An Initiative to Increase Residency Program Diversity

The Nationwide Children’s Hospital residency program and its faculty are academically affiliated with The Ohio State University College of Medicine but are supported exclusively by Nationwide Children’s Hospital, one of the largest pediatric health campuses in the United States. Our recruitment team included leaders from the department of pediatrics training program directors, URM faculty, and URM residents. Motivated by children’s rights activist Marian Wright Edelman’s assertion that “You can’t be what you can’t see,” we set out to increase visibility of URM faculty and trainees in multiple venues related to resident recruitment. We targeted efforts that could be sustained year after year, ensuring a diverse and engaged applicant pool motivated to match and train in our program.

Figure 1 is our cause-and-effect diagram identifying key drivers in workforce parity with our patient population. Funds were allocated from the department chair’s annual discretionary budget to support travel of residents to the Student National Medical Association (SNMA) and Latino Medical Student Association (LMSA) national meetings to increase visibility in these venues. Similarly, the chair’s annual discretionary budget was accessed to support 50% of a visiting student elective (VSE), with the other half coming from the hospital’s graduate medical education office. Marketing materials and booths at the SNMA and LMSA national meetings were supported by the marketing department’s annual budget. These initiatives targeted factors that align with many of those described in the literature.^{11–16 ,18–20} 

During the first intervention year (cycle 1; 2017–2018 recruitment year), we increased the visibility of URM faculty during recruitment, both during the interviews and in the prerecruitment season via participation in SNMA programming. For the first time, residency program directors received formal implicit bias training before the recruitment season, URM applicants were intentionally assigned a face-to-face interview by at least 1 URM faculty member, and a more holistic review of applicants was initiated, with increased emphasis on leadership, research, and community service experiences relative to board scores.

During cycle 2 (2018–2019 recruitment year), we implemented a VSE for fourth-year URM medical students targeting 4 to 5 visiting students annually. This program included a month-long subinternship on an inpatient clinical service, participation in residency teaching conferences, and assigned mentorship with both URM and interest-specific faculty. All eligible applicants met the Association of American Medical Colleges definition as racially and ethnically URM, and admission criteria were based on the applicant’s curriculum vitae, personal statements with focus on previous advocacy work, extracurricular activities, Gold Humanism awards, previous research experience, evidence of previous adversity, and grades from medical school. Standardized test scores were considered but not prioritized, and applicants from disadvantaged backgrounds were given additional consideration.

In addition, URM residents were visibly involved in recruitment efforts beyond traditional interview activities, such as interview dinners, hospital tours, and interview-day lunches in which all applicants typically participate. A URM Resident Recruitment Task Force advised the training program and participated extensively in recruitment. Each URM applicant was invited to preinterview dinner social engagements with existing URM residents. Interactions emphasized visibility, connectedness, and a candid discussion of the experiences of URM residents in the residency program, in the hospital, and in Columbus. Topics of discussion included treatment of URM residents by peers, ancillary staff, attending physicians, and patients; the importance of diversity to hospital administration and the residency program; the perceived inclusivity of URM residents with coresidents; and ethnic and cultural activities around the city, among others.

During cycle 3 (2019–2020 recruitment year), our URM Resident Recruitment Task Force transitioned to a more robust resident-led minority housestaff organization (MHO), named the Resident Diversity Inclusion Coalition, which was expanded to include lesbian, gay, bisexual, transgender, and queer faculty and trainees. The MHO emphasized the importance of year-long initiatives to improve the URM applicant pool. Visibility-oriented efforts included development of new print and online marketing materials and a social media account showcasing our diverse workforce. A formal diversity mission statement was developed for the MHO and approved by departmental leadership. We increased visibility at national meetings by hosting booths at the national SNMA and LMSA conferences and by participating in roundtables with local medical schools. During this period, the former residency program director was promoted to an institutional leadership post, and an associate program director who was well-positioned for the role and happened to be URM became program director.

MHO members became standing members of the Pediatric Residency Recruitment Committee. URM faculty continued to interview URM candidates, and URM applicants and residents continued to meet at preinterview dinner social engagements. We added a strategy of drop-in visits by additional URM faculty mentors from the MHO who were not already involved in the day’s interview. This intervention was designed to publicly highlight our commitment to diversity and inclusion with all candidates together in a group setting. This was considered an important element of our URM recruitment strategy. Finally, a second-look weekend, designed to facilitate networking, and to allow candidates to learn more about living in Columbus, was offered to URM applicants, and was organized and attended by URM residents.

During cycle 4 (2020–2021 recruitment year), our efforts became virtual because of the coronavirus disease 2019 pandemic. The MHO held virtual roundtables with URM medical students interested in pediatrics. Rather than having in-person social gatherings before interview dinners, preinterview social engagements were virtual. Additionally, because of more applicant interviews, the MHO held several URM-specific interview gatherings throughout the interview season. In collaboration with The Ohio State University College of Medicine, there were virtual URM gatherings for multiple programs, including pediatrics, internal medicine-pediatrics, and internal medicine.

Our primary outcome measure was the percentage of matched residents who were American Indian or Pacific Islander, Black or African American, or Hispanic, Latino, or of Spanish origin, as defined by self-identifiers provided by the Electronic Residency Application Service.

The percentage of URM residents who matched into our program before our dedicated initiatives was 5% (2 of 37) in the 2015 match, 3% (1 of 37) in the 2016 match, and 5% (2 of 37) in the 2017 match. After we implemented our targeted initiatives, the percentage increased to 16% (6 of 37) in the 2018 match, 26% (11 of 43) in the 2019 match, 19% (8 of 43) in the 2020 match, and 21% (9 of 43) in the 2021 match (Fig 2). Eighteen percent (2 of 11) of URM residents recruited in 2019, 13% (1 of 8) in 2020, and 11% (1 of 9) in 2021 were VSE students, respectively. The percentage of URM residents matched during our baseline period increased significantly over the intervention period (P = .0002, Fisher’s exact test) (Fig 3).

Interestingly, compared with previous years, we observed only a minor increase in the percentage of applications submitted by URM candidates to our program during the years after our intervention (Table 1): URM applicants were 8% of the total applicant pool during the baseline period, and 10% during the intervention period. The average percentage of URM applicants who accepted an invitation to interview was 76% during our baseline period, and 84% during our intervention period.

The costs of our interventions were ∼$7000 per year for VSE participants and $17 000 to support booths, sponsorship, and resident travel to the SNMA and LMSA national conferences. Because only a small number of URM trainees and faculty were available for activities during interview season, human resource costs were also assessed. Starting in cycle 2, we tracked the number of recruitment dinners attended by 2 URM residents (residents A and B) as a reflection of time burden, a balancing measure. During cycle 2, residents A and B attended 33% (5 of 15) of dinners each. Conversely, during cycle 3, resident A attended 6.7% (1 of 15) of dinners and resident B attended 13% (2 of 15). In these later years, the increased number of URM residents in our program decreased the individual obligation to these dinners.

We learned through the 2015, 2016, and 2017 recruitment seasons that well-meaning but (in retrospect) ambiguous intentions to recruit URM residents were ineffective to achieve our goal. During recruitment season 2017–2018 (cycle 1), we developed goal-directed, structured approaches to improve URM recruitment, resulting in improved outcomes.

More initiatives added in recruitment season 2018–2019 (cycle 2) further increased successful URM resident recruitment. We believe the pivotal difference was substantially increased URM resident participation in every phase of the recruitment process. Specifically, a group of enthusiastic, engaged URM residents led outreach efforts during the residency interview season, including organized, informal discussions regarding the lived experiences of URM residents working in our hospital and living in our city. Many applicants expressed misconceptions about a lack of diversity in our residency program, in our patient population, and in Columbus, which aligns with evidence that geographic region and the perceived geo-cultural environment impact the ability to recruit a diverse workforce.^11,21  Our URM residents provided a comfortable forum for discussion of concerns specific to URM applicants and to address any misconceptions.²²  Examples of discussion topics included feelings of tokenism, such as being singled out for differential treatment during their interview day because of their URM status; stereotype threat, such as being presented with representations of URM faculty and trainees in stereotypical contexts; and the extent of the microaggression culture at Nationwide Children’s.¹⁹  Additionally, the VSE allowed fourth-year URM medical students the opportunity to interact with URM residents and get first-hand exposure to the culture of the residency program, institution, and city. Cumulatively, these interventions, led by an engaged and dedicated team of residents sophisticated in the unique challenges of URM trainees, were thought to be the most impactful of the programs implemented.

Iterative changes to our recruitment strategies during cycle 3 (recruitment season 2019–2020) involved optimizing previous efforts and increasing our public-facing initiatives, such as conference attendance and thoughtful diversification of marketing materials. A well-designed print piece specifically highlighting our diverse faculty, URM trainees, and our culture of inclusion was developed and distributed to all applicants during the interview day.

Continuing our recruitment strategies during cycle 4 (recruitment season 2020–2021) was challenged by the pandemic. Virtual meeting fatigue decreased attendance of events outside of the standard interview day. The genuineness and sincerity of our diversity culture proved difficult to convey in a virtual format void of the high-touch practices in previous cycles. Notwithstanding, we were able to recruit URM residents at a level greater than our baseline.

Our work and experience builds on previously published URM workforce diversity efforts in 2 ways. First, drawing on our institution’s experience in the application of QI science, we are among the first to use an iterative process with repeat tests of change based on a QI model, rather than implementing large-scale change at once. Second, to our knowledge, our efforts represent the first to benchmark URM recruitment efforts to a region’s patient population demography.

Amid these efforts, an MHO was created, designed to provide support and a sense of community to our URM trainees. On the basis of evidence that burnout has a dose-dependent relationship with loneliness, a sense of community might have had a positive effect on resident resiliency, which was perceived by applicants as an institutional culture of inclusion.²³  Furthermore, creation of an MHO has been associated with an increase in the number of minority trainees at other institutions.²⁴  Although it is unlikely that our MHO alone is responsible for the improved URM recruitment results, it likely augmented these efforts. To create institutional memory, there is now an executive board for our MHO, which holds annual leadership nominations to ensure efforts are sustained. Finally, our efforts to diversify our resident workforce were focused not only on recruitment but also on increasing the pipeline of URM trainees pursuing pediatrics as a career, with the aim to increase the pool of URM applicants applying to pediatric residencies nationwide.²⁰ 

The expectation that minority trainees and faculty take on additional responsibilities by participating in diversity initiatives is known as the minority tax.^25–29  For trainees, this expectation may compete with training or disrupt rest and self-study, and for faculty, it may hinder professional advancement.²⁴  Throughout our project, as more interventions were added, additional duties accrued to URM residents. Successfully increasing the number of URM residents in our program helped mitigate the minority tax, but this is not a desirable final solution to this burden. The most desirable outcome is robust institutional, faculty, and graduate medical education commitment to equality and equity such that URM trainees may seek, learn, and flourish in the same full array of pursuits available to all trainees. In this ideal scenario, the tax now incurred by URM physicians becomes more equitably borne by everyone.

One way to ease the minority tax is for senior leaders is to acknowledge diversity efforts in full-time equivalent calculations and academic promotion criteria. Promotion criteria in the Department of Pediatrics at The Ohio State University recognizes the varied and diverse contributions of faculty members, including participation in URM recruitment activities. Moreover, efforts to advance diversity, equity, and inclusion within our department routinely receive institutional recognition in the form of awards that support academic promotion. Non-URM faculty and resident team members must help alleviate the burden on URM colleagues by creating minority tax–exempt opportunities. As a simple example, if a URM colleague is asked to prepare a lecture on diversity and inclusion, their supervisor should provide coverage of clinical duties and protected time for the URM colleague to complete the added assignment. The publication of this article is a concrete example of how dedicated efforts of URM faculty and trainees can be converted into academic output supporting promotion. Doing so provides a motivating career opportunity for more faculty participation in diversity initiatives, which in-turn alleviates some of the burden differentially placed on trainees.²⁴ 

More broadly, there is a pervasive expectation that individuals from marginalized communities, such as those who are URM or lesbian, gay, bisexual, transgender, and queer; those with disabilities; and others, serve as ambassadors of their groups, frequently precipitating their absence from other platforms.^30,31  It is important for institutions to understand that while striving for diversity, inclusion, and equity, everyone can and should play a role in creating a just culture. Ultimately, the most perfect iteration of a diverse, equitable, and inclusive culture must refute the perception that only individuals from marginalized communities can be experts on topics relating to their underrepresented status and that majority colleagues are equally capable of promoting diversity initiatives as allies.³² 

There was little resistance to the adoption of this initiative within the pediatric residency program. There were, however, challenges in integrating our approach with our combined residency programs because this involves increased buy-in and coordination from multiple stakeholders from outside the pediatric residency program. Because of the minority tax, the greatest hurdle was obtaining consistent faculty involvement. As our institution finds ways to mitigate this burden, faculty engagement continues to increase.

Our residency program was able to effectively improve its diversity, although the URM applicant pool increased only 2% over the course of our project (Table 1). However, our gains likely have come at the cost of another program’s losses. This so-called 0 sum game has been recently described as related to recruitment of URM subspecialty pediatrics fellows.³³  Consequently, our local success accomplished little to improve diversity of the national pediatric general and subspeciality workforce and occurs against the backdrop of 4 decades of stagnate matriculation of URM medical students in the United States.³⁴  In 2020, the self-reported URM composition of US medical school graduates included only 6.59% Black or African American students; 5.88% Hispanic, Latino, or Spanish students; and 0.04% Native Hawaiian or another Pacific Islander students (far short of equivalence with the general population). Collectively, these data clearly indicate a URM pipeline problem into medical school and consequently into pediatric residency training.

Academic medical centers have an obligation to develop sustainable pipeline initiatives at the precollegiate, undergraduate, and medical school levels at the same time they commit to recruit and retain URM faculty. Only then will they fully meet their patient care, research, and teaching responsibilities to a diverse public. A plethora of these pipeline initiatives, many of which are creative, exist across the United States. These have not led to increases in enrollment of URM students into medical school. A comprehensive evaluation and analysis of these programs is needed to identify innovative new strategies to enhance the entry of URM students into health career. Consideration of these is beyond the scope of this commentary.

Child health leaders must better understand what factors contribute to URM medical students choosing the field of pediatrics. Characteristics of specialty choice among medical students include parental profession, status, altruism, and patient characteristics.^35,36  Other factors, such as career guidance during medical school from mentors, access to high-quality clinical experiences, and early and accurate information regarding training and career opportunities provide better understanding of the profession and the populations served. Interestingly, intention to work with patients who are medically underserved predicts interest in primary care disciplines, such as pediatrics.³⁷ 

Cycle 4 was an unprecedented time for training programs, faculty, residents, and applicants because of the coronavirus disease 2019 pandemic and concurrent movements for social justice in the United States and around the world. Each of these experiences disproportionately impact URM residents, faculty, and applicants nationally and locally. The disproportionate effect of severe acute respiratory syndrome coronavirus 2 on the health of minoritized populations likely contributed additional stress for URM faculty and trainees.³²  Medical students had fewer opportunities for away rotations. We did not accommodate second-look visits. Collectively, these factors had the potential to compromise our high-touch, personal approach to URM recruitment. Moreover, like all programs, virtual interviews resulted in far more applications than usual, potentially compromising the time-consuming process of holistic review.³⁸  Despite these potentially adverse influences, 21% of our matched residents for the 2020 recruitment season were URM. Some experts have theorized that virtual interviews contribute to a more equitable interview process, a concept that merits additional study.³⁹ 

A physician workforce demographic that reflects the patient population served is associated with improved outcomes and promotes health equity.^1,2  However, as the US population grows more diverse, physician-patient demographic parity has worsened.¹⁵  Using iterative tests of change and a quasi-experimental time series design to implement interventions, we were able to significantly improve the diversity of our resident workforce, which led to a shift in how our institution approaches inclusive excellence and strives for equitable outcomes. Our focus was strong leader support; increased visibility and engagement of the URM workforce, primarily at the resident level; and deployment of a modest departmental and institutional financial investment.^11,13  Although we have not yet met our primary aim, these encouraging early results reveal statistically significant and sustained improvements after 4 years of recruitment. Implementing these efforts comes at the cost of a minority tax, which requires attention, acknowledgment, and mitigation.

We acknowledge Kathy Trace, Nationwide Children’s Hospital Pediatric Residency Program coordinator, for her commitment to workforce diversity and her adaptability to new recruitment strategies.

LMSA

Latino Medical Student Association

MHO

minority housestaff organization

QI

quality improvement

SNMA

Student National Medical Association

URM

underrepresented in medicine

VSE

visiting student elective