Thomson, William A. PhD; Ferry, Pamela MHS; King, Jason PhD; Wedig, Cindy Martinez PhD; Villarreal, Graciela B. MD
There has been long and widespread agreement that increasing diversity in medicine and across the entire health care workforce is in the best interest of the nation's health.1–3 The Association of American Medical Colleges (AAMC) and the American Medical Association have both made diversity a priority goal.4–5 Indeed, evidence indicates that a racially and ethnically diverse student body contributes to preparing medical students for practice among an increasingly diverse population.6 For over 30 years, major programmatic efforts, many known as “pipeline programs,” have been undertaken to address recognized needs.7–9 They include the AAMC's Project 3000 by 2000, federal programs such the Health Careers Opportunity Program and the Centers of Excellence initiative, the Robert Wood Johnson Foundation's Summer Medical and Dental Education Program, and the Health Professions Partnership Initiative (funded by the Robert Wood Johnson and WK Kellogg Foundations).
Although these laudable programs seem to have a role in helping minority students succeed in the health professions pathway, they have met with limited success in increasing the diversity of students entering the health professions.2,3,10 In medical education, underrepresented minority (URM) students reached a high of 15.5% of total first-year matriculants in 1994 before falling to 13.8% in 2000.1 Recent data are not encouraging. The proportion of first-year URM matriculants who identified themselves as either American Indian/Alaskan Native, black/African American, or Latino between 2002 and 2008 remained relatively flat with respect to total first-year enrollment in U.S. medical schools. For example, in 2002, 15.6% of entering first-year medical students were from these three groups; in 2008 they accounted for 16.0% of entering students.11
For many observers these data are not surprising in that obstacles to implementing diversity-oriented programs have been considerable. Court rulings and political initiatives of the 1990s banning the use of race in admission decisions have negatively influenced diversity efforts and forced many medical schools to turn to “race-neutral” strategies that focus on socioeconomic and geographic factors rather than race to recruit a diverse student body.3,12 The policy-level barriers, combined with gross inequalities in K–16 educational opportunities across racial and ethnic groups, have made achievement of diversity goals in medical education a challenge.2 Unfortunately, the lack of education research has made it difficult to conclusively determine the effectiveness of diversity-promoting programs and to identify the specific intervention strategies that may most effectively—including cost-effectively—promote participation of URMs in the health professions.3,7,8,12
As of July 1, 2009, recently adopted Liaison Committee on Medical Education (LCME) standards will require all medical schools to develop programs or partnerships aimed at broadening diversity among qualified applicants for medical school admission. In the new annotation for this standard, the LCME specifically encourages pipeline programs and partnerships with other institutions to make admission to medical education more accessible to potential applicants from diverse backgrounds.13 Even before the new LCME requirement, many medical schools had developed partnerships with school districts and communities to strengthen the educational pipeline and increase the number of qualified medical school applicants. These efforts included the Health Professions Partnership Initiative,14 summer enrichment activities,15–19 and baccalaureate/MD programs.20–22
The lack of progress toward increasing the numbers of medical school matriculants from traditionally underrepresented backgrounds has been discouraging. Therefore, it is surprising that there is also a paucity of education research that documents the long-term effectiveness of existing pipeline strategies or that recommends one approach over another. Unfortunately, medical schools must continue to rely on limited published examples of success and anecdotal evidence that current diversity strategies have a positive effect on the challenging problem.
Baylor College of Medicine (BCM) has a long history of educational outreach initiatives that have been innovative and effective.20,23,24 One of BCM's most successful pipeline program strategies—the Premedical Honors College (PHC)—targets students from a medically underserved and predominantly Latino region of the state, including the Texas–Mexico border. A general description of the PHC and initial results have been published previously.23,25 In this article, we present data on the program's first 15 years of operation (1994–2008), focusing on program retention and outcomes related to participant career pathways, in the hopes that the experiences of BCM and the PHC will add to the growing body of evidence on successful strategies to increase diversity in medical education.
A Pipeline Program Experiment for South Texas Students
In 1983, BCM began to partner with educational institutions and school districts in South Texas to address the shortage of physicians and students academically prepared to enter the medical profession in the area. BCM has defined the South Texas target region as a 13-county area (consisting of Brooks, Cameron, Duval, Hidalgo, Jim Hogg, Jim Wells, Kenedy, Kleberg, Nueces, Starr, Webb, Willacy, and Zapata counties) representing a population of approximately 1.8 million, which is over 83% Latino according to 2006 census estimates.26 All 13 counties have all or part of their populations designated as medically underserved, and 12 of the 13 also contain one or more primary care Health Professional Shortage Areas (as defined by the U.S. Department of Health and Human Services).27
Physician shortages in South Texas are due in part to the fact that too few students from South Texas colleges and universities enter medical school. Before the inception of the PHC, there were five undergraduate colleges in the region with an annual cumulative enrollment of approximately 30,000 students (80% Mexican American). In 1994, 26 students from these institutions applied to medical school, and only 10 matriculated, including 6 students from The University of Texas–Pan American (UTPA) (Franc J. Slapar, MA, database specialist, Data Resources and Studies, AAMC, December 31, 2009). However, since graduating its first undergraduate cohort in 1998, the PHC alone has produced a total of 134 medical school matriculants (averaging 12 per year).
The PHC was established in 1994 as a partnership between BCM and UTPA to create an eight-year, high-school-through-medical-school pathway for promising students from South Texas counties. The program provides academic support, a rigorous undergraduate curriculum, enrichment experiences, scholarships, and conditional acceptance to BCM on successful completion of the BS degree in biology or chemistry. Specific details of the curriculum and program offerings are provided elsewhere.23,25 The main purpose of the program is to produce more physicians for needy areas of the state of Texas. However, additional aims of the program are to increase access to medical education for students from underrepresented and disadvantaged backgrounds (taking into account race, ethnicity, and socioeconomic and geographic factors) and to enhance the diversity, and thus the cultural awareness and competency, of the BCM medical school body and ultimately the physician population of the state. In other words, the PHC could be considered both a race-neutral and a “pipeline” approach to the diversity issue.
Now in its 15th year of operation, the PHC is beginning to demonstrate its long-term impact as some of its first graduates complete their training and enter medical practice. Between 1994 and fall 2008, a total of 242 students entered the PHC. As of 2008, 134 (87% of those who completed the undergraduate portion of the program) have matriculated to medical schools in Texas, 99 at BCM. Among those matriculating to medical schools, 110 (82%) are URMs and 106 (79%) are Latino. Table 1 depicts the overall characteristics and outcomes of PHC participants from 1994 to 2008.
Recruitment and Retention Strategies and Challenges
One of the first issues program planners confronted was the need to identify and recruit academically strong students from the region to become participants in an experimental program. Initially, recruitment strategies focused on students attending BCM-affiliated magnet health science high schools in the region, as well as using direct media advertisements, special presentations to school boards and community groups in South Texas, and direct and repeated contacts with high school counselors by UTPA recruiters. The success of these efforts is evidenced by the increasing average SAT scores of entering cohorts from 1994 to the present (see Table 1). During the first 12 years of the program, incentives of full tuition and fee support at both the undergraduate and medical school levels helped achieve the desired number of entering first-year students. In recent years, financial support has been provided based on demonstrated economic need. High school seniors apply for admission to the PHC, and applications are initially screened and ranked by UTPA and BCM faculty and staff on the basis of high school grades, courses taken, SAT scores, completed application, and letters of reference. Between 20 and 25 finalists are selected each year and flown to Houston for interviews with the BCM Medical School Admissions Committee. Applicants are evaluated using the same criteria applied to traditional BCM medical school applicants.
Retention of students in the pipeline
Another indicator of success of pipeline programs is retention of students in the pipeline. A recent study28 of factors associated with pipeline “dropout” among minority premedical students suggests that at the undergraduate level, one of the most common reasons for early dropout is academic challenge, particularly courses such as chemistry (most often mentioned), which are notoriously difficult early premed courses. Peer support, high-quality advisors, and supportive faculty are factors that encourage retention. Interventions are needed and must be sustained during the undergraduate prehealth curricula (gateway courses) to enhance minority student academic achievement.29 At the medical school level, a study of factors related to success among minority medical students found that social and community support, financial assistance, advisors, mentors, and faculty were considered important contributors to success, while lack of social support, discrimination, self-doubt, and financial pressures were obstacles to success.30 The PHC experience lends evidence to confirm the conclusions of this prior research.
Attrition in the PHC has been highest during the undergraduate portion of the program, mainly as a result of academic difficulty, with students voluntarily withdrawing or being dismissed after a number of academic warnings for less-than-expected performance. While student achievement is monitored throughout any semester, students who fail to meet grade point and course load requirements receive a written warning from BCM's dean of medical education. Students who receive three warnings, unless extraordinary circumstances are present, are dismissed from the program. Of the 34 students who have left the program during the undergraduate portion, at least 21 (62%) were dismissed for academic reasons, and the other 13 (38%) withdrew for nonacademic reasons (moving, marriage, change of career interests), although undoubtedly for some of these students the decision to withdraw from the program was influenced by academic performance. The PHC undergraduate students who leave the program are not replaced by alternate students.
The rate of attrition was highest in the initial years of the program. Retention rates for entering students from 1994 to 1996 who completed the undergraduate program and successfully matriculated into a medical school were 50%, 49%, and 50%, respectively. Undergraduate retention rates improved as the program matured and gained experience in the recruitment, selection, and support of students, and the number of withdrawals/dismissals has decreased. Factors that have helped to increase retention in the program include having a full-time faculty advisor whose primary responsibility is to oversee the program at UTPA and provide support for students, including tutoring for tough science classes; the close-knit support group formed among program participants over their four years at UTPA; and the network of mentoring and advising provided by upper-level program participants, including those in medical school and residency, to those coming along the pipeline behind them.
More recently, in addition to a required six-week summer program at BCM between the freshman and sophomore years, a 100-hour-per-year community service requirement has been added to the program. This requirement will be primarily addressed through clinical experiences in hospitals and clinics in South Texas. It is hoped that these experiences will promote a greater understanding of health and health care delivery in the region and continue to nurture individual motivation to become a physician.
Academic performance indicators
A more challenging obstacle to program retention has been ensuring that students meet the minimum GPA, science GPA, and MCAT requirements for admission to BCM for medical school. These requirements have evolved over the course of the program. Currently, students must maintain a semester and cumulative GPA of 3.4 (on a 4.0 scale) and science GPA of 3.2 and attain a total MCAT score of 28, with no section score less than 8. When the program began in 1994, the overall and science GPA requirements were 3.2 and 3.0, respectively, and the required MCAT score was 24. Unfortunately, there are no national guidelines to inform these requirements, and minimum performance standards have been developed and modified for the PHC based on program experience and on national, state, and institutional data for medical school matriculants.
Average cumulative undergraduate GPA, science GPA, and MCAT scores for students who entered the PHC from 1994 to 2004, and matriculated into a Texas medical school completed the BS degree at UTPA through 2008 are shown in Table 2. Overall, these data suggest that the majority of PHC participants are able to maintain a competitive GPA and science GPA while majoring in biology or chemistry (with a minor in the other discipline). Despite the provision of a full-length MCAT preparatory course for PHC participants, the MCAT remains a significant limiting factor in student retention in the program. PHC participants who fail to meet the required MCAT score are provided additional tutoring and take the test again. They are also encouraged to apply to all other Texas medical schools and have been very successful in gaining acceptance. The average MCAT score for all PHC graduates matriculating into BCM to date is 27.9 (N = 99; SD = 2.78), compared with an average score of 21.9 (N = 34; SD = 2.45) for PHC graduates who matriculated to other Texas medical schools (after failing to meet the BCM minimum MCAT requirement). As of this writing (2008), 31% of PHC students have not met the minimum MCAT score for admission to BCM. Of these, 59% matriculated into other Texas medical schools. Others have pursued degrees in public health, physician assisting, optometry, graduate science, and teaching.
Interestingly, as seen in Table 2, the mean MCAT score for PHC participants jumped from 25.1 in 2001 to 29.5 in 2002. This is a remarkable increase that cannot be attributed to any change in program curriculum or activities. However, this represents the first year that the minimum MCAT requirement was increased to 27, perhaps motivating participants to spend more time and effort on MCAT preparation. It is also possible that as the program has matured, more academically strong students and their families have been attracted to a pipeline opportunity with demonstrated success.
Retention through medical school has been much higher than that for the undergraduate portion of the program. Of the 134 PHC graduates who have matriculated into medical school in Texas (at BCM and other Texas medical schools), 65 have graduated, 15 have withdrawn, and 54 are still enrolled. Nine of the 99 (9%) PHC students who matriculated to BCM have withdrawn from medical school, 4 for personal reasons and 5 for a mixture of personal and academic reasons. Information regarding the reasons for withdrawal of the six students at other Texas medical schools is not available, but likely represents both academic and personal factors as well (one student decided to complete a PA degree rather than the MD). Performance on the various parts of the United States Medical Licensing Exam (USMLE) does not seem to be a major obstacle for those PHC students matriculating to BCM. To date, 51 PHC matriculants to BCM have taken the USMLE Step 1, and all but one student received passing scores (average score 214, range 171–248). USMLE scores for PHC graduates enrolled in other Texas medical schools were not available.
Standardized examinations play an important role in student access, progression, and retention throughout the medical education pathway. Previous findings provide evidence of relationships between standardized test scores (i.e., scores on the SAT, MCAT, and USMLE Step 1 examination) and undergraduate GPA.31–37 Our preliminary analysis of PHC data supports these findings, demonstrating a relationship between MCAT total score and each of the following: SAT (r2 = 0.274; P < .001), GPA (r2 = 0.230; P < .001), science GPA (r2 = 0.243; P < .001), and high school rank (r2 = 0.067; P < .011). In addition, in a regression of USMLE Step 1 scores on a number of variables (i.e., MCAT total and subscales, SAT, GPA, science GPA, and other academic indicators), the biology subscale of the MCAT (MCAT-B) emerged as the strongest single predictor of USMLE (r2 = 0.207; P < .001), exceeding by over 3% the predictive power of MCAT total (r2 = 0.172; P < .001). Adding other variables to the regression equation did not statistically significantly improve the predictive power of the MCAT-B. Further investigations into these and other relationships involving nonacademic predictors of USMLE are currently under way and will be reported in detail at a later date.
Although others have undertaken more comprehensive explorations into predictors of MCAT and USMLE, we see value in reporting these findings because it is unknown whether observed relationships will hold for this unique subpopulation of students. Replication studies will be undertaken as additional students complete the PHC, which should improve the accuracy of parameter estimates. In addition, longitudinal collection of data on dropouts will allow us to explore with greater specificity indicators of success/failure in medical school.
Where Graduates Choose to Practice
Although the PHC does not require its graduates to practice in South Texas or even to practice in Texas, we anticipated that a significant number of graduates would return to the region to provide health care for medically underserved Texans. In fact, of the 20 graduates who are now in practice, 12 are practicing in or near South Texas: 4 in San Antonio, 2 in Victoria, 2 in Weslaco, 2 in McAllen, 1 in Edinburg, and 1 in Alamo. Another 9 PHC graduates are attending residency programs in South Texas (5 in San Antonio, 2 in Corpus Christi, and 2 in Harlingen). Table 3 shows the areas of specialty that PHC graduates have pursued from 1994 to 2008. Thirteen have completed or are currently completing subspecialty fellowships. While time will provide a greater opportunity to compare PHC graduates with other BCM graduates in terms of practice outcomes, experience to date suggests that more PHC graduates have elected to pursue primary care residency programs (about 60%) compared with other BCM graduates who enter primary care programs (between 45% and 50%) in recent years. Anecdotally, from conversations with these young physicians, it is clear that many intend to return to South Texas eventually to practice. Others have expressed a desire to continue in academia with a role in teaching and mentoring. Still others have chosen to practice among other underserved populations, including one family physician who spent a year with the BCM Pediatric AIDS Corp in Africa and who is now working as a faculty member in a family medicine residency program in northern California. Six PHC graduates are academic faculty members (four clinical instructors and two assistant professors), all of them Latino.
Pipeline Program Strategies Can Work
As a pipeline program strategy to increase opportunities for students from underrepresented and disadvantaged backgrounds to access medical education, the PHC has been highly successful. The program has increased the number of South Texas students matriculating to medical school from 10 medical school matriculants from all South Texas undergraduate institutions in 1994 (6 at UTPA) to an average of 12 matriculants per year from the PHC at UTPA. However, the impact of the PHC has been much broader than these numbers suggest. The number of students matriculating to medical school from UTPA has increased from 6 students in 1994 (at the inception of the PHC) to an average of more than 30 students annually between 2001 and 2009, including PHC graduates and students who are not a part of the BCM–UTPA partnership. At the time the PHC was established, UTPA instigated a major overhaul of the science departments, which, together with the opportunities afforded by the PHC, attracted more academically able students to the university. Other Texas medical schools subsequently implemented similar pipeline programs at UTPA, further increasing medical school access for interested South Texas students. In 2002, the PHC was recognized by the Texas Higher Education Coordinating Board and awarded its “Star Award” for exceptional contributions to Texas higher education. Also in 2002, the Texas legislature created the Joint Admission Medical Program, a program modeled after the PHC to increase access to medical education for economically disadvantaged students statewide. These programs have resulted in even more opportunities for minority and economically disadvantaged students to attend medical school at UTPA and other universities across the state.
As far as increasing the diversity of the student body at BCM, the PHC has indeed contributed significantly to the number of URM students in recent entering classes. From 1998 through 2008, the PHC has been responsible for 38% of all BCM Latino matriculants (75 of 196). In 1997, the year before PHC students began to enter BCM, 7% of BCM's entering class was Latino. The PHC has helped to improve the numbers of entering Latino students from 7% in 1998 to 13% of total entering enrollment in 2008. In addition, most PHC graduates are bilingual, and 40% of them are economically disadvantaged, thus enhancing not only ethnic and geographic but also socioeconomic diversity among the BCM student body. PHC graduates have contributed significantly to the BCM community as leaders of organizations, student body presidents, chief residents, researchers, mentors for other students, and recipients of numerous awards and scholarships. Many have participated in educational outreach activities in South Texas to recruit and inspire high school and college students by sharing their stories, serving as role models and mentors, thereby encouraging the next generation of students to continue on the pathway to medicine.
Finally, early results regarding graduate practice patterns are encouraging in that many Latino graduates of the PHC are returning to South Texas to practice medicine. As of 2008, the PHC has produced a total of 65 MDs (83% Latino), 20 of whom have completed training and are currently in practice (all Latino), with 12 in South Texas. This is a significant outcome in a state whose population is 40% Latino, given that only 11% of the state's physicians are Latino.38 Six PHC graduates, all Latino, are practicing at academic medical centers and have faculty appointments. This last statistic is especially encouraging, as only 4.2% of faculty members in U.S. academic medical centers are Latino.39 The PHC may well represent a new and growing source of academically talented Latino medical professionals who may be recruited to academic medicine to increase diversity among medical faculty, help train culturally competent medical professionals in the future, and address the health care and research needs of our nation's diverse population.
The data presented in this article demonstrate long-term outcomes of a unique and successful pipeline program. Although partially anecdotal, these results are important in validating the role of pipeline programs to make a meaningful contribution to the diversity of the medical workforce. The success of such programs depends largely on the commitment of the partner institutions to maintain ongoing support and adapt program offerings and services as evidence of needs arises.
The authors gratefully acknowledge the support of the U.S. Department of Health and Human Services, Bureau of Health Professions, The Houston Endowment, Inc., The Robert Wood Johnson Foundation, the Long Foundation, The Brown Foundation, and several anonymous donors.
1Institute of Medicine. The Right Thing to Do: The Smart Thing to Do: Enhancing Diversity in Health Professions. Washington, DC: The National Academies Press; 2001.
2Institute of Medicine. In the Nation's Compelling Interest: Ensuring Diversity in the Health-Care Workforce. Washington, DC: The National Academies Press; 2004.
6Saha 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.
7Patterson DG, Carline JD. Promoting minority access to health careers through health profession–public school partnerships: A review of the literature. Acad Med. 2006;81(6 suppl):S5–S10.
8Carline JD, Patterson DG. Characteristics of health professions schools, public school systems, and community-based organizations in successful partnerships to increase the numbers of underrepresented minority students entering health professions education. Acad Med. 2003;78:467–482.
9See special edition of Academic Medicine: Educational Programs to Strengthen the Medical School Pipeline. Acad Med. 1999;74:305–407 and 431–452.
10California Endowment. Strategies for Improving the Diversity of the Health Professions. San Francisco, Calif: Center for California Health Workforce Studies, University of California, San Francisco; 2003.
12Steinecke A, Beaudreau J, Bletzinger RB, Terrell C. Race-neutral admission approaches: Challenges and opportunities for medical schools. Acad Med. 2007;82:117–126.
14Cleveland EF, Steinecke A, eds. Lessons Learned From the Health Professions Partnership Initiative, 1996–2005. Special supplement. Acad Med. 81(6 suppl):S1–S61.
15Carline JD, Hunt DD, Patterson DG, Garcia C. Participation in enrichment programs and its effect on interview scores of applicants to the University of Washington School of Medicine. Acad Med. 1999;74:360–362.
16Cantor 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.
17Fang WL, Woode MKA, Carey RM, Apprey M, Schuyler JM, Atkins-Brady TL. The medical academic advancement program at the University of Virginia School of Medicine. Acad Med. 1999;74:366–369.
18Ferry PG, King JE, Thomson WA. Retrospective pre/post survey design to assess student gains in knowledge, confidence and motivation in a premedical summer enrichment program. The Advisor. 2004;24(4):5–10.
19Parrish AR, Daniels DE, Hester RK, Colenda CC. Addressing medical school diversity through an undergraduate partnership at Texas A&M Health Science Center: A blueprint for success. Acad Med. 2008;83:512–515.
20Thomson WA, Denk JP. Promoting diversity in the medical school pipeline: A national overview. Acad Med. 1999;74:312–314.
21Cosgrove EM, Harrison GL, Kalishman S, et al. Addressing physician shortages in New Mexico through a combined BA/MD program. Acad Med. 2007;82:1152–1157.
22Roman SA, McGanney ML. The Sophie Davis School of Biomedical Education: The first 20 years of a unique BS–MD program. Acad Med. 1994;69:224–230.
23Thomson WA, Ferry PG, King JE, Martinez-Wedig C, Michael LH. Increasing access to medical education for students from medically underserved communities: One program's success. Acad Med. 2003;78:454–459.
24Moreno N. Science education partnerships: Being realistic about meeting expectations. Cell Biol Educ. 2005;4:30–32.
25Thomson WA, Denk JP, Ferry PG, Martinez-Wedig C, Michael LH. Producing physicians for South Texas. Tex Med. 1999;95:52–57.
28Barr AB, Gonzalez ME, Wanat SF. The leaky pipeline: Factors associated with early decline in interest in premedical studies among underrepresented minority undergraduate students. Acad Med. 2008;83:503–511.
29Alexander CA, Chen E, Grumbach K. How leaky is the health career pipeline? Minority student achievement in college gateway courses. Acad Med. 2009;84:797–802.
30Odom KL, Roberts LM, Johnson RL, Cooper LA. Exploring obstacles to and opportunities for professional success among ethnic minority medical students. Acad Med. 2007;82:146–153.
31Koenig JA, Sireci SG, Wiley A. Evaluating the predictive validity of MCAT scores across diverse applicant groups. Acad Med. 1998;73:1095–1106.
32Roth KS, Riley WT, Brandt RB, Seibel HR. Prediction of students' USMLE Step 2 performances based on premedical credentials related to verbal skills. Acad Med. 1996;71:176–180.
33Veloski JJ, Callahan CA, Xu G, Hojat M, Nash DB. Prediction of students' performance on licensing examinations using age, race, sex, undergraduate GPAs, and MCAT scores. Acad Med. 2000;75(10 suppl):S28–S30.
34Julian ER. Validity of the Medical College Admission Test for predicting medical school performance. Acad Med. 2005;80:910–917.
35Basco WT, Way DP, Gilbert GE, Hudson A. Undergraduate institutional MCAT scores as predictors of USMLE Step 1 performance. Acad Med. 2002;77(10 suppl):S13–S16.
36Montague JR, Frei JK. A twelve-year profile of students' SAT scores, GPAs, and MCAT scores from a small university's premedical program. Acac Med. 1993;306–308.
37Silver B, Hodgson CS. Evaluating GPAs and MCAT scores as predictors of NBME I and clerkship performance based on students' data from one undergraduate institution. Acad Med. 1997;72:394–396.
39Association of American Medical Colleges. Diversity in Medical Education: Facts & Figures 2008. Washington, DC: Association of American Medical Colleges; 2008.