Secondary Logo

Journal Logo

Primary Care

Primary Care Specialty Choices of United States Medical Graduates, 1997–2006

Jeffe, Donna B., PhD; Whelan, Alison J., MD; Andriole, Dorothy A., MD

Author Information
doi: 10.1097/ACM.0b013e3181dbe77d
  • Free


The U.S. health care system is faced with a growing shortage of primary care physicians,1,2 which is particularly disconcerting in light of recent health care reform efforts to extend health insurance coverage to all. Proposed legislation to expand coverage and increase access to care will likely result in even greater demand for primary care specialists.3–5 To meet the growing national need for primary care physicians, our medical education system is challenged to develop strategies that will increase the numbers of U.S. medical graduates who pursue primary care careers, particularly as generalists rather than as subspecialists. Efforts to meet this challenge can be informed by an understanding of the trends in primary care specialty choice among recent U.S. graduates of Liaison Committee on Medical Education (LCME)-accredited medical schools in the context of these graduates' concurrent increases in debt loads and changing demographic characteristics. Figure 1 illustrates specialty choice trends by using annual programmatic data from the Graduation Questionnaire (GQ) of the Association of American Medical Colleges (AAMC).6 On the basis of a meta-analysis of the literature, Bland and colleagues7 described three components of a medical career decision model for primary care specialty choice: student characteristics, medical school characteristics, and students' perceptions of the medical specialty. A more recently published single-institution study used the Bland–Meurer model to test variables from the AAMC GQ that were predictive of primary care specialty choices; these authors concluded that primary care specialty choices should be examined on a specialty-specific basis because some variables, such as female gender, race/ethnicity, and plans to practice in an underserved area, were not uniformly predictive of all primary care specialty choices (which included internal medicine, family medicine, and pediatrics).8

Figure 1
Figure 1:
Programmatic data from the Association of American Medical Colleges' Graduation Questionnaire (GQ)6 from 1997 through 2008, showing the percentages of graduates who indicated that they did not plan to become board certified in a specialty or that they had chosen primary care or non-primary-care specialties. Primary care specialties included general internal medicine, general pediatrics, family medicine, obstetrics–gynecology, internal medicine subspecialties, and pediatrics subspecialties.

Factors that may influence primary care specialty choice may have a differential influence on the choice of a specific primary care specialty. An extensive review identified attendance at a public medical school, female gender, and interest in working in medically underserved communities as among the factors associated with greater likelihood of primary care specialty choice, particularly family medicine.2 That report noted that the relationship between debt and specialty choice was complex, particularly in the context of the increasing proportions of medical students who come from affluent backgrounds. In a recent cross-sectional study, male gender and attendance at a private medical school were found to be associated with fourth-year students' greater likelihood of choosing an internal medicine specialty, but a high level of debt was not independently associated with internal medicine specialty choice in the authors' multivariate model.9 In contrast to widespread concerns about the impact of declining U.S. student interest in family medicine and internal medicine on each specialty's workforce, a recent report reiterated the need to understand primary care workforce issues on a specialty-specific basis. For example, the shortage of physicians who deliver primary care to adult populations is not matched by such a shortage for pediatric populations.10

Studies in the literature to date have had some limitations. Studies that combine several primary care specialties (e.g., internal medicine, pediatrics, and family medicine) into a single primary care specialty category preclude the identification of factors that are associated with individual primary care specialties.10–12 On the other hand, results of studies that analyze factors associated with only one primary care specialty are not necessarily generalizable to other primary care specialties.8,9,13,14 Many studies have not distinguished between generalist and subspecialty career paths, but that is an important distinction, given the growing numbers of U.S. graduates of LCME-accredited medical schools who are pursuing subspecialty training in internal medicine and pediatrics.15–18 Finally, although an association between primary care specialty choices at matriculation and those at graduation has been reported,19,20 many students enter medical school without having decided on a specialty choice,21 and students often change their specialty choice during medical school.7,22–24 However, no study has fully explored the student characteristics (e.g., attitudes and beliefs) at the time of matriculation to medical school that could influence students' specialty choices at graduation.2

Hence, whereas a substantial body of work on factors associated with particular primary care specialty choices exists, these factors have not been examined in the context of the full range of students' planned primary care specialty choices among a nationally representative sample of all U.S. graduates of LCME-accredited medical schools. We conducted the present study to address this gap in the literature.

We hypothesized that, during the recent period of declining interest in primary care specialties, the characteristics of students who have chosen to enter one of those specialties might vary considerably on a specialty-specific basis. To test this hypothesis, we conducted a retrospective study of a nationally representative, population-based cohort of persons who graduated from LCME-accredited medical schools between 1997 and 2006. We sought to identify demographic, attitudinal, and career intention variables that were associated with each of six primary care specialty choices during this decade of change in specialty choice preferences among U.S. medical graduates.


We analyzed individualized, linked data for all 1997–2006 U.S. graduates of LCME-accredited medical schools who completed two AAMC questionnaires, the Matriculating Student Questionnaire (MSQ) and the GQ. The AAMC administers both questionnaires annually; survey completion is voluntary and confidential,6,25 and all data we received had been deidentified by the AAMC. The 1997–2006 graduate cohort comprises the emerging U.S.-trained physician workforce, which has been marked by a decline in primary care specialty choice since 1998.1 The institutional review board at Washington University School of Medicine approved this study.


We examined demographic, attitudinal, and career intention variables on the MSQ and GQ in association with the specialty choice outcomes of interest. The specialties that encompass primary care specialties are variably defined in the literature. We defined primary care specialties in accordance with the primary medical care designation criteria of the Bureau of Health Professions of the Health Resources and Services Administration; these specialties include family medicine, internal medicine (both general and subspecialty), obstetrics–gynecology, pediatrics (both general and subspecialty), and combined internal medicine/pediatrics.18

On the basis of graduates' responses to GQ items pertaining to their intended specialty and their further intent to subspecialize, we assigned graduates to one of eight primary care specialty choice categories for analysis. Six of these categories were family medicine, obstetrics–gynecology, general internal medicine (including internal medicine/pediatrics), internal medicine subspecialties, general pediatrics, and pediatrics subspecialties. Because internal medicine/pediatrics was not a specialty choice on the GQ until 2000, graduates who made this specialty choice were included in the general internal medicine specialty choice category. Because of the substantial increase since 1995 in the participation of U.S. graduates of LCME-accredited medical schools in Accreditation Council on Graduate Medical Education (ACGME)-accredited subspecialty training programs in internal medicine and pediatrics,18 we created separate categories for general internal medicine (internal medicine specialty without subspecialty intent), internal medicine subspecialties (internal medicine with subspecialty intent), general pediatrics (pediatrics specialty without subspecialty intent), and pediatrics subspecialty (pediatrics specialty with subspecialty intent), according to the graduates' responses to the items pertaining to their plans for subspecialty training. The family medicine category was not subdivided into general and subspecialty categories for several reasons: Family practice or subspecialty already was a single GQ specialty choice option, there are relatively few ACGME-accredited positions for family medicine subspecialty training, and only small proportions of residents completing training in family medicine plan any additional training.6,26,27 Similarly, because there are no ACGME-accredited subspecialty training programs in obstetrics–gynecology,27 we did not further classify into general or subspecialty categories the graduates who chose obstetrics–gynecology. Graduates who made all other specialty choices composed the “other specialty” group. Only those GQ respondents who indicated they planned to become board certified in a specialty could respond to the next question about specialty choice. Thus, graduates who responded “no” or “undecided” to that question on the GQ were grouped in the no-board-certification specialty group.

We examined 13 variables in association with specialty choice. Demographic variables included year of graduation, gender (female versus male), self-identified race/ethnicity (Asian/Pacific Islander, other/unknown, or underrepresented minority [URM; refers to minorities who are underrepresented in medicine, including persons who are black, Hispanic/Latino, and American Indian/Alaska Native] versus white), total debt at graduation, and type of medical school (private versus public); parent occupation (at least one parent is a physician or at least one parent is a professional but not a physician versus all other parent occupations) was created from two MSQ items about the occupations of the student's mother and father. We included race/ethnicity because it has been found to be associated with medical students' specialty choice.28 Students who self-identified as “other” or multiple races or who chose not to respond to these items on the GQ were combined in the “other/unknown” race/ethnicity category for analysis. We included type of medical school because public medical schools reportedly produce greater proportions of primary care physicians than do privately owned medical schools.2

GQ items about career intentions included students' plans to practice in an underserved area (yes or undecided versus no) and their preferred career setting. We created four categories of preferred career settings from a list of the career settings listed on the GQ: undecided, full-time university faculty (including teaching/research in both basic science and clinical science), full-time (nonuniversity) clinical practice (the reference group), and other (including state/federal or medical/health care administration and nonuniversity research scientist).

Twenty-two MSQ items inquired about the importance of various factors in students' choice of medicine as a career; answers were given on a 5-point scale ranging from 0 (not at all) to 4 (very important). In addition, 17 MSQ items available in all years of our study asked about students' perceptions of the profession of medicine; answers were given on a 5-point scale from 1 (strongly agree) to 5 (strongly disagree). Responses to these items were reverse-coded for analysis so that higher scores indicated stronger agreement with an item, and exploratory principal components analysis was used for data reduction. In a previous study by our group29 that included a subsample of 39,462 GQ respondents from this same cohort, 6 of these 17 items loaded on two factors that were relevant to the present study. The items on these two factors had high internal consistency reliability as measured by standardized Cronbach α coefficients: (1) altruistic beliefs about access to health care and provision of care, regardless of a patient's ability to pay (α = .70), and (2) perceptions of how the demands of medicine can interfere with family and other interests (α = .92).

Statistical analysis

For data reduction, we used an iterative process of principal components analysis with varimax rotation of each of the two sets of attitudinal items. We used eigenvalues >1.0 to determine the number of factors and Lautenschlager's30 tables to confirm the number of factors. We used Cronbach α to measure the internal consistency of items on each factor and computed mean scores for each factor. We measured the associations between categorical variables using chi-square tests and those between continuous and categorical variables using analysis of variance. We used multivariate logistic regression to identify MSQ and GQ variables that were independently associated with each of the six primary care specialties and the no-board-certification specialty group that were compared with the “all other specialties” group (reference group). We report descriptive statistics for each variable of interest, as well as adjusted odds ratios and 95% confidence intervals from the multivariate logistic regression models. All tests were performed with SPSS software (version 16.0; SPSS, Inc., Chicago, Illinois). Two-sided P values <.05 were considered statistically significant.


Of all 1997–2006 U.S. graduates of LCME-accredited medical schools (N = 158,091), 129,545 (81.9%) completed the GQ (in full or in part). Response rates varied from a low of 60.1% in 2005 to a high of 91.7% in 2000.6,31 MSQ records were identified for 112,117 of the students who completed the GQ (70.9% of all 1997–2006 graduates and 86.5% of GQ respondents). We excluded from analysis 2,377 combined MD/PhD program graduates, because these graduates' characteristics and career intentions differ markedly from those of their MD-degree counterparts.32 Of the remaining 109,740 graduates with linked MSQ and GQ records, 7,067 did not respond to one or more items of interest and were excluded from analysis, which left 102,673 graduates to be included in our analysis (64.9% of all 1997–2006 U.S. graduates of LCME-accredited medical schools).

From 1997 to 2006, there was an overall decrease in the proportions of GQ respondents who chose general internal medicine (from 15.7% to 6.7%), general pediatrics (from 10.2% to 6.6%), family medicine (from 17.6% to 6.9%), and obstetrics–gynecology (from 8.2% to 6.1%), whereas there was an overall increase in the proportions who chose internal medicine (from 6.8% to 11.4%) and pediatrics (from 2.2% to 4.4%) subspecialties. However, the increasing proportions of GQ respondents who chose these subspecialties did not offset the declines in the choices of the four generalist specialties, so the combined total proportion of all graduates in our study sample who chose these six primary care specialties showed a net decrease from 60.7% in 1997 to 42.1% in 2006. The proportions of GQ respondents who chose other specialties increased steadily, from 36.9% in 1997 to 51.5% in 2006. The proportion of graduates in the no-board-certification specialty group fluctuated from a low of 2.4% in 1997 to a high of 13.0% in 2003 and then declined again to 6.3% in 2006. Table l shows the overall proportions of graduates who entered each specialty category and provides descriptive statistics for each categorical variable grouped by specialty choice.

Table 1
Table 1:
Sample Characteristics by Specialty Choice at Graduation Among 1997–2006 U.S. Medical Graduates Who Completed Both the Association of American Medical Colleges' Matriculating Student Questionnaire and Graduation Questionnaire

The decline in generalist–primary care specialty choices coincided with other trends in medical education over the study period. In the years from 1997 to 2006, in our study sample, the proportion of female graduates reached parity with the proportion of male graduates, growing from 42.5% to 50.8%; the proportion of Asian/Pacific Islander students rose from 14.4% to 17.6%, and the proportion of medical school graduates with at least $150,000 in total debt at graduation rose from 6.7% to 35.9%. However, the change in the proportion of URM graduates (from 10.8% to 9.0%) was not substantive.

Principal components analysis of the 22 MSQ items on the importance of factors in the choice of medicine as a career resulted in 15 items that loaded on three factors: measuring students' attribution of the importance of innovation and research (6 items), the importance of social responsibility (5 items), and the importance of prestige (4 items); higher scores indicate that the students attributed a greater importance to these items (Table 2). We excluded seven items from further analysis because they loaded on more than one factor or did not load on any of these three factors. A separate principal components analysis of the six previously used MSQ items measuring students' beliefs about the medical profession yielded two factors; higher scores indicated more altruistic beliefs about health care and beliefs that the demands of medicine interfere with family and other interests (Table 2). Factor loadings, Cronbach α coefficients, and mean scores (SDs) of items on each attitudinal factor are shown in Table 2. Descriptive statistics for each of these attitudinal factors, grouped by students' specialty choice, appear in Table 3.

Table 2
Table 2:
Factor Loadings and Descriptive Statistics of Items Measuring 102,673 Students' Perceptions on the AAMC Matriculating Student Questionnaire and Cronbach α Showing Internal Consistency of Items Loading on Each Factor
Table 3
Table 3:
Attitudinal Variables on the 1997–2006 AAMC Graduation Questionnaire by Specialty Choice Categories (N = 102,673)

Table 4 shows the results of the multivariate logistic regression model. The Pearson chi-square goodness-of-fit test indicated that the model was a good fit to the data (P = .416). Students who, at matriculation, attributed greater importance to social responsibility in their choice of a medical career and who had more altruistic beliefs about health care were more likely to choose one of the primary care specialties at graduation, whereas students who attributed greater importance to prestige in medicine were less likely to do so. Female graduates and those who responded “yes” or “undecided” about plans to practice in underserved communities were more likely to choose general internal medicine, general pediatrics, family medicine, or obstetrics–gynecology; graduates in more recent years (e.g., 2005 or 2006) and graduates who attributed greater importance to intellectual challenge, innovation, and research in their choice of a medical career, who had a physician parent, and who planned to pursue full-time academic medicine careers were less likely to do so. Graduates who were planning full-time academic medicine careers were more likely than were those planning full-time nonuniversity clinical practice to choose internal medicine subspecialties and pediatric subspecialties.

Table 4
Table 4:
Multivariate Logistic Regression Models of Variables on the 1997–2006 AAMC Graduation Questionnaire Associated With Primary Care Specialties and “No-Board-Certification Specialty Choice,” Each Compared With All Other Specialty Choices (N = 102,673)
Table 4
Table 4:

URM graduates were less likely than were white graduates to choose general internal medicine, general pediatrics, or family medicine but more likely to choose obstetrics–gynecology. Asian/Pacific Islander graduates were less likely than white graduates to choose family medicine, general pediatrics, and obstetrics–gynecology but more likely than white graduates to choose general internal medicine and internal medicine subspecialties.

The associations between debt and each specialty choice category varied. Graduates with higher levels of debt were less likely to choose general internal medicine, internal medicine subspecialties, general pediatrics, pediatric subspecialties, or a no-board-certification specialty. Graduates with higher debt were more likely to choose obstetrics–gynecology. Debt was not significantly associated with family medicine specialty choice.


Our findings regarding overall primary care workforce trends are consistent with those of other studies, which have documented substantial shifts away from primary care specialty choice among U.S. graduates of LCME-accredited medical schools since 1997.2,16,17 It is notable that the decline in choice of primary care specialties (i.e., family medicine, general internal medicine, general pediatrics, and obstetrics–gynecology) that we observed in the 1997–2006 cohort of U.S. medical graduates continues. As shown in Figure 1, even with increases in the proportions of graduates choosing internal medicine subspecialties and pediatrics subspecialties, only 30.3% of all GQ respondents in 2008 chose one of the six primary care specialties included in our analysis.6 Because analyzing all primary care specialties together as one group can mask potential differential effects that particular variables might have on the choice of different specialties,11 an understanding of various demographic, attitudinal, and career intention variables associated with each of the six primary care specialty choices can inform both our understanding of primary care workforce trends and the design of tailored interventions to address shortages in the primary care workforce on a specialty-specific basis.

Unlike previous studies using GQ data, which systematically excluded from analysis graduates not definitely committed to specialty board certification,16,33,34 our analysis included the GQ respondents who did not definitely plan to become specialty board certified and who did not make a specialty choice on the GQ. This growing proportion of graduates not committed to board certification in a specialty coincided with the declining proportions of U.S. graduates who chose primary care specialties. Although the rate of change in proportions of graduates in this group fluctuated somewhat over time, programmatic GQ data indicate that the size of this no-board-certification specialty group increased fivefold, from 3.1% of all GQ respondents in 1997 to 17.5% of all GQ respondents in 2008 (Figure 1).6 Our finding that growing proportions of graduates did not plan to become board certified in any specialty may seem paradoxical, given the growing de facto necessity of specialty board certification for physicians in virtually all clinical settings.35,36 Some of the associations between the no-board-certification specialty group and some of the predictor variables were similar to associations we observed with predictors of the different primary care specialties we examined, but the no-board-certification specialty group did not consistently look like any primary care specialty group. Therefore, that group seems to represent an important graduate group to study. Not only is the no-board-certification specialty group growing, but, according to our findings, at least some graduates in the group might become primary care physicians; yet, their career paths remain unknown. Exploration of the career paths pursued by this relatively underrecognized and understudied group of graduates seems warranted in the context of growing concerns about the projected shortage of practicing physicians in primary care.

Our findings regarding the association between female gender and greater likelihood of choosing one of the four generalist–primary care specialties suggest that the steady increase in the proportion of women among U.S. medical school graduates over the study period played a critical role in limiting the overall decline in proportions of graduates who planned generalist–primary care careers. Since the 2003–2004 academic year, when gender parity among matriculants was reached, the proportion of women entering medical school has not increased further, and the proportion of female graduates from LCME-accredited medical schools is plateauing.31 In this context, the decline in graduates' choice of generalist–primary care specialties may be exacerbated by a lack of continued growth in the proportion of women among all graduates.

Compared with white graduates, URM graduates were less likely to choose general internal medicine, general pediatrics, and family medicine specialties, and Asian/Pacific Islander graduates were less likely to choose family medicine, obstetrics–gynecology, general pediatrics, and pediatrics subspecialties. Thus, efforts to increase levels of racial/ethnic diversity among U.S. graduates of LCME-accredited medical schools will not, alone, necessarily result in greater numbers of generalist–primary care physicians, especially in general pediatrics and family medicine.

We observed that a student's having a physician parent had a pervasive negative effect on graduates' choice of any generalist–primary care specialty, but graduates with at least one parent in other professional occupations were more likely than were graduates without a parent who was a physician or another type of professional to choose generalist internal medicine, generalist pediatrics, and internal medicine subspecialties. Physicians in practice may hold particularly negative perceptions about generalist–primary care careers, and they may convey these perceptions to their children. Because 16% of graduates in our sample reported having a physician parent and 24% reported having a parent who was a professional (but not a physician), the impact of parents' occupations on graduates' specialty choices may be considerable. This finding regarding the occupations of graduates' parents, along with the observation that graduates' attitudes at matriculation were among the predictors of primary care specialty choice, is consistent with the thesis that primary care specialty choices reflect the impact of a wide range of variables, including medical students' perspectives about primary care at the time of matriculation.2 That several attitudinal factors at the time of matriculation (e.g., more altruistic beliefs and attribution of greater importance to social responsibility and of lesser importance to prestige) were associated with a greater likelihood of primary care specialty choice at graduation may be of particular interest to medical school admissions committees as they seek to matriculate students with attitudes and career aspirations that are well aligned with the institution's missions and goals.37

The relationships between total medical school debt and primary care specialty choice varied and were generally modest. Graduates with higher debt were less likely to choose general pediatrics, pediatrics subspecialties, general internal medicine, or internal medicine subspecialties but more likely to choose obstetrics–gynecology, which is characterized by higher expected median incomes than those for physicians practicing in other primary care specialties, especially general internal medicine, general pediatrics, or family medicine.38 Debt was not significantly associated with a choice of family medicine. Therefore, loan repayment and/or debt-forgiveness programs may have a role in promoting internal medicine and pediatrics specialty choices among indebted graduates, but they may not have any impact on the choice of family medicine.

The observed associations between specialty choice and the type of medical school (public or private) suggest that variations in schools' educational missions and strengths (e.g., in research, teaching, and service to the community) play a role in graduates' specialty choices. Our observation that private-school graduates were less likely than public-school graduates to choose family medicine or obstetrics–gynecology is consistent with the mandate of many public medical schools to educate future primary care physicians. The association between graduating from a public medical school and a greater likelihood of choosing family practice has previously been reported.2 Some authors suggested that the effect of a school's mission on students' specialty choice is likely mediated by the faculty, who influence students' specialty choice through policy decisions, curriculum, and faculty attitudes toward primary care.7 Schools that have primary care missions and that historically have produced more generalists are more likely to encourage their students to pursue primary care specialties.39 Graduates intending to practice in underserved communities also were more likely to choose one of the primary care specialties, particularly family medicine, which is consistent with previous reports.2,20 It is important that the predictive validity of planning to practice in an underserved community at graduation has been established.2,40 Thus, our results suggest that strategies to increase the low numbers of U.S. medical graduates who choose family medicine careers could be especially critical to the success of ongoing national efforts to address the projected physician workforce shortage and could also help the United States meet its projected health care needs, particularly in underserved communities.2,41

Graduates who preferred full-time academic medicine careers were more likely to choose internal medicine subspecialties and pediatrics subspecialties and less likely to choose the other primary care specialties, particularly family medicine. Efforts to inform students about the wide range of professional activities pursued by faculty physicians could help change students' perceptions that family medicine may be incompatible with an academic medicine career.42 Because specialty choice decisions may be influenced to some extent by the composition of the academic medicine workforce, a more visible presence of generalist–primary care physicians among the faculty may be a strategy to increase interest in generalist–primary care specialties. This strategy may be particularly true for family medicine: The numbers of physicians with full-time faculty appointments in family medicine departments are very low relative to the numbers of family medicine physicians in the overall U.S. physician workforce.43,44 These findings, along with our findings about the offspring of physicians, collectively suggest that physicians themselves, either as parents or as academicians, may have contributed to the decline in interest in generalist specialties among recent U.S. medical school graduates.

Our study has several strengths and limitations. Among its strengths was the use of linked MSQ and GQ records for a population-based cohort of individual students. Although we examined specialty choice at the time of graduation only, the GQ is administered after graduates complete the residency application process and after they have largely finalized their specialty choice decisions. Therefore, the specialty choices reported on the GQ likely reflect the specialties that graduates indeed entered for their graduate medical education.

One limitation of our study was that, because completion of the MSQ and GQ was voluntary, our study sample did not include all U.S. graduates of LCME-accredited medical schools. However, the sample was demographically representative of all 1997–2006 U.S. medical school graduates.31 Another limitation was that our analysis pertained only to the U.S. graduates of LCME-accredited medical schools. Our findings cannot be generalized to all physicians, particularly to osteopathic graduates or graduates of international medical schools; the latter group, who did not graduate from an LCME-accredited medical school, currently represents 34% of the physicians in ACGME-accredited residency training positions.27 However, current efforts to address the anticipated U.S. physician workforce shortage include strategies to increase the number of physicians by increasing enrollment at existing U.S. LCME-accredited medical schools and by creating new U.S. LCME-accredited medical schools. Thus, examination of the specialty choices of graduates of these U.S. LCME-accredited medical schools informs our understanding of the largest component (and a still-growing component) of the contemporary physician workforce in the United States.

Although this fact is not a limitation per se, there are variables besides the ones that we included in our analysis that influence specialty choice outcomes. Those additional variables include a graduate's personality, interests, skills, perceptions about the lifestyle of physicians in certain specialties,9,33,45,46 and medical school academic performance measures, particularly the standardized United States Medical Licensing Examination Step l and Step 2CK (clinical knowledge) scores. There is considerable self-selection among U.S. senior students on the basis of Step l scores in the specialties to which they apply for graduate medical education positions and considerable selectivity on the basis of Step l scores demonstrated by program directors in choosing applicants whom they consider acceptable for their training programs.47 Because student perceptions about the income potential of various specialties were not measured on the MSQ or GQ during the study period, we also could not determine the extent to which growing income disparities between generalist–primary care physicians and physicians in other specialties may have contributed to the declining interest in generalist–primary care specialties among U.S. graduates of LCME-accredited medical schools (regardless of debt load) since 1997.38,45


Our study findings further our understanding of factors that have contributed to declining trends in primary care specialty choices among U.S. graduates of LCME-accredited medical schools. Two observations may be particularly relevant in the context of primary care physician workforce shortages. First, graduates in all six primary care categories included in our study and graduates in the no-board-certification specialty group shared altruistic beliefs and perceptions about social responsibility in medicine at the time of matriculation, which suggests that students do indeed enter medical school with particular beliefs and attitudes that can predict their specialty choices at graduation. It is also important, however, that we identified differences in predictors of choice of the six primary care categories in our study. Our observation that the six primary care categories did not share a uniform set of predictors reinforces the critical importance of understanding primary care workforce issues in a specialty-specific manner.10 This consideration is particularly notable in the context of the variable magnitude and direction of the associations we observed between students' characteristics, such as gender, race/ethnicity, debt, and primary care specialty choices. Second, the career intentions of contemporary graduates appear to be expanding beyond the traditional paradigm of specialty choice for board certification. Growth in this understudied and underrecognized group of graduates, who are not definitely planning board certification in any specialty, has coincided with the steadily declining proportions of graduates planning board certification in general primary care specialties, and thus further research is warranted to better understand the career paths of this group.


The authors wish to thank our colleagues Jason Cantow, MS, MBA, and David Matthew, PhD, at the AAMC, Washington, DC, for provision of the data and assistance with coding, and Heather Hageman, MBA, director, Educational Planning and Program Assessment, Office of Medical Student Education, Washington University School of Medicine, for assistance with acquisition of the data from the AAMC and review of the manuscript.


Funding for this project was provided by the Office of the Dean of Washington University School of Medicine.

Other disclosures:


Ethical approval:

The institutional review board at the Washington University School of Medicine provided ethical approval of the study.


The conclusions of the authors are not necessarily those of the AAMC or AAMC staff.


1Whitcomb ME, Cohen JJ. The future of primary care medicine. N Engl J Med. 2004;351:710–712.
2Phillips RL, Dodoo MS, Petterson S, et al. Specialty and Geographic Distribution of the Physician Workforce: What Influences Medical Student and Resident Choices? Available at: Accessed November 20, 2009.
3Bodenheimer T, Grumbach K, Berenson RA. A lifeline for primary care. N Engl J Med. 2009;360:2693–2696.
4Chokshi DA. Ensuring progress in primary care—What can health care reform realistically accomplish? N Engl J Med. 2009;361:e43.
5Massachusetts Medical Society. MMS Physician Workforce Study. Available at: Accessed November 18, 2009.
6Association of American Medical Colleges. Graduation Questionnaire All Schools Reports. Available at: Accessed November 20, 2009.
7Bland CJ, Meurer LN, Maldonado G. Determinants of primary care specialty choice: A non-statistical meta-analysis of the literature. Acad Med. 1995;70:620–641.
8Lawson SR, Hoban JD, Mazmanian PE. Understanding primary care residency choices: A test of selected variables in the Bland–Meurer model. Acad Med. 2004;79(10 suppl):S36–S39.
9Hauer KE, Durning SJ, Kernan WN, et al. Factors associated with medical students' career choices regarding internal medicine. JAMA. 2008;300:1154–1164.
10Freed GL, Stockman JA. Oversimplifying primary care supply and shortages. JAMA. 2009;301:1920–1922.
11Senf JH, Campos-Outcalt D, Kutob RM. Lessons not learned from the generalist initiatives. Acad Med. 2002;77:774–775.
12Connelly MT, Sullivan AM, Peters AS, et al. Variation in predictors of primary care career choice by year and stage of training. J Gen Intern Med. 2003;18:159–169.
13McDonald FS, West CP, Popkave C, Kolars JC. Educational debt and reported career plans among internal medicine residents. Ann Intern Med. 2008;149:416–420.
14Kutob RM, Senf JH, Campos-Outcalt D. Declining interest in family medicine: Perspectives of department heads and faculty. Fam Med. 2003;35:504–509.
15Kahn MJ, Markert RJ, Lopez FA, Specter S, Randall H, Krane NK. Is medical student choice of a primary care residency influenced by debt? MedGenMed. October 24, 2006;8:18.
16Newton DA, Grayson MS. Trends in career choice by US medical school graduates. JAMA. 2003;290:1179–1182.
17Dorsey ER, Jarjoura D, Rutecki GW. Influence of controllable lifestyle on recent trends in specialty choice by US medical students. JAMA. 2003;290:1173–1178.
18Brotherton SE, Rockey PH, Etzel SI. US graduate medical education, 2004–2005: Trends in primary care specialties. JAMA. 2005;294:1075–1082.
19Babbott D, Baldwin DC, Jolly P, Williams DJ. The stability of early specialty preferences among US medical school graduates in 1983. JAMA. 1988;259:1970–1975.
20Senf JH, Campos-Outcalt D, Kutob R. Factors related to the choice of family medicine: A reassessment and literature review. J Am Board Fam Pract. 2003;16:502–512.
21Association of American Medical Colleges. Matriculating Student Questionnaire (MSQ): 2008: All Schools Summary Report. Available at: Accessed November 19, 2009.
22Compton MT, Frank E, Elon L, Carrera J. Changes in US medical students' specialty interests over the course of medical school. J Gen Intern Med. 2008;23:1095–1100.
23Fincher RM, Lewis LA. Influences on specialty choice: Are they different for switchers and non-switchers? Acad Med. 1999;74(10 suppl):S121–S123.
24Xu G, Hojat M, Brigham TP, Veloski JJ. Factors associated with changing levels of interest in primary care during medical school. Acad Med. 1999;74:1011–1015.
25Association of American Medical Colleges. Matriculating Student Questionnaire 2008. Available at: Accessed November 20, 2009.
26American Medical Association. Graduate career plan statistics. FRIEDA Online. Available at: Accessed November 19, 2009.
27Brotherton SE, Etzel SI. Graduate medical education, 2007–2008. Appendix II. JAMA. 2008;300:1228–1244.
28Rosenblatt RA, Andrilla HA. The impact of U.S. medical students' debt on their choice of primary care careers: An analysis of data from the 2002 Medical School Graduation Questionnaire. Acad Med. 2005;80:815–819.
29McAlister RP, Andriole DA, Brotherton SE, Jeffe DB. Are entering obstetrics/gynecology residents more similar to the entering primary care or surgery workforce? Am J Obstet Gynecol. 2007;197:536.e1–536.e6.
30Lautenschlager GJ. A comparison of alternatives to conducting Monte Carlo analyses for determining parallel analysis criteria. Multivariate Behav Res. 1989;24:365–395.
31Terrell C, Castillo-Page L, eds. Diversity in Medical Education Facts and Figures 2008. Washington, DC: Association of American Medical Colleges; 2008.
32Andriole DA, Whelan A, Jeffe DB. Characteristics and career intentions of the emerging MD–PhD workforce. JAMA. 2008;300:1165–1173.
33Dorsey ER, Jarjoura D, Rutecki GW. The influence of controllable lifestyle and sex on the specialty choices of graduating U.S. medical students, 1996–2003. Acad Med. 2005;80:791–796.
34Lambert EM, Holmboe ES. The relationship between specialty choice and gender of US medical students, 1990–2003. Acad Med. 2005;80:797–802.
35Brennan TA, Horwitz RI, Duffy FD, Cassel CK, Goode LD, Lipner RS. The role of physician specialty board certification status in the quality movement. JAMA. 2004;292:1038–1043.
36Sharp LK, Bashook PG, Lipsky MS, Horowitz SD, Miller SH. Specialty board certification and clinical outcomes: The missing link. Acad Med. 2002;77:534–542.
37Association of American Medical Colleges. Holistic Review: Aligning Admissions to Mission. Available at: Accessed November 20, 2009.
38Bodenheimer T, Bernson RA, Rudolf P. The primary care–specialty income gap: Why it matters. Ann Intern Med. 2007;146:895–896.
39Block SD, Clark-Chiarelli N, Singer JD. Mixed messages about primary care in the culture of U.S. medical schools. Acad Med. 1998;731:1087–1094.
40Ko M, Edelstein RA, Heslin KC, et al. Impact of the University of California, Los Angeles/Charles R. Drew University Medical Education Program on medical students' intentions to practice in underserved areas. Acad Med. 2005;80:803–808.
41Freeman J, Ferrer RL, Greiner KA. Viewpoint: Developing a physician workforce for America's disadvantaged. Acad Med. 2007;82:133–138.
42Senf JH, Campos-Outcalt D, Kutob R. Family medicine specialty choice and interest in research. Fam Med. 2005;37:265–270.
43Smart DR. Physician Characteristics and Distribution in the US: 2007 Edition. Chicago, Ill: American Medical Association; 2008.
44Campbell K, Horn A, Schoolcraft S, Youngclaus J, eds. AAMC Data Book: Medical Schools and Teaching Hospitals by the Numbers. Washington, DC: Association of American Medical Colleges; 2006.
45Newton DA, Grayson MS, Thompson LF. The variable influence of lifestyle and income on medical students' career specialty choices: Data from two U.S. medical schools, 1998–2004. Acad Med. 2005;80:809–814.
46Manuel RS, Borges NJ, Jones BJ. Person-oriented versus technique-oriented specialties: Early preferences and eventual choice. Med Educ Online [serial online]. 2009;14:4. Available at: Accessed November 20, 2009.
47National Resident Matching Program; Association of American Medical Colleges. Charting Outcomes in the Match. Available at: Accessed November 20, 2009.
© 2010 Association of American Medical Colleges