Perverse Incentives: A Challenge for Graduate Medical Education

Over the past two decades, the residency placement process has seen an inexorable intensification. In 2005, the average U.S. medical school graduate submitted 30 Electronic Residency Application Service (ERAS) applications. By 2015, that number rose to 46. The most recent (2022-23) residency application cycle featured an astounding average of 78 ERAS applications submitted per U.S. medical school graduate.1ERAS Statistics, Association of American Medical Colleges (https://www.aamc.org/data-reports/interactive-data/eras-statistics-data)Google Scholar This ever-increasing number of applications feeds the perception that attainment of a residency match has become increasingly challenging, while simultaneously resulting in an increasingly stressful, costly, and inefficient process for all the stakeholders involved.2Mott NM Carmody JB Marzano DA Hammoud MM. Breaking down the residency match rate.New Engl J Med. 2022; 386: 1583-1586Crossref Scopus (2) Google Scholar The root cause of this residency placement fever is not a relative glut of applicants. The ratio of first-year resident positions offered per applicant has been stable. Between 2018 and 2022, that ratio hovered between 0.80 and 0.85 despite a marked increase in the number of applications filed.1ERAS Statistics, Association of American Medical Colleges (https://www.aamc.org/data-reports/interactive-data/eras-statistics-data)Google Scholar Since multiple applications confer a relative advantage upon an individual over other applicants, it becomes an understandably preferred strategy among medical students. This dynamic epitomizes the so-called “prisoner's dilemma,” a situation in game theory in which individual decision-makers are incentivized to make choices that create a suboptimal outcome for the group. It is an extant incentive structure that largely accounts for residency placement fever,3Warm EJ Kinnear B Pereira A Hirsh DA. The residency match: Escaping the prisoner's dilemma.J Grad Med Educ. 2021; 13: 616-625Crossref Scopus (3) Google Scholar with poorly calibrated incentives that reward individuals or institutions for behavior that ultimately harms applicants and residency programs. Consider the financial implications of the overapplication to residency programs. The Association of American Medical Colleges (AAMC), which operates ERAS, collects a fee for each submitted application. Assuming a conservative 78 applications submitted per U.S. MD applicant in the context of 49,985 ERAS-registered applicants for the 2023 Match, the aggregate cost to U.S. medical students is in excess of $90 million. In addition, overapplication increases labor market congestion, rendering the selection process inefficient through the accumulation of time spent on application review. This, in turn, renders the holistic review of applicants challenging if not impossible. Another consequence is the disproportionate assignment of interview slots to a limited number of outstanding applicants. These circumstances may undermine longstanding efforts on the part of residency programs to diversify, a goal with additional implications given the recent Supreme Court ruling on affirmative action and its likely negative effect on diversity of the U.S. medical student body. In addition, market congestion may contribute to less favorable Match outcomes. The increasing length of applicant rank order lists over the past decade, a reasonable proxy for the number of interviews conducted by residency programs, has coincided with a decline in the proportion of applicants obtaining one of their top four residency choices.2Mott NM Carmody JB Marzano DA Hammoud MM. Breaking down the residency match rate.New Engl J Med. 2022; 386: 1583-1586Crossref Scopus (2) Google Scholar Meanwhile, the Match rate among U.S. MD applicants remained stable at approximately 94% for the last twenty years, further illustrating a lack of benefit to applicants and programs alike. A system permissive of overapplication also creates perverse incentives for students. In a congested labor market, applicants face more pressure to convey two types of signals so as to distinguish themselves, namely, that one is interesting and also interested in a residency program.4Roth AE Marketplaces, markets, and market design.Am Econ Rev. 2018; 108: 1609-1658Crossref Scopus (57) Google Scholar The recent incorporation of “program signaling” into the ERAS application for most specialties may assist in conveying the latter message. After all, the submission of an application no longer functions as a signal of strong interest given the ease of filing numerous electronic applications. To convey that one is interesting, medical students are confronted with ever-mounting pressure to showcase presumed indicators of their “applicant quality.” A problem arises when the incentives to generate these indicators are not calibrated to the most desirable training outcomes. For example, there has been a startling increase in the volume of medical student research output in the past decade.5Ahmed AK Adashi EY. Research fever-an ever more prominent trend in the residency match.JAMA Surg. 2023; 158 (Jan 1): 6-8Crossref Scopus (3) Google Scholar As a result, the sheer volume of research produced and listed by applicants to some highly competitive specialties now rivals that of a junior faculty member. A parallel trend has occurred in all extracurricular qualifications listed on ERAS. Between 2009 and 2022, the total number of volunteer and work experiences, along with research experiences, listed on ERAS has increased from 11 to 16. But is more necessarily better? The ramping up of extracurricular qualifications appears to have become a self-reinforcing requirement for a “competitive” residency application. While there may be genuine value in these endeavors to the development of physicians, there are also drawbacks and diminishing returns. For instance, there are financial costs in the form of research gap years, a cost with limited benefit for those who undertake research for its contribution to their residency application rather than a genuine commitment to research. Given that meaningful involvement in multiple scholarly and extracurricular activities appear at odds with the demanding nature of a medical school curriculum, it is noteworthy that the upcoming 2023-24 ERAS application cycle will feature a new cap of 10 extracurricular experiences. We consider this a welcome development that aims to incentivize quality over the quantity of experiences. Nonetheless, we also believe that additional steps are required to get the incentives right. First, overapplication is only possible because of the current infrastructure that supports the submission of an unlimited number of applications. The primary and most direct solution to the high volume of applications filed would be an across-the-board cap on the number of applications filed to be implemented by AAMC/ERAS. This would be similar to the limitation imposed by the Common Application, which permits high school students to file up to 20 college applications on its platform. Limiting the permitted number of applications could yield clear benefits. U.S. medical graduates who judiciously apply to a limited number of programs would likely be invited to interviews by a high proportion of programs. In addition, reduced congestion in the labor market may well enhance the prospects of attaining one's preferred outcomes. An application cap enables residency programs to infer a student's strong interest, thereby eliminating the need for “preference signaling.” Costs to applicants would be lowered, a particular benefit for those applicants who are financially less advantaged. Second, residency programs should help promote incentives that are better aligned with program goals. Accomplishing this would require greater clarity in the selection criteria used by residency programs, criteria that are at present murky.2Mott NM Carmody JB Marzano DA Hammoud MM. Breaking down the residency match rate.New Engl J Med. 2022; 386: 1583-1586Crossref Scopus (2) Google Scholar More transparent selection criteria from residency programs would be helpful in allowing medical students to apply accordingly, a change of particular importance if an application cap is implemented. Ideally, increased transparency would help mitigate the pressure on students to make themselves interesting on paper, promote a learning environment in medical school focused on growth instead of performance, and optimize the fit between a self-selected applicant pool and the values espoused by a residency program. The rules governing the resident labor market are important because they reflect the way that medical education imparts its values. By setting caps on applications and extracurricular activities, and by promoting transparency in the applicant qualities favored by residencies, medical education can encourage endeavors that matter most to students and residency training programs. We believe that the redresses suggested here can serve as a steppingstone towards the long-term promise of competency-based medical education, in which the incentives should focus on the development of a highly competent physician rather than a competitive applicant. Only through such actions can the system align student and graduate medical education program incentives and goals. None.