Secondary Logo

Journal Logo

Featured Topic: Specialty Choice: Featured Topic Research Report

Impact of the University of California, Los Angeles/Charles R. Drew University Medical Education Program on Medical Students’ Intentions to Practice in Underserved Areas

Ko, Michelle MD; Edelstein, Ronald A. EdD; Heslin, Kevin C. PhD; Rajagopalan, Shobita MD; Wilkerson, LuAnn EdD; Colburn, Lois; Grumbach, Kevin MD

Author Information
  • Free


Strategies to address physician shortages in the United States usually fall into one of three categories: the applicant pool, medical education, and physicians’ practice environment.1 Multiple studies have identified characteristics of physicians practicing in underserved communities, including underrepresented minority race/ethnicity background,2–6 growing up poor7 or in an underserved area,4,8 interest prior to medical school,8 and incentives such as scholarships and loan repayment programs offered by the National Health Service Corps.8 However, less is understood about the impact of medical education on physicians’ eventual career choices.

A number of educational programs have been developed to encourage medical students to practice in underserved communities. In general, short-term clinical clerkships appear to have little effect on students’ plans or eventual practice.4,9,10 Longitudinal rural programs that combine selective admissions, preceptorships, required clerkships, and sometimes a community service or research requirement have shown the most success in producing physicians for rural areas.11–15

Furthermore, most of these efforts have focused on rural areas, not urban or other types of communities. In two studies, African American graduates of Howard University College of Medicine reported that 64% to 72% of their patients were African American, and one-half of minority graduates of the Sophie B. Davis/City University of New York program practiced in low-access areas of New York.16,17 However, these studies did not include comparisons to other programs. Keith et al.18 found that although African American graduates of Howard University College of Medicine and Meharry Medical College School of Medicine reported caring for a higher percentage of African Americans, there was no significant difference in the number of graduates practicing in designated Health Professional Shortage Areas or caring for medically indigent patients.

These studies illustrate the difficulty in discerning the impact of programs versus the impact of premedical and postgraduate factors. In specialty choice research, students interested in primary care at matriculation are more likely to retain these intentions upon graduation, and ultimately pursue primary care careers.19,20 Rabinowitz et al.8 determined that a strong desire to practice in an underserved community prior to medical school predicts future service. To our knowledge, no investigation has determined to what degree, if any, medical school experiences affect students’ career plans.

In this study, we examined the relationships between participating in the University of California, Los Angeles/Charles R. Drew University (UCLA/Drew) Medical Education Program, a comprehensive, inner-city-based program, and medical students’ intentions to practice in underserved areas. We included the potential confounding predictors of students’ demographics, interests prior to medical school, socioeconomic status, and other factors such as debt. We hypothesized that enrollment in the UCLA/Drew program is associated with increased intention to practice in underserved areas at the time of graduation from medical school, accounting for premedical and educational factors.


The UCLA/Drew Medical Education Program

The UCLA/Drew Medical Education Program began in 1981 to train students to fulfill the mission of the Charles R. Drew University: “to conduct education and research in the context of community service in order to train physicians and allied health professionals to provide care with excellence and compassion, especially to underserved populations.”21 A joint effort of the Charles R. Drew University in Los Angeles, California, and the University of California, Los Angeles (UCLA), David Geffen School of Medicine at UCLA, the program selects applicants for their demonstrated commitment to this mission. Individuals apply to the programs separately (prior to year one of medical school) for admission into the UCLA/Drew program. Once accepted, students participate in a two-week orientation at the Martin Luther King Jr./ Charles R. Drew Medical Center in South Central Los Angeles. UCLA/Drew students then complete the first two years of basic science training at UCLA alongside those enrolled in the UCLA School of Medicine. In years three and four, UCLA/Drew students undergo clinical training and complete a research project at the King/Drew Medical Center in South Central Los Angeles, one of the most impoverished areas in Los Angeles County. By graduation, Drew students complete a curriculum that is equivalent to the clinical curriculum at UCLA, but with an emphasis on minority and multicultural health, a longitudinal primary care rotation, and a health disparities research thesis.

Study sample

Our study sample consisted of 1,088 graduates of the UCLA School of Medicine and the UCLA/Drew Medical Education Program from 1996 to 2002, who completed the Graduation Questionnaire (GQ) of the Association of American Medical Colleges (AAMC).

Data sources

We conducted an analysis of secondary data collected by the AAMC for program improvement. Data were obtained from the AAMC biographical files, the Prematriculation Questionnaire (PMQ), Matriculating Student Questionnaire (MSQ), and GQ. The PMQ is administered with the Medical College Admissions Test, prior to a student’s admission into medical school. The MSQ and GQ are administered at the start and end of medical school, respectively. The institutional review boards of both UCLA and Charles R. Drew University approved the use of these archived data for research purposes.

Outcome variables

We measured intent to practice in an underserved area using the MSQ and the GQ data. In 1996–97, both questionnaires asked, “Do you plan to locate your practice in a socioeconomically deprived area?” In 1998–2002 both questionnaires asked, “Do you plan to locate your practice in a socioeconomically underserved area?” Our primary outcome measure was the student’s response to these questions on the GQ: “yes,” “no,” or “undecided.” For the analysis, we collapsed the response categories of “undecided” and “no,” because our objective was to identify students with an expressed intention to practice in underserved areas. In addition, we created variables to measure changes in practice intentions from matriculation to graduation: increased interest (students who answered “undecided/no” at matriculation and “yes” at graduation); maintained interest (students who answered “yes” at both time points); and decreased interest (students who answered “yes” at matriculation and “undecided/no” at graduation).

Predictor variables

Based on prior studies, we chose independent variables for their association with an intent to practice in underserved areas.4,8,22 We chose additional medical education variables for their potential relevance.

Demographic variables.

We obtained demographic variables from the AAMC biographical file. Age at matriculation was categorized as 22 years or younger or as older than 22 years. We dichotomized race/ethnicity into underrepresented minorities (URMs) and non-URMs. In the biographical file, URMs were defined as African American, Latino, and Native American/Alaskan/Hawaiian.

Social background variables.

Social background variables collected from the PMQ included parents’ education and occupation, and applicants’ high school community and region of origin when they applied to medical school. Because of data limitations, we were unable to ascertain whether a respondent was raised in an underserved area.

Medical education variables.

Medical education variables taken from both the MSQ and GQ included year of graduation, total educational debt, degree program, and enrollment in the UCLA School of Medicine or the UCLA/Drew Medical Education Program.

Interaction variables.

We tested several interaction terms between the UCLA/Drew program and URM race/ethnicity, gender, age, and matriculation plans. We identified no significant interactions in multivariate analysis, and ultimately these terms were not included in the final models.

Statistical analysis

We performed bivariate analyses to examine relationships between the predictor variables and the outcomes of interest, and calculated odds ratios (ORs) with 95% confidence intervals (CIs). We used the chi-square test for categorical variables and t tests for continuous variables. Independent variables were included in multivariate logistic regression analysis if they were associated with the outcome in bivariate analyses, or if the literature review suggested they may be important for their content.

Forty percent of graduates were missing data on the social background variables collected in the PMQ. Because the missing PMQ data led to a proportional reduction in sample size, two regression models were fitted. In the first model, regression without the social background variables was performed. In the second model, we repeated the regression with the addition of the social background variables.

Validation of the GQ practice intention item

Because the aim of these programs is to influence physicians to care for underserved populations upon completing training, we also examined whether intent at graduation translated into eventual practice in an underserved area. We attempted to validate the GQ intention item using a different sample: graduates of the UCLA School of Medicine from 1985 to 1995 who completed the GQ, finished training, and were practicing in California in 2003. We obtained practice activity and locations from the American Medical Association Physician Masterfile. Those physicians with an office address in California (67.8% of all UCLA graduates) were mapped and geocoded to a California Medical Service Study Area (MSSA).23 We defined a medically underserved MSSA as one that meets any one of the following criteria, described previously by Grumbach et al.23:

  • ▪ a federally designated Health Professional Shortage Area or Medically Underserved Area,
  • ▪ a rural area,
  • ▪ a high proportion of Latino or African American residents, or
  • ▪ a high proportion of low-income residents.

We divided the UCLA graduates into those who planned to practice in an underserved community on their GQ and those who did not, and compared the proportions of physicians in each group actually practicing in an underserved area in California.

We used SPSS 11.0 (SPSS Inc., Chicago, Illinois) for Windows and Mac OSX to perform statistical analyses.


Characteristics of the 1996–2002 graduates of the UCLA/Drew Medical Education Program and the UCLA School of Medicine are shown in Table 1. A total of 302 graduates (28%) were members of underrepresented minorities. Nine hundred sixty-eight (89%) graduates were enrolled in the UCLA School of Medicine, and 120 (11%) in the UCLA/Drew program. At matriculation, 291 (32.3% of 902 survey respondents) expressed an intention to practice in underserved areas. A larger proportion, 49.6% (447 graduates) was undecided, and 164 (18.2%) indicated that they did not plan to practice in underserved areas.

Table 1:
Characteristics of 1,088 Graduates of the University of California at Los Angeles (UCLA)/Charles R. Drew University Medical Education Program and the David Geffen School of Medicine at UCLA Who Responded to the Medical School Graduation Questionnaire, 1996–2002

Between 1996 and 2002, 1,088 students at UCLA/Drew and UCLA School of Medicine completed the GQ, for a response rate of 93.6%. Nine hundred two of these students (83%) completed both the GQ and MSQ, which allowed us to measure change in these students’ career plans from matriculation to graduation (see Figure 1). At graduation, 26.6% of respondents expressed an intention to practice medicine in underserved areas, and 73.4% were undecided or did not. Measures of change in intention constructed with matriculation and graduation data showed that 9.2% of students had an increase in interest in practicing in underserved areas and 14.9% of students lost interest. Of those initially interested, 53.9%, or 32.2% of the total sample, maintained their intention to practice in underserved areas (see Figure 1).

Figure 1:
Intentions at matriculation and graduation to work in an underserved area of medical students in the University of California, Los Angeles (UCLA)/Charles R. Drew University Medical Education Program and in the UCLA David Geffen School of Medicine, 1996–2002. Responses are from the AAMC’s Matriculating Student Questionnaire and Graduation Questionnaire administered at the start and end of medical school, respectively.

We fitted several multivariate models to estimate the independent association of enrollment in the UCLA/Drew program with intentions to practice medicine in underserved communities (multivariate results not provided). Compared with students of the UCLA School of Medicine, UCLA/Drew program students had greater adjusted odds of reporting an intention to work in underserved communities at graduation (OR: 9.40; 95% CI: 4.66–19.96). Between matriculation and graduation, UCLA/Drew students had greater odds of reporting maintained or increased intentions to work in underserved communities (OR: 8.84; 95% CI: 3.33–20.9) than did students of the UCLA School of Medicine (OR: 7.18; 95% CI: 2.79–18.48). Further, graduates of the UCLA/Drew program had substantially lower odds of decreased intention to work in underserved communities (OR: 0.24; 95% CI: 0.086–0.66).

For the validation of GQ practice intention measure, of the UCLA graduates from 1985 to 1995 who were practicing in California in 2003, 656 completed the GQ (69.6%). Graduates who planned to practice in an underserved area at graduation were significantly more likely to be practicing in an underserved community than were their counterparts who did not state this intention (OR: 1.80; 95% CI: 1.19–3.17).


Our study provides evidence that a medical education program can have a positive, reinforcing effect on student goals to practice in underserved areas. After accounting for two major premedical factors—URM race/ethnicity and intent at matriculation—as well as other important variables, participation in the UCLA/Drew program was independently associated with intention at graduation. Furthermore, participation in the program predicted, maintained, and increased intention during medical school.

As expected, we found the intent to practice in underserved areas at the start of medical school to be a strong predictor of eventual intent at graduation. This finding is consistent with an earlier study suggestive of school effect on practice plans.24 In discussing specialty choice, some educators argue that medical education bears little influence relative to premedical factors.25 However, it would be a mistake to conclude that medical training experiences have no impact on students’ career choices, and may actually be negative. Our results confirm previous observations that interest in underserved communities declines markedly from the first to the fourth year of medical school.26 During medical school, students encounter the challenges of practicing in underserved areas and may reconsider, especially when presented with a broad array of career options. However, a small percentage also developed an interest in practicing in underserved areas by the end of medical school. Students may have been inspired by clinical experiences with minority patient populations, role models, or other motivating factors experienced in medical school. These shifts during medical school, both toward and away from practice in underserved areas, suggest that educators have definite opportunities to guide students’ career choices.

Medical programs can provide a supportive environment for those students already inclined to practice in underserved communities. The cumulative number of experiences prior to and during medical training was a predictor of family medicine residents’ plans to practice in underserved areas.27 Our findings are consistent with those of the Jefferson Physician Shortage Area Program (PSAP), which found that students from rural areas and interested in family medicine in the first year of medical school were more likely to practice in rural family medicine than were non-PSAP students with the same qualities.15 We propose that an inner-city-based program can likewise reinforce the commitment of service among students predisposed to work in urban underserved areas. A multicomponent longitudinal program appears to be more effective than short-term exposures. UCLA students have required and elective clinical rotations in underserved settings, but as seen in other short-term efforts, this did not seem to have an effect on students’ interests. In the Drew program, students spend the entire third year in an urban underserved setting, with a primary care continuity experience, allowing time to develop ties to the patient population. Whereas the UCLA students in our study lost interest over time, Drew students maintained or increased their commitment (see Figure 1).

Students may also become motivated by interactions with faculty and other students who have similar goals. Thirty-six percent of the students in the Drew program are Hispanic/Latino and 42% are African American; this diversity may reinforce the already strong norm of community service in the institution. Medical students from schools with more racially diverse student bodies have more favorable attitudes toward underserved populations.28

As the first longitudinal study of the Drew program, our study had a number of limitations. We attempted to capture the short-term effects of medical education by focusing on plans at graduation. We recognize that graduates’ career intentions may not necessarily lead to eventual practice. However, our examination of the practice locations of earlier UCLA graduates found that the GQ practice intention item had reasonable validity as a predictor of actual practice location. Our study was also constrained by the limitations of a secondary data analysis of student surveys originally conducted for administrative purposes. The loss of social background data from the PMQ limits the conclusions that can be drawn about these predictors. However, if we consider these premedical factors as likely to function as predictors of intention at matriculation, we could conclude that a program effect remains. Even though the Drew educational experience appears to have had an appreciable influence, students’ underlying characteristics and dispositions remain the most powerful predictors of intentions at graduation. Although our ability to control for practice intentions at matriculation is a unique strength of the study and helps mitigate the problem of selection bias, it is possible that unobservable student characteristics still influenced intentions. Information on Drew students who applied to UCLA and were not accepted would be useful for exploring these types of unmeasured confounders, but this information was not readily available. Although the AAMC questionnaires are anonymous, both URM and Drew students may have been more likely to express intentions to work in underserved communities due to the social desirability of a positive response. However, as noted above, our validation study and prior research on URM physicians suggest that intentions are associated with actual practice behavior.4–7 Because we observed only two medical education programs, our findings are not necessarily generalizable to other programs. Nevertheless, given similar results produced by rural medical education programs, we are encouraged that the basic components—selective admissions and longitudinal experiences—can be applied to inner-city, predominantly minority communities.

Finally, we need to acknowledge that the most important outcome to examine, a medical student’s actual choice of practice location, is not included in our comparison. Future research will need to follow up with medical school graduates to estimate the impact of educational experience on practice over time.

As the “control valve” of physician supply, academic institutions are in a unique position to address the nation’s health care disparities. Our study points to the potential benefits of programs similar to UCLA/Drew. If more medical schools were to develop these types of service-oriented programs, educators could make a long-term contribution to alleviating physician shortage in our most disadvantaged communities.

The authors would like to thank Deborah Danoff, Carol Hodgson, Kehua Zhang, Deyu Pan, and Magda Shaheen for their assistance in the design of the study. Renee Taylor provided administrative assistance. The authors would also like to thank Keith Norris, Marcelle Willock, Neil Parker, and Theodore Miller for institutional support.


1Grumbach K, Coffman J, Liu R, Mertz E. Strategies for Increasing Physician Supply in Medically Underserved Communities in California. Berkeley: California Program on Access to Care, California Policy Research Center, 1999.
2Petersdorf RG, Turner KS, Nickens HW, Ready T. Minorities in medicine: past, present, and future. Acad Med. 1990;65:663–70.
3Kassebaum DG, Szenas PL, Caldwell K. Educational debt, specialty choices, and practice intentions of underrepresented-minority medical school graduates. Acad Med. 1993;68:506–11.
4Xu G, Veloski J, Hojat M, Politzer RM, Rabinowitz HK, Rattner SL. Factors influencing primary care physicians’ choice to practice in medically underserved areas. Acad Med. 1997;72(10 suppl 1):S109–11.
5Moy E, Bartman BA. Physician race and care of minority and medically indigent patients. JAMA. 1995;273:1515–20.
6Komaromy 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;334:1305–10.
7Cantor JC, Miles EL, Baker LC, Barker DC. Physician service to the underserved: implications for affirmative action in medical education. Inquiry. 1996;33:167–80.
8Rabinowitz HK, Diamond JJ, Veloski JJ, Gayle JA. The impact of multiple predictors on generalist physicians’ care of underserved populations. Am J Public Health. 2000;90:1225–28.
9Lynch DC, Willis SE. Can a 3-day preceptorship change first year medical students’ opinions about living and working in small towns? Fam Med. 2000;32:495–500.
10Dobie SA, Carline JD, Laskowski MB. An early preceptorship and medical students’ beliefs, values, and career choices. Adv Health Sci Educ. 1997;2:35–47.
11Adkins RJ, Anderson GR, Cullen TJ, Myers WW, Newman FS, Schwarz MR. Geographic and specialty distributions of WAMI program participants and nonparticipants. J Med Educ. 1987;62:810–7.
12Ramsey PG, Coombs JB, Hunt DD, Marshall SG, Wenrich MD. From concept to culture: the WWAMI program at the University of Washington School of Medicine. Acad Med. 2001;76:765–75.
13Stearns JA, Stearns MA, Glasser M, Londo RA. Illinois RMED: a comprehensive program to improve the supply of rural physicians. Fam Med. 2000;32:17–21.
14Dever GEA, Eveland AP, Tedders SH, Fehlenberg RD, Laurens MB, Harrelson JC. Impact of a population-based medical curriculum on specialty choice. J Health Care Poor Underserved. 2001;12:261–71.
15Rabinowitz HK. Critical factors for designing programs to increase the supply and retention of rural primary care physicians. JAMA. 2001;286:1041–48.
16Cregler LL, McGanney ML, Roman SA, Kagan DV. Refining a method of identifying CUNY Medical School graduates practicing in underserved areas. Acad Med. 1997;72:794–97.
17Johnson DG, Lloyd SM, Miller RL. A second survey of graduates of a traditionally black college of medicine. Acad Med. 1989;64:87–94.
18Keith SN, Bell RM, Williams AP. Affirmative action in medical education and its effect on Howard and Meharry: a study of the class of 1975. J Natl Med Assoc. 1988;80:153–58.
19Xu G, Hojat M, Brigham TP, Veloski JJ. Factors associated with changing levels of interest in primary care during medical school. Acad Med. 1999;74:1011–15.
20Owen JA, Hayden GF, Connors AF Jr. Can medical school admission committee members predict which applicants will choose primary care careers? Acad Med. 2002;77:344–49.
21Charles R. Drew University of Medicine and Science. Mission statement 〈〉. Accessed 10 June 2005.
22Keith SN, Bell RM, Swanson AG, Williams AP. Effects of affirmative action in medical schools: a study of the class of 1975. N Engl J Med. 1985;313:1519–25.
23Grumbach K, Hart G, Mertz E, Coffman JM, Palazzo L. Who is caring for the underserved? A comparison of primary care physicians and nonphysician clinicians in California and Washington. Ann Fam Med. 2003;1:97–104.
24Barnhart J, Shekelle P, Lewis C. The effect of a medical school’s admission and curriculum policies on increasing the number of physicians in primary care specialties. Acad Med. 1996;71:293–95.
25Pathman DE. Medical education and physicians’ career choices: are we taking credit beyond our due? Acad Med. 1996;71:963–68.
26Crandall SJ, Volk RJ, Loemker V. Medical students’ attitudes toward providing care for the underserved. Are we training socially responsible physicians? JAMA. 1993;269:2519–23.
27Tavernier LA, Connor PD, Gates D, Wan JY. Does exposure to medically underserved areas during training influence eventual choice of practice location? Med Educ. 2003;37:299–304.
28Elam CL, Johnson MMS, Wiggs JS, Messmer JM, Brown PI, Hinkley R. Diversity in medical school: perceptions of first-year students at four southeastern US medical schools. Acad Med. 2001;76:60–65.
© 2005 Association of American Medical Colleges