learning environment. With the increasing recognition of burnout in residents and how it affects regret over specialty choice,13 we need to understand how to create better work and learning environments that will align with our needs for our future workforce. The Match, with its envelopes of winners and losers, is too reminiscent of “The Lottery.” I would much prefer the more joyful image of Fiddler on the Roof, where the decisions about a medical student’s future were connected to our community, to excellence, and to commitment to service. For those attributes to drive our selection process we need to know who our students are, what brought them to medicine, and what goals they have and how they could align with their career options. No multiple-choice test will answer those questions. But a humanistic educational environment, with faculty who truly get to know their students and establish trust with the students and the GME community, could.
I think our community knows what it needs to do, and there is literature that can help guide us.14 We can solve this problem if we have the motivation and courage to do so. We may not be able to guarantee a perfect match, but I think we can guarantee one that is compassionate, fair, and better than what we have now.
David P. Sklar, MD