Page, Kathleen Raquel MD; Castillo-Page, Laura PhD; Wright, Scott M. MD
The Institute of Medicine strongly endorses diversity in the medical profession as a strategy to improve the overall health of the public.1 However, despite some progress in diversifying the medical student body over the last several decades, the proportion of African American and Latino medical students still lags behind corresponding U.S. demographic changes. Among the faculty at academic health centers, the proportions of racial and ethnic minorities are even lower. For example, although African Americans and Latinos represent 28% of the U.S. population, they account for only 15.2% of all medical students and 7.2% of all medical faculty.2–4
The reasons for the lack of diversity among faculty in academic medical institutions are multiple and complex. Pipeline issues and the high cost of medical education are obvious challenges, with relatively low numbers of racial and ethnic minority students entering medical schools. Approaches to stimulate interest in academic medicine include increasing exposure of minority students to role models and to structured mentored experiences early in their educational journey, as well as increased exposure to biomedical sciences in general.5
However, the scarcity of racial and ethnic minority trainees is not the only obstacle to diversifying academic medical institutions' faculties. Multiple studies have shown that racial and ethnic minority health care professionals who have chosen careers in academic medicine encounter barriers that negatively influence their likelihood for success.6–8 The underrepresentation of a diverse faculty cohort can lead to feelings of isolation for racial and ethnic minority faculty within the academic community. Further, Cohen6 described a disproportionate demand on this group of faculty to meet obligations that do not traditionally translate into products recognized by promotion processes, such as serving on committees, mentoring racial and ethnic minority students, or volunteering in the community. Racial and ethnic minority faculty are also more likely to experience conscious and subconscious bias, report lower career satisfaction, and transition out of academia sooner than other faculty.9,10 Even more troubling, racial and ethnic minority faculty at the assistant and associate professor levels are less likely to be promoted than are Caucasian faculty, even after accounting for confounding factors such as gender, degree, tenure status, grant support, and research productivity.7,11
In an attempt to diversify the faculty in medicine, many institutions have established diversity programs that address pipeline challenges and help improve racial and ethnic minority faculty recruitment, retention, and academic success. Descriptive reports from selected faculty development initiatives and input from racial and ethnic minority faculty have identified several key components to improving faculty diversity, such as setting institutional goals and targets, holding leaders accountable for progress, providing specific mentoring programs, improving networking and leadership opportunities, limiting obligations that do not contribute to promotion, and supporting protected research time.4,12–14 However, diversity programs vary in scope, funding, and institutional support. Also, racial and ethnic minority faculty report a gap between institutional intention and implementation of practical initiatives aiming to promote a diverse academic environment.8
To learn more about these programs, we surveyed diversity program leaders at U.S. MD-granting medical schools to obtain information about the composition and support of diversity programs for faculty. We hypothesized that programs with more funding and strong institutional support (defined as receiving funds from the institution) would have higher racial and ethnic minority faculty representation than the national average reported by the Association of American Medical Colleges (AAMC). We also hypothesized that (1) greater diversity in the demographics of an institution's state and (2) the presence of a diverse medical student body 10 years earlier would positively influence faculty diversity.
We targeted leaders of diversity initiatives in all U.S. medical schools (except those in Puerto Rico) for inclusion in the study. We reviewed the AAMC's Directory of Diversity Affairs Officers (http://services.aamc.org/minorityaffairsdirectory/) to obtain contact information for medical schools' diversity leaders (defined as deans, associate or assistant deans, and directors of diversity, minority, or multicultural affairs programs). In some cases, we contacted medical schools by e-mail to request contact information for the institution's diversity leader. In addition, the following terms were used to search the Internet content of all accredited MD-granting schools of medicine in the United States: diversity, minority, minorities, and multicultural affairs. Links of interest were explored to obtain contacts for diversity leaders at each institution. We obtained at least one contact for 106 of the 127 U.S. MD-granting institutions that existed during the study period. Of the remaining 21 schools, some did not seem to have any diversity programs that could be identified, other programs did not seem to have a diversity leader, and a few refused to share any contact information, which meant that we could not include their programs in the study. The majority of the institutions where we could not identify a diversity program or leader were public (16/21; 76.2%). Minority faculty representation in these institutions was similar to that in institutions that participated in our survey: 2.9% versus 2.6% African American and 3.2% versus 3.1% Latino faculty, respectively. Racial and ethnic minority representation in the states of institutions where we could not identify a diversity leader was slightly higher than such representation in the states of institutions that participated in our survey: 14.5% versus 13.4% African American, and 12.1% versus 11.4% Latino, respectively. We conducted the survey in 2010.
From five institutions, more than one individual responded. In those cases, we included the individual with the highest-ranking position in his or her diversity program, and we excluded individuals who worked exclusively on student diversity. In addition, 11 leaders from distinct institutions did not answer many of the questions and were thus excluded from the analysis because of incomplete data.
The study was considered exempt by the Johns Hopkins institutional review board.
We developed a 26-item questionnaire focusing on the characteristics, leadership, and scope of diversity programs in schools of medicine. Program characteristics included date of establishment, target populations, funding sources, and institutional support. Characteristics of the leaders included gender, self-identified race and ethnicity, title, years of experience, time devoted to diversity initiatives, and salary support for diversity efforts. We sought information about the scope of diversity programs by asking about the types of diversity initiatives supported, the definition of underrepresented minority (URM), and the eligibility of racial and ethnic minority groups for participation in diversity initiatives. The majority of the questions were multiple-choice or yes/no/don't know answers, with the opportunity to add text for “other” options. The items selected for inclusion in the questionnaire were chosen based on a review of the literature and consultations with experts in the field.
After pilot testing the questionnaire at the Johns Hopkins School of Medicine Diversity Council and with selected diversity leaders from the AAMC, the instrument was sent electronically to diversity leaders in February 2010. Monthly reminder e-mails and phone calls one month prior to closing the study were used to encourage full participation.
To evaluate the association between diversity program characteristics and faculty diversity, we obtained the 1998 and 2008 data from the AAMC Faculty Roster System, a data system that tracks U.S. medical school faculty. (The 2008 data were the most current at the time of the study.) We also obtained the AAMC's 1998 list of medical students by medical school, which includes race and ethnicity data. We used 1998 medical student data because we estimated that it would take approximately 10 years for these students to complete training and assume faculty positions. For each medical school that responded to our survey, we compared the proportions of African American and Latino faculty in 1998 and 2008, and the proportions of African American and Latino medical students in 1998. To evaluate the impact of state demographics on medical school faculty diversity, we added to the database the U.S. Census Bureau 2008 estimates of the African American and Latino population for each state of the participating schools of medicine.15
Survey responses were reviewed to examine the frequency of distributions and identify outliers, nonnormality, and other data irregularities. Descriptive statistics (mean, median, range, and standard deviation [SD]) were used to summarize the responses to all the questions from the institutions that we included in the descriptive analysis. Linear regression and chi-square analysis were used to evaluate differences in the diversity programs, leadership, and funding characteristics between schools of medicine with high versus low faculty diversity. In this analysis, we excluded any institutions that were outliers. High diversity was defined as higher than the mean racial and ethnic minority faculty representation reported by the AAMC (>3% for African American faculty and >4.2% for Latino faculty).16 Other outcomes explored included change in faculty diversity during a 10-year period (ratio of 2008 to 1998 minority faculty representation), overall minority representation (high overall minority faculty representation defined as >7.2% based on AAMC data),16 and minority faculty representation relative to state demographics (ratio of faculty minority percentage to state minority percentage). Pearson correlation was used to assess the relationship between state racial/ethnic demographics and minority faculty representation. All analyses were done using STATA version 10.0 (College Station, Texas).
Characteristics of diversity leaders
Of the 106 individuals who were sent our survey via e-mail, 89 (84.0%) completed the survey; 82 (77.4%) were identified as the main diversity leaders at their institutions, and their responses were included in the study. The majority of the diversity leaders were deans (3/82; 3.7%), associate deans (38/82; 46.3%), or assistant deans (15/82; 18.3%) (see Table 1). Other titles included directors of minority, multicultural, or diversity programs (12/82; 14.6%) and other titles such as vice president, chief diversity officer, assistant or associate vice chancellor, chair, or assistant vice provost (14/82; 17.1%). The majority of the study respondents were women (52/82; 63.4%) and from underrepresented racial and ethnic backgrounds (African Americans: 54/82 [65.9%]; Latinos: 12/82 [14.7%]). On average, diversity leaders had served in their current position 6.7 years and dedicated approximately half of their FTE to diversity initiatives. More than half (47/82; 57.3%) of the diversity leaders received full (100%) salary support for the FTE dedicated to diversity.
Characteristics of diversity programs
The average age of diversity programs was 18.7 years (SD 13.1) with a median age of 17 years (see Table 2). Forty percent (33/82) of the programs were relatively new (<10 years), but one-third (27/82) had been established before 1985 (≥25 years). The majority of the programs (52/82; 63.4%) targeted both trainees (students, residents, and fellows) and faculty, whereas 30.5% (25/82) primarily targeted only trainees, and 6.1% (5/82) targeted faculty alone. Approximately one-fourth of the programs had a budget of less than $100,000 (22/82; 26.9%), a third between $100,000 and $500,000 (30/82; 36.6%), and a fourth over $500,000 (20/82; 24.4%). Only four programs had no allocated funds, whereas three programs had an annual budget in excess of $1,000,000. The majority of the programs (71/82; 86.6%) received institutional financial support from the dean's, provost's, or chancellor's office. Approximately one-fourth (22/82; 26.8%) received funding from extramural grants, and 15.9% (13/82) from endowments or charitable contributions. On average, diversity offices had 4.2 salaried staff (SD 3.5, median 3), including the diversity leader.
The top three initiatives supported by diversity programs were outreach to racial and ethnic minorities in the community (58/82; 70.7%), recruitment of racial and ethnic minority faculty (58/82; 70.7%), and retention efforts (53/82; 64.3%). More than half of the programs had specific activities to improve the institutional environment for racial and ethnic minorities, such as cultural competency training, leadership education to improve awareness about conscious and unconscious bias, and formal networking opportunities for racial and ethnic minority faculty. Specific professional development programs for racial and ethnic minority faculty such as mentorship and leadership training were common (42/82 [51.2%] and 35/82 [42.7%], respectively). One-third of the programs (28/82; 34.1%) had an established mechanism to track racial and ethnic minority faculty representation. Only 11 out of 82 (13.2%) of the programs restricted their activities to only African Americans, Mexican Americans, Native Americans, or mainland Puerto Ricans (the previous AAMC definition of URM until 2003).17
Predictors of faculty diversity
To analyze factors associated with diversity, we evaluated the data from 79 institutions, but we excluded data from the three historically black colleges. Higher racial and ethnic minority faculty representation was associated with being in a state with high African American and/or Latino populations (see Table 3). Among schools with >3% African American faculty, the average state African American population was 22.5% (versus 9.9% in states of schools with <3% African American faculty). Schools with >4.2% Latino faculty had an average state Latino population of 21.1% (versus 8.9% in states of schools with <4.2% Latino faculty). Correlations were also noted between African American or Latino state population and African American or Latino faculty representation in the responding medical schools (African American ρ2 = 0.46, Latino ρ2 = 0.34, both P < .001) (see Figure 1).
None of the diversity program characteristics measured, such as funding, institutional support for diversity, race/ethnicity of the program leader, size of program, type of diversity initiative, or whether the institution was public or private, were associated with higher racial and ethnic minority faculty representation.
High racial and ethnic medical student representation in 1998 was associated with higher faculty diversity in 2008. Schools with higher-than-average African American faculty representation had a larger proportion of African American students in 1998 (8.0% versus 5.7%, P = .03). Likewise, in 1998 there were more Latino medical students in schools with >4.2% faculty representation in 2008 (6.8% versus 3.9%, P = .004). The average proportion of African American faculty was higher in schools with more than 4.2% Latino faculty (3.3% African American faculty versus 2.4% African American faculty, P = .03), but there was no difference in Latino faculty representation between schools with higher- or lower-than-average African American faculty (3.5% Latino faculty versus 3.0% Latino faculty, P = .17).
We also evaluated trends in diversity by calculating the changes in African American and Latino faculty representation in the 79 institutions in our sample (we excluded the three historically black colleges from the calculation). Because of the relatively small number of racial and ethnic minority faculty at each institution, these data were somewhat labile. For example, a large proportional increase or decrease in racial and ethnic minority faculty could be due to the recruitment or loss of only one individual. On average, there was a 61% (SD = 49%) increase in African American and a 46% (SD = 50%) increase in Latino faculty representation during that decade. However, approximately one-fourth of the 79 institutions had a decline in African American and Latino faculty representation during that period (22% and 25%, respectively). In bivariate analysis, we did not identify any factors associated with the temporal changes in faculty representation (data not shown).
Because of the strong associations between state demographics and faculty representation, we evaluated the relationship between program characteristics and faculty diversity after controlling for local demographics. The mean ratio of African American faculty and African American populations in states was 25% (interquartile range 13%–28%), and the mean ratio of Latino faculty and Latino populations in states was 48% (interquartile range 18%–62%). None of the measured diversity program characteristics was associated with higher racial and ethnic minority faculty representation relative to state demographics (data not shown).
Discussion and Conclusion
To our knowledge, this is the first study to report a comprehensive description of programs aimed at improving faculty diversity in U.S. schools of medicine. We found that the majority of diversity programs receive monetary support from their institution, many are long established (>10 years old), and most target racial and ethnic minorities across the institutional spectrum, from students to residents to faculty. None of the program characteristics that we measured were associated with either racial and ethnic minority faculty representation in 2008 or a change in minority faculty representation between 1998 and 2008. Local state demographics and high medical student diversity 10 years prior were associated with African American and Latino faculty representation. The finding that faculty diversity is influenced more by the environment and the minority medical student pipeline than by diversity programs' efforts directed at faculty suggests that serious efforts to enhance faculty diversity require early intervention and a long-term commitment. More early intervention programs that target high school and middle school students should be implemented, and all such programs should be evaluated.
In our study, diversity in medical schools correlated strongly with local demographics. The finding that schools in states with higher proportions of racial and ethnic minorities in the population have more racial and ethnic minority faculty has not been previously documented. This result, however, is congruous with the AAMC's schema for diversity wherein there is an emphasis on evaluating diversity within the context of the local populations.16–18 Our data highlight the challenges of diversifying the academic faculty in states with low racial and ethnic minority populations. Schools in states with low minority representation may need to proactively recruit minority faculty from other states and create an environment where diverse faculty can thrive in order to overcome their demographic disadvantage. It is worth remembering, however, that even among schools with higher racial and ethnic minority faculty representation, there is a vast difference between minority faculty and local minority representation. Therefore, relying solely on faculty recruitment from high-minority states or on the growth of minority populations is unlikely to lead to greater representation of African American or Latino faculty in U.S. medical schools.
This study's results also suggest that improving racial and ethnic minority student enrollment in medical schools may be an effective strategy to improve faculty diversity. Schools with higher racial and ethnic minority medical student representation in 1998 had better racial and ethnic faculty representation 10 years later, perhaps because some of these students went on to join the faculty at the same institution. The association between racial and ethnic minority medical student and faculty diversity may also be driven by state demographics. According to the AAMC's 2008 Matriculating Student Questionnaire, 48.5% of Latinos and 43.4% of African Americans that matriculated into medical school emphasized the importance of the geographical location of the school in deciding which medical school to attend.19 Although the reason for this association may be subject to speculation, the findings have implications for racial and ethnic minority faculty. In the short term, schools with low racial and ethnic minority student and demographic representation may need to develop programs to attract faculty from elsewhere. In the longer term, providing incentives for racial and ethnic minority medical students to attend out-of-state institutions may help diversify the student body and provide the seeds for the next generation of faculty.
However, improving racial and ethnic minority student enrollment remains a challenge for all institutions, and efforts to improve the pipeline and promote an inclusive academic environment need to be redoubled and sustained. The proportion of racial and ethnic minority medical students has remained constant during the past decade despite the growth of minority populations nationwide. In 1998, African Americans and Latinos accounted for 14.2% of all medical students, whereas in 2008 they composed 15.1%.3,20 Successful programs that have improved matriculation of racial and ethnic minorities in medical school often intervene early in the educational trajectory. Examples of such programs include the Premedical Honors College, an eight-year, high-school-through-medical-school program created by Baylor College of Medicine and the University of Texas–Pan American, and the intensive Summer Medical and Dental Education Program that prepares undergraduate students in science, writing, and test-taking skills.5,21 The long-term impact of these programs on faculty diversity should be evaluated because our results suggest that investment in this type of program may be one of the most effective strategies in improving faculty diversity.
The lack of association between specific characteristics of diversity programs and higher-than-average African American or Latino faculty representation does not mean that these programs are ineffective. In 2006, the AAMC Diversity Research Forum18 identified three components of diversity that depend on each other to maximize the institutional environment of equity and inclusion. The key elements to a diverse climate are (1) compositional diversity (the proportion of racial and ethnic minority faculty and trainees), (2) diversity of interactions through which individuals gain exposure to alternative views, and (3) institutional diversity-related activities that address issues of diversity, such as cultural competency training. Our study outcome focused solely on compositional diversity as a measurable indicator of diversity. Although we did not find an association between the diversity program characteristics we measured and compositional diversity, it is probable that these programs are having an impact on the other components of diversity.
Descriptive reports of racial and ethnic minority faculty development programs at 12 institutions identified several ingredients for success, such as setting clear goals, careful selection of target population, strong mentoring programs for racial and ethnic minority junior faculty, providing an environment for success, and promoting leadership among senior racial and ethnic minority faculty.12,22 We quantified some of these factors, but some program characteristics such as culture, hospitality, relationships and commitments to racial and ethnic minority faculty may be best evaluated using qualitative methods. For example, we measured funding from the dean's or provost office as a surrogate for institutional support, but strong institutional commitment includes not only dollars but also the explicit support from a critical mass of individuals and leaders that can affect the climate of the institution. To this end, we found that one of the predictors of higher Latino faculty representation was a higher proportion of African American faculty, suggesting that improved representation of one racial and ethnic minority group may positively change the environment for other minority groups.
Several limitations of this study should be considered. First, our ability to detect statistically significant differences was limited by a relatively small sample size. Although our response rate was high, 44 institutions were not included in our analysis because we were unable to identify a diversity leader or the identified diversity leader did not participate in the survey. The average African American or Latino faculty representation among institutions that did not participate was slightly higher than among participating institutions (3.1% versus 2.6% African American faculty, and 4.0% versus 3.1% Latino faculty, respectively). There was no difference in the proportion of African Americans in the states of institutions that did not participate compared with those that did (13.2% versus 13.4%, respectively), but the Latino representation was slightly higher in the states of institutions that did not participate (14.3% versus 11.4%). Therefore, our sample may not represent some of the institutions that have been more successful in compositional diversity.
Second, it should be noted that international medical graduates (IMGs) account for approximately one-fifth of the U.S. physician workforce and may play a significant role in diversifying academia, but we were unable to distinguish between IMG minority faculty and U.S.-trained minority faculty.23 In addition, although our survey focused on faculty diversity initiatives, we were not able to differentiate between funding for faculty diversity versus funding for trainee diversity and did not measure the resources allocated for specific activities. Because some programs may focus more of their attention on racial and ethnic minority student recruitment, information on the funding for activities that directly affect faculty, such as retention and recruitment initiatives, may have provided valuable information about how to manage limited resources. Finally, we developed the questionnaire used to collect data because there had not been a previously validated measure. Evidence of the questionnaire's content validity was conferred by the experts and literature review. Further, the questions inquired about facts that the respondents would be expected to know, given their positions, rather than asking about attitudes and opinions.
There are many practical arguments for promoting diversity in academic medicine. Particularly compelling is the need to provide optimal and culturally sensitive care to an increasingly diverse U.S. population. Members of diverse groups are more likely to provide care to underserved populations, specifically of their own racial or ethnic background. Patients receiving care from physicians of their same cultural background seem to benefit from enhanced communication and improved outcomes.24–26 Further, a diverse academic faculty is more likely to drive the expansion of the research agenda to address the health care matters and concerns that are germane to the needs of diverse populations. An equally important argument in favor of diversity is the impact of diversity on academic excellence.16 The innovation that drives progress in research and education is rooted in diverse experiences and points of view. The contribution of diverse faculty to academia should not be judged only by its impact on health disparities but, rather, embraced as a necessary component to enhance the overall experience and mission of the academic institution.
Our findings have practical implications for institutional and diversity leaders who wish to meet the physician workforce needs of the 21st century. At the faculty level, African Americans and Latinos are highly underrepresented. The finding that faculty diversity is strongly associated with state demographics highlights the challenges faced by medical schools in low minority states and suggests that these programs may need to work that much harder to attract and retain racial and ethnic minority faculty. To meet this demand, schools across the board will need to create an inclusive environment that is attractive to diverse students and faculty. The impact of strengthening pipeline programs on medical student diversity and the educational experience is well documented.27,28 Our findings demonstrate that enhancing student diversity, a modifiable factor that is within the control of medical schools and other academic medical institutions, seems to be associated with enhanced faculty diversity. This is yet another reason to expand initiatives that recruit racial and ethnic minority students into careers in medicine.
The authors would like to thank Brian Gibbs, Myron Weisfeldt, Roy C. Ziegelstein, Robert Bollinger, Norma Poll-Hunter, and Marc Halushka for their insightful suggestions, and Lisa Roberts for her excellent administrative support.
Dr. Wright is a Miller-Coulson Family Scholar, and this work is supported by the Miller-Coulson family through the Johns Hopkins Center for Innovative Medicine. Dr. Page is supported by a K23 training grant (K23HD056957) from the National Institute of AIDS and Infectious Diseases and the National Institutes of Health.
This study was reviewed by the Johns Hopkins IRB committee (NA_00033321), which determined that the study qualified for an exemption under the DHHS regulations.
1 Smedley BD, Stith Butler A, Bristow LR. In the Nation's Compelling Interest: Ensuring Diversity in the Health Care Workforce. Washington, DC: National Academies Press; 1994.
3 Association of American Medical Colleges. AAMC Data Book: Statistical Information Related to Medical Education. Washington, DC: Association of American Medical Colleges; 2010.
4 Nivet M, Cora-Bramble D, Butts G. Striving Toward Excellence: Faculty Diversity in Medical Education. Washington, DC: Association of American Medical Colleges; 2009.
5 Cantor JC, Bergeisen L, Baker LC. Effect of an intensive educational program for minority college students and recent graduates on the probability of acceptance to medical school. JAMA. 1998;280:772–776.
6 Cohen JJ. Time to shatter the glass ceiling for minority faculty. JAMA. 1998;280:821–822.
7 Fang D, Moy E, Colburn L, Hurley J. Racial and ethnic disparities in faculty promotion in academic medicine. JAMA. 2000;284:1085–1092.
9 Peterson NB, Friedman RH, Ash AS, Franco S, Carr PL. Faculty self-reported experience with racial and ethnic discrimination in academic medicine. J Gen Intern Med. 2004;19:259–265.
10 Price EG, Gozu A, Kern DE, et al. The role of cultural diversity climate in recruitment, promotion, and retention of faculty in academic medicine. J Gen Intern Med. 2005;20:565–571.
11 Palepu A, Carr PL, Friedman RH, Amos H, Ash AS, Moskowitz MA. Minority faculty and academic rank in medicine. JAMA. 1998;280:767–771.
12 Daley SP, Palermo AG, Nivet M, et al. Diversity in academic medicine no. 6 successful programs in minority faculty development: Ingredients of success. Mt Sinai J Med. 2008;75:533–551.
13 Merchant JL, Omary MB. Underrepresentation of underrepresented minorities in academic medicine: The need to enhance the pipeline and the pipe. Gastroenterology. 2010;138:19–26.e1–e3.
14 Wesson DE, King TE Jr, Todd RF, et al. Achieving diversity in academic internal medicine: Recommendations for leaders. Am J Med. 2006;119:76–81.
16 Association of American Medical Colleges. Striving Toward Excellence: Faculty Diversity in Medical Education. Washington, DC: AAMC; 2009.
18 Association of American Medical Colleges. The Diversity Research Forum: Tools for Assessing and Measuring Diversity in Medical Education. Washington, DC: AAMC; 2005.
19 Association of American Medical Colleges. Matriculating Student Questionnaire. Washington, DC: Association of American Medical Colleges; 2008.
20 Association of American Medical Colleges. AAMC Data Book: Statistical Information Related to Medical Education. Washington, DC: Association of American Medical Colleges; 1998.
22 Nivet MA, Taylor VS, Butts GC, et al. Diversity in academic medicine no. 1 case for minority faculty development today. Mt Sinai J Med. 2008;75:491–498.
24 Cooper-Patrick L, Gallo JJ, Gonzales JJ, et al. Race, gender, and partnership in the patient–physician relationship. JAMA. 1999;282:583–589.
25 Komaromy M, Grumbach K, Drake M, et al. The role of black and Hispanic physicians in providing health care for underserved populations. N Engl J Med. 1996;334:1305–1310.
26 Rabinowitz HK, Diamond JJ, Veloski JJ, Gayle JA. The impact of multiple predictors on generalist physicians' care of underserved populations. Am J Public Health. 2000;90:1225–1228.
27 Saha S, Guiton G, Wimmers PF, Wilkerson L. Student body racial and ethnic composition and diversity-related outcomes in US medical schools. JAMA. 2008;300:1135–1145.