ARE MD-PHD PROGRAMS MEETING THEIR GOALS? AN ANALYSIS OF CAREER CHOICES MADE BY GRADUATES OF 24 MD-PHD PROGRAMS

Page created by Anne Webster
 
CONTINUE READING
Career Choice

Are MD–PhD Programs Meeting Their Goals?
An Analysis of Career Choices Made by
Graduates of 24 MD–PhD Programs
Lawrence F. Brass, MD, PhD, Myles H. Akabas, MD, PhD, Linda D. Burnley,
David M. Engman, MD, PhD, Clayton A. Wiley, MD, PhD, and Olaf S. Andersen, MD

Abstract
Purpose                                                 Results                                         appointments in nonclinical departments,
MD–PhD training programs provide an                     The average program enrolled 90 trainees,       increasing time to graduation, and
integrated approach for training                        required 8.0 years to complete, and had an      expanding residency choices that include
physician–scientists. The goal of this                  attrition rate of 10%. Nearly all (95%) of      disciplines historically associated with
study was to characterize the career path               those who graduated entered residencies.        clinical practice rather than research.
taken by MD–PhD program alumni                          Most (81%) were employed in academia,
during the past 40 years and identify                   research institutes, or industry; 16% were      Conclusions
trends that affect their success.                       in private practice. Of those in academia,      Most MD–PhD program graduates follow
                                                        82% were doing research and at least            career paths generally consistent with their
                                                        61% had identifiable research funding.          training as physician–scientists. However,
Method                                                  Whereas two-thirds devoted more than            the range of their professional options is
In 2007 and 2008, 24 programs                           50% effort to research, only 39% devoted        broad. Further thought should be given to
enrolling 43% of current trainees and                   more than 75% effort. Many with                 designing their training to anticipate their
representing half of the National                       laboratory-based PhDs reported doing            career choices and maximize their
Institutes of Health-funded MD–PhD                      clinical, as well as basic and translational,   likelihood of success as investigators.
training programs submitted anonymous                   research. Emerging trends include
data on 5,969 current and former                        decreasing numbers of graduates who
trainees.                                               forego residencies or hold primary              doi: 10.1097/ACM.0b013e3181d3ca17

The concept of the MD–PhD program                       academic medical centers, universities,         Sciences (NIGMS)4 and competitively
as an integrated approach to training                   and research institutes such as the             awarded individual fellowships that are
physician–scientists dates back to the                  National Institutes of Health (NIH). It         offered by some of the NIH institutes. In
late 1950s. Although a long tradition                   was also expected that links would exist        addition to assisting with program costs,
exists of physicians becoming                           between program graduates’ medical              NIH funding has helped standardize
investigators as well as clinicians, MD–                training, clinical activities, and research     training approaches and provides a
PhD programs were established with the                  interests and that each of these would          regular source of external review as
realization that the standard four-year                 inform the others in ways that could not        programs compete to obtain and
medical school curriculum is neither                    be experienced by scientists who were not       continue their MSTP status.
intended nor sufficient to train                        trained as physicians.1
physician–investigators who are as                                                                      Because MD–PhD programs commonly
proficient in the lab as they are in the                MD–PhD programs are not the only                provide full tuition and stipend support
clinic. As the term is now commonly                     available approach for training future          to their trainees, they collectively
applied, a physician–scientist or, more                 physician–scientists, but they have, in many    represent a large investment of
broadly, a physician–investigator, is a                 respects, become the most visible.2–4 In an     institutional, federal, and societal
physician who is committed to the quest                 MD–PhD curriculum, students complete            resources. Therefore, it is reasonable to
for new knowledge and new approaches                    the requirements for both the MD and            ask whether they are meeting their goals.
to disease diagnosis, treatment, and                    the PhD, usually by starting with the first     This was last done in the 1990s.4 –6 Recent
prevention, and who devotes far more of                 two years of medical education and then         debates at national conferences and in the
his or her time to these activities than to             focusing on graduate school before              literature about the best ways and times
routine clinical care. It was expected from             returning to complete the medical degree.       to train physicians to be investigators7 led
the start that most graduates of MD–PhD                 Although initially available at only a few      us to believe it is time to evaluate MD–
programs would be employed by                           universities, MD–PhD programs are now           PhD programs again. To our knowledge,
                                                        offered at most U.S. and a few Canadian         the present study, which analyzes the data
                                                        medical schools. These programs vary            provided to us by 24 MD–PhD programs
Please see the end of this article for information      considerably in size and resources. Since       in 2007 and early 2008, represents the
about the authors.
                                                        1964, NIH support has been available            most exhaustive look to date at the career
Correspondence should be addressed to Dr. Brass,        through institutional Medical Scientist         paths of MD–PhD program alumni. The
University of Pennsylvania, Room 915 BRB-II, 421
Curie Blvd., Philadelphia, PA 19104; telephone: (215)   Training Program (MSTP) grants from             participating programs collectively enroll
573-4669; e-mail: brass@mail.med.upenn.edu.             the National Institute of General Medical       approximately 40% of the estimated

Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17                                                                                               1
Career Choice

5,000 MD–PhD students currently in                      required to track their graduates and           Program characteristics and trainees
training and together represent nearly                  report on their activities every five years.    The 24 programs that participated in the
half of the 42 programs that were                       In other words, they were more likely           survey ranged in size in AY2008 from 31
receiving NIGMS MSTP grants at the                      than other programs to have the                 to 184 trainees; the average was 90
time the data were collected. The data                  necessary data on hand.                         trainees. Thirty-seven percent of trainees
that were provided allow us to address                                                                  in AY2008 were women, but this
issues raised in recent editorials calling              Each program submitted basic                    proportion varied considerably (range
for an evaluation of MD–PhD training                    information, including the number of            20%–60%). The unweighted average
programs,7,8 as well as concerns that                   trainees in AY2008, the number of               time to complete both degrees for
physician–investigators are a vanishing                 students entering the program in                graduates in AY1998 –2007 was 8.0 ⫾ 0.4
species.1                                               AY1998 –2007, the number of students            years (mean ⫾ 1 SD; 7.8 years when
                                                        resigning without one or both degrees           weighted by program size, range 7.2–8.5
                                                        from AY1998 to AY2007, the percentage           years). On average, 10% of the 2,562
Method                                                  of trainees in AY2008 who were women,           students who enrolled from AY1998 to
We sent a request for data to the directors             the average number of years to complete         AY2007 withdrew without completing
and administrators of 25 MD–PhD                         both degrees for graduates in AY1998 –          both degrees (range 3%–34%, 1 SD ⫽
programs, which we selected from among                  2007, and the discipline in which the PhD       7%). We found no relationship between
participants at the 2007 national                       was awarded.                                    program size and either the time to
conference of MD–PhD programs. We                                                                       degree or the attrition rate from the same
tried to achieve as much diversity as                   Program directors and administrators
                                                                                                        program (Figures 1A and 1B). Although
possible in program size and location.                  also provided information about alumni,
                                                                                                        we were not provided information about
This was not a prospective study; rather,               including all graduates from program
                                                                                                        students’ reasons for withdrawing, our
we asked the program leaders to provide                 inception when possible. Program leaders
                                                                                                        personal experience as program directors
as much data as possible from the                       obtained this information from a
                                                                                                        and administrators suggests that most of
inception of their program through the                  combination of alumni questionnaires,
                                                                                                        those who withdraw from MD–PhD
end of academic year (AY) 2007 (the 12                  alumni Web sites, and searches of public
                                                                                                        programs complete medical school. Fewer
months ending in June 2007). Current                    databases such as the NIH Computer
                                                                                                        complete graduate school alone or drop out
students were defined as those enrolled at              Retrieval of Information on Scientific
                                                                                                        completely. Despite a recent report that
the time that the survey was completed,                 Projects database. The identities of
                                                                                                        women are less likely than men to complete
which was during AY2008. Twenty-four                    individual trainees and alumni were not
                                                                                                        MD–PhD programs,9 we found no
programs agreed to participate*; one                    available to us, and therefore we were not
                                                                                                        relationship between the attrition rate from
program was unable to comply within                     able to obtain additional primary data
                                                                                                        individual programs and the percentage of
the designated time frame. Twenty of the                ourselves.
                                                                                                        trainees enrolled in the programs who were
participating programs were among the                                                                   women (Figure 1C).
42 programs receiving NIGMS MSTP                        Statistical analysis
grants. Our decision to overrepresent                   We calculated values in Figure 1 for r2,        Information on PhD disciplines was
NIGMS-funded programs arose in part                     the square of the Pearson correlation           provided for 1,957 trainees enrolled in
because they tend to be the larger                      coefficient, using functions built into         AY2008, including 306 first- and second-
programs and in part because they are                   Microsoft Excel 2008 (Microsoft                 year students who had not yet declared a
                                                        Corporation, Redmond, Washington).              discipline (a common practice in many of
                                                                                                        the programs). Of the 1,651 students who
* The 24 programs that participated are located at
the Albert Einstein College of Medicine of Yeshiva      Results                                         had declared a discipline, 1,422 (86%)
University; Baylor College of Medicine; Case Western                                                    were enrolled in PhD programs within
Reserve University School of Medicine; Emory            Twenty-four MD–PhD programs                     the broad arc of biomedical disciplines.
University School of Medicine; Harvard Medical          provided information about 5,969                Another 151 (9%) were in engineering.
School; Johns Hopkins University School of Medicine;    individuals, including 2,023 current
Medical College of Wisconsin; Northwestern                                                              The others were working in diverse
University Feinberg School of Medicine; University of
                                                        trainees, 1,143 recent graduates who were       disciplines including health policy,
California, San Diego, School of Medicine; University   still in residencies or postdoctoral            epidemiology, public health,
of California, San Francisco, School of Medicine;       fellowships, and 2,803 older alumni who
University of Colorado Denver School of Medicine;
                                                                                                        anthropology, sociology, chemistry,
University of Iowa Roy J. and Lucille A. Carver
                                                        had completed all phases of postgraduate        mathematics, philosophy, marine
College of Medicine; University of Maryland School      training. This represents 43% of the 4720       biology, population health, psychology,
of Medicine; University of Medicine and Dentistry of    trainees enrolled in 2008 (www.aamc.org/        and the history and sociology of science.
New Jersey–Robert Wood Johnson Medical School;          data/facts/enrollmentgraduate/start.htm)
University of Michigan Medical School; University
of Pennsylvania School of Medicine; University of       and may represent a similar proportion
Pittsburgh School of Medicine; University of            of alumni, although the total number of         Positions of program alumni
Rochester School of Medicine and Dentistry;             graduates is currently unknown. Because         Twenty-two programs provided
University of Texas Medical School at Houston;
University of Texas Southwestern Medical Center at
                                                        some of the programs were unable to             information about then-current or last
Dallas Southwestern Medical School; University of       supply answers to all of the survey             known positions of 2,413 alumni who
Wisconsin School of Medicine and Public Health;         questions, in the following presentation        had completed all phases of postgraduate
Vanderbilt University School of Medicine;               of the results we have indicated the            training. Our results show that 80% of
Washington University in St. Louis School of
Medicine; and Weill Cornell Medical College of          number of programs and/or alumni that           graduates were employed full-time in
Cornell University.                                     answered each question.                         academic centers (1,625, or 67%),

2                                                                                            Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17
Career Choice

Figure 1 Relationships among MD–PhD program size, time to degree, attrition rates, and the percentage of trainees who are women. (A) The
relationship between program size in academic year (AY) 2008 and time to degree for students entering in AY1998 –2007 using data from 23 MD–
PhD programs (r2 ⫽ 0.018). (B) The relationship between program size in AY2008 and attrition rate over the period from AY1998 to AY2007 using
data from 24 programs (r2 ⫽ 0.034). (C) The proportion of trainees in AY2008 who are women and the attrition rate over the period from 1998 to
2007 using data from 24 programs (r2 ⫽ 0.043).

research institutes such as the NIH (105,        to research: Nearly two-thirds (521, or                                                            Interestingly, even though the vast
or 4%), or in industry (189, or 8%), each        64%) reported committing at least half of                                                          majority of trainees completed their PhD
of which is an outcome consistent with           their time to research activities, but only                                                        studies in basic biomedical or
the goals of MD–PhD training. Of the             39% (317) devoted at least three-quarters                                                          engineering disciplines, 736 alumni of 14
remainder, 386 (16%) were in private             of their time; 19% (155) reported                                                                  programs gave a variety of answers to a
practice.                                        spending one-quarter or less (Figure 2A).                                                          question about the kinds of research that
                                                                                                                                                    they do. The choices provided were basic,
Because of concerns about possible                                                                                                                  translational, clinical, and health services.
ascertainment bias arising from the                                                                                                                 More than one answer was allowed, and
greater ease in tracking individuals                                                                                                                more than one was commonly given.
employed in universities and research
                                                 Table 1
                                                 Primary Department of MD–PhD                                                                       Although relatively few program
institutes, we also analyzed the data            Program Alumni in Academia*                                                                        graduates were engaged in health services
focusing solely on the 16 programs able                                                                                                             research at the time of the survey, nearly
to provide information on at least 98%                                                             MD–PhD program                                   equal numbers reported doing basic,
(n ⫽ 1,927) of their alumni. The                                                                            alumni
                                                                                                                                                    translational, and clinical research
                                                 Department                                                No. (%)
numbers we obtained are essentially                                                                                                                 (Figure 2B).
identical to those derived from the larger       Internal medicine                                                         427 (26.3)
                                                  ...............................................................................................
data set: 68% employed in academia, 4%           Pediatrics                                                                203 (12.5)               Information about research funding was
                                                  ...............................................................................................
in research institutes, 8% in industry, and      Pathology                                                                 192 (11.8)               available for 1,120 alumni of 17 programs
                                                  ...............................................................................................
16% in private practice.                         Neurology                                                                    153 (9.4)             who were in academia. Of these, at least
                                                  ...............................................................................................
                                                 Surgery                                                                      116 (7.2)             685 (61%) had funding; 154 (14%) said
Information about primary                         ...............................................................................................
                                                 Psychiatry                                                                     85 (5.2)            that they did not. Of the 281 (25%) for
appointments was provided to us for               ...............................................................................................   whom no data were available, nearly half
1,621 alumni of 22 programs working in           Ophthalmology                                                                  61 (3.8)
                                                  ...............................................................................................   (131, or 12%) reported research activities
academia (Table 1). Collectively, internal       Anesthesiology                                                                 50 (3.1)
                                                  ...............................................................................................   and may have had research support that
medicine, pediatrics, pathology, and             Radiology                                                                      43 (2.7)
                                                  ...............................................................................................   was not reported, which means that the
neurology accounted for 975 (60%) of             Dermatology                                                                    43 (2.7)            correct proportion of those in academia
the alumni in academia, but most alumni           ...............................................................................................
                                                 Radiation oncology                                                             24 (1.5)            who have research funding may be as
(1,428, or 88%) held primary                      ...............................................................................................
                                                 Obstetrics–gynecology                                                          19 (1.2)            much as 73% (61% plus 12%). Note that
appointments in clinical departments.             ...............................................................................................
                                                 Emergency medicine                                                                7 (0.4)          many of the individuals working at
Many also had secondary appointments              ...............................................................................................   institutes or in industry would also be
in basic science departments.                    Physical medicine and                                                             5 (0.3)
                                                 rehabilitation                                                                                     expected to be doing research, but they
                                                  ...............................................................................................
Research activities of program alumni            Nonclinical                                                               175 (10.8)
                                                                                                                                                    were not included in this analysis, so the
                                                  ...............................................................................................   overall percentage of alumni doing
At least 921 (82%) of 1,118 alumni (of 17        Unknown                                                                        18 (1.1)
                                                  ...............................................................................................   research may be even higher than
programs) in academia indicated that                  Total                                                         1,621 (100.0)                   suggested by our analysis of those in
they were doing research; 141 (13%)             * Summary of data provided by 22 MD–PhD programs                                                    academia.
reported that they were not, and no data          on 1,621 alumni, each of whom had completed
were available for 56 (5%). However,              postgraduate training and was employed full-time in
considerable variation existed in the             academia at the time that the survey data were                                                    Trends over time
                                                  collected in 2007–2008. “Surgery” includes all of
amount of time that 814 alumni of 16              the related disciplines. “Nonclinical” includes, but is                                           As noted above, our data show that most
programs were willing or able to devote           not limited to, basic science departments.                                                        MD–PhD graduates enter academia and

Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17                                                                                                                                            3
Career Choice

Figure 2 MD–PhD program graduates’ responses to questions about their recent research activities. (A) How much time do you devote to research?
Responses from 814 alumni of 16 programs who are now in academia. (B) Which kinds of research do you do? Data from 736 alumni of 14
programs; more than one answer was allowed.

have their primary appointment in a              For an analysis of changes over time, we                                                          There has also been a decline in the
clinical department. The choices that            divided the data on program alumni into                                                           number of graduates choosing
senior students make for their next step         arbitrary cohorts of approximately 10                                                             residencies in internal medicine,
after graduation strongly affect where           years based on graduation year. Several                                                           neurology, pathology, and pediatrics—
they are likely to end up and may affect         trends were evident. The percentage of                                                            disciplines that have historically provided
the likelihood that they will choose and         graduates choosing to forgo a residency                                                           a protected environment for the
succeed in a research-oriented career.           in favor of a postdoctoral fellowship,                                                            development of physician–scientist
Table 2 summarizes the choices made by           which has never been a common choice,                                                             careers—and an increase in those
939 recent graduates of 21 programs. Of          has declined to 4% to 5% (Figure 3A).                                                             selecting dermatology, ophthalmology,
these, 892 (95%) chose to continue their                                                                                                           radiation oncology, and surgery (Figure
clinical training. Only 47 (5%) elected to                                                                                                         3B). Others have noted similar trends in
do a postdoctoral fellowship without             Table 2                                                                                           residency choice by all medical school
postgraduate clinical training. Among            Residency Choice of Recent MD–PhD                                                                 graduates.9 –11
those opting for a residency, internal           Alumni Still in Training*
medicine was the most popular choice                                                                                                               Is choice of residency field predictive of
                                                                                                      MD–PhD alumni
(270, or 29%), and, collectively, internal                                                                in training
                                                                                                                                                   whether MD–PhD program graduates
medicine plus pediatrics, pathology, and         Department                                                   No. (%)                              will stay on track to become
neurology accounted for 518 (55%)                                                                                                                  investigators or enter private practice?
                                                 Internal medicine                                                        270 (28.8)
recent graduates. The next most popular          ...............................................................................................   In Table 3, we present the outcomes
choice was surgery (107, or 11%), a              Surgery                                                                  107 (11.4)               data for 1,862 alumni of 22 programs
                                                 ...............................................................................................
category that combines all of the surgery-       Pediatrics                                                                  98 (10.4)             who had completed all phases of
                                                 ...............................................................................................
related disciplines.                             Pathology                                                                     79 (8.4)
                                                 ...............................................................................................
                                                                                                                                                   postgraduate training, asking
                                                 Neurology                                                                     71 (7.6)            retrospectively whether those who
                                                 ...............................................................................................
Irrespective of their choice of residency,       Radiology                                                                     61 (6.5)            chose clinical training in a particular
                                                 ...............................................................................................
the vast majority of the alumni included         Psychiatry                                                                    54 (5.8)            field eventually ended up in private
in Table 2 completed their PhD training          ...............................................................................................   practice. There was considerable
                                                 Dermatology                                                                   51 (5.4)
in a biomedical laboratory discipline. It        ...............................................................................................   variability: 8% of pathology graduates
                                                 Radiation oncology                                                            32 (3.4)
is interesting to compare the residency          ...............................................................................................   ended up in private practice compared
choices of the whole group with the              Ophthalmology                                                                 33 (3.5)            with 62% of those who completed
                                                 ...............................................................................................
choices made by the small number (21)            Anesthesiology                                                                19 (2.0)            family medicine residencies. Overall,
                                                 ...............................................................................................
of recent graduates in the survey who did        Obstetrics–gynecology                                                         10 (1.1)            14% (165 of 1,220) of graduates who
                                                 ...............................................................................................
their PhD training in anthropology,              Emergency medicine                                                               7 (0.7)          chose residencies in internal medicine,
                                                 ...............................................................................................
demography, English, epidemiology,               Postdoctoral training                                                         47 (5.0)            neurology, pediatrics, or pathology
health policy, history of science, or public     without residency                                                                                 ended up in private practice, compared
                                                 ...............................................................................................
health. In this group, internal medicine             Total                                                             939 (100.0)                 with 36% (120 of 338) of graduates
was overwhelmingly the most popular                                                                                                                who chose dermatology,
                                                * Summary of data provided by 21 MD–PhD programs
choice, with 62% (13) choosing it                                                                                                                  ophthalmology, or surgery (three of the
                                                  on 939 recent graduates who were still in
compared with 29% (Table 2) in the total          postgraduate training at the time that the survey                                                four areas noted above as showing an
pool of recent graduates.                         data were collected in 2007–2008.                                                                increase in popularity).

4                                                                                                                             Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17
Career Choice

Figure 3 Emerging trends in choices made by MD–PhD program alumni and in the length of time required to graduate from MD–PhD programs. (A)
Trends in the choice to do a postdoctoral research fellowship and forego doing a residency. Analysis drawn from data on 3,172 alumni from 23
programs divided into cohorts: 1965–1978 (n ⫽ 144), 1979 –1988 (n ⫽ 551), 1989 –1998 (n ⫽ 1,160), and 1999 –2007 (1,317). (B) Trends in
choosing a residency in internal medicine, neurology, pathology, or pediatrics compared with choosing a residency in dermatology, ophthalmology,
radiation oncology, or surgery. Analysis drawn from data on 3,172 alumni from 23 programs divided into cohorts as shown in (A). (C) Trends in
having a primary appointment in a basic science department. Analysis drawn from data provided by 22 programs divided into cohorts: 1965–1978
(n ⫽ 97), 1979 –1988 (n ⫽ 383), 1989 –1998 (n ⫽ 820), and 1999 –2007 (n ⫽ 306). (D) Trends in choosing a career in academia, a research institute,
industry, or private practice. Analysis drawn from data provided by 22 programs divided into cohorts: 1965–1978 (n ⫽ 163), 1979 –1988 (n ⫽ 601),
1989 –1998 (n ⫽ 1,198), and 1999 –2007 (n ⫽ 306). (E) Average time to graduation. Trainees who graduated between 1998 and 2007 from the
programs included in the present study required an average of 7.8 years (weighted) to complete both degrees. The data from studies completed in
1980, 1985, and 1995 are from the National Institute of General Medical Sciences.4

Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17                                                                                             5
Career Choice

                                                                                                                                                                                                             data. Nonetheless, a number of the
 Table 3                                                                                                                                                                                                     results that we obtained are in good
 Residency Choice as a Predictor of Eventually Choosing Private Practice*                                                                                                                                    agreement with those from the earlier
                                                                                                                                                                                                             studies cited above. Given the size of the
                                                                                                                  MD–PhD program alumni
                                                                                                                                                                                                             sample, our conclusions are likely to be
                                                                                         Total                         No. in private                                    % in private                        applicable to MD–PhD programs as a
 Department                                                                                no.                               practice                                        practice
                                                                                                                                                                                                             whole and especially to those programs
 Family medicine                                                                                13                                                     8                                            62       with NIGMS MSTP funding, nearly half
 .........................................................................................................................................................................................................
 Emergency medicine                                                                             13                                                     6                                            46       (20 of 42) of which were included in our
 .........................................................................................................................................................................................................
 Dermatology                                                                                    70                                                  31                                              44       survey. To what extent this heavy reliance
 .........................................................................................................................................................................................................
 Ophthalmology                                                                                101                                                   44                                              44       on data from programs with MSTP
 .........................................................................................................................................................................................................   grants has affected the results will remain
 Radiology                                                                                      69                                                  28                                              41
 .........................................................................................................................................................................................................   unclear until a fully inclusive MD–PhD
 Obstetrics–gynecology                                                                          26                                                     8                                            31
 .........................................................................................................................................................................................................   program graduates database—such as the
 Surgery                                                                                      167                                                   45                                              27       one under consideration by the
 .........................................................................................................................................................................................................
 Anesthesiology                                                                                 57                                                  13                                              23       Association of American Medical
 .........................................................................................................................................................................................................
 Physical medicine and rehabilitation                                                              5                                                   1                                            20       Colleges (AAMC)—is launched. Until
 .........................................................................................................................................................................................................
 Internal medicine                                                                            578                                                   93                                              16       then, important questions about
 .........................................................................................................................................................................................................
 Radiation oncology                                                                             27                                                     4                                            15       comparative outcomes between programs
 .........................................................................................................................................................................................................
 Neurology                                                                                    173                                                   23                                              13
                                                                                                                                                                                                             that receive MSTP support and those that
 .........................................................................................................................................................................................................   do not will have to wait.
 Pediatrics                                                                                   243                                                   32                                              13
 .........................................................................................................................................................................................................
 Psychiatry                                                                                     94                                                  11                                              12       Nevertheless, we can draw several clear
 .........................................................................................................................................................................................................
 Pathology                                                                                    226                                                   17                                                 8     conclusions from the data that are
* Summary of data provided by 22 MD–PhD programs on 1,862 alumni, each of whom had completed                                                                                                                 available. The first is that most MD–PhD
  postgraduate training at the time of this study in 2007–2008. Note that overall, 16% of the alumni who had                                                                                                 program graduates (80%) are in careers
  completed training were in private practice at the time that the survey data were collected. The columns indicate                                                                                          generally consistent with the goal of MD–
  the total number of alumni who completed a residency in each department and the number of those who
  subsequently entered private practice. The percentage for each department is the number who entered private                                                                                                PhD training, which is to train physicians
  practice divided by the total number who completed a residency in that department.                                                                                                                         who are committed to the quest for new
                                                                                                                                                                                                             knowledge and new approaches to
                                                                                                                                                                                                             disease diagnosis, prevention, and
 An additional trend that emerged shows a                                                                 fraction of total medical school                                                                   treatment. This number is essentially the
 steady decline in the proportion of                                                                      graduates. In 2007, there were only 1,721                                                          same as the 81% of 2000 –2006 MD–PhD
 graduates working in academia whose                                                                      applicants to MD–PhD programs, which                                                               graduates who reported in the AAMC
 primary appointment is in a basic science                                                                represents about 4% of total medical                                                               graduation survey that they planned
 department (Figure 3C). Overall,                                                                         school candidates. Of these 1,721                                                                  substantial career involvement in
 however, the proportion of graduates                                                                     individuals, 536 matriculated.12                                                                   research.9 Approximately 67% are in
 choosing careers in academia, research                                                                                                                                                                      academia, a number that is also
 institutes, or industry has changed little                                                               As program numbers increased, some                                                                 remarkably similar to what has been
 over almost 50 years (Figure 3D).                                                                        schools and groups of schools made                                                                 reported previously.4,6,13,14
                                                                                                          efforts to analyze the careers of their
 Finally, we asked whether there has been                                                                 graduates.5,6,13,14 In 1998, NIGMS made                                                            A second conclusion is that most (82%)
 a change in the time required to complete                                                                available some of the outcomes data that                                                           of the program graduates at academic
 both degrees over the years that MD–                                                                     had been reported by MSTP-funded                                                                   medical centers are doing research and
 PhD programs have been in existence. As                                                                  institutions as part of their training grant                                                       have funding to support their efforts.
 noted above, the unweighted average                                                                      renewals.4 The results of these studies                                                            Presumably at least as many of the
 time for graduates in AY1998 –2007 was                                                                   showed, much as we have found, that the                                                            graduates at research institutes such as
 nearly 8 years, which represents a                                                                       majority of MD–PhD graduates are in                                                                the NIH are doing research as well. The
 substantial increase from the 6.6 years                                                                  academia. However, none of the previous                                                            data, however, also clearly show that the
 noted in 19804 (Figure 3E).                                                                              studies are recent, and none of them                                                               range of career choices among MD–PhD
                                                                                                          address recent trends. Although an all-                                                            program graduates is very broad with
                                                                                                          encompassing prospective database about                                                            respect to both research interests and
 Discussion                                                                                               MD–PhD program graduates would be                                                                  time spent on research. This diversity is
 MD–PhD programs have expanded                                                                            useful, none currently exists.                                                                     not currently reflected in the curriculum
 greatly in size and have become nearly                                                                                                                                                                      design and admissions policies of MD–
 ubiquitous since their beginnings about                                                                  In an effort to obtain a snapshot within a                                                         PhD programs, which typically focus on
 50 years ago at a handful of medical                                                                     reasonable length of time, we asked the                                                            the recruitment and training of bench
 schools. However, the number of                                                                          24 participating programs to provide                                                               scientists. The unstated assumption is
 applicants to these programs each year                                                                   information that they had already                                                                  that if MD–PhD graduates start in
 remains a small fraction of those applying                                                               collected. As a result, not all programs                                                           laboratory-based research, that is what
 to medical school, and the number of                                                                     could answer all questions about all                                                               they will continue to do—an assumption
 MD–PhD graduates remains a small                                                                         alumni, a recognized limitation of the                                                             that is challenged by the number of

 6                                                                                                                                                                                      Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17
Career Choice

graduates who have decided to do            unsurprising given the challenges of            Alternative paths to research careers
translational and clinical research. The    predicting what a 21-year-old applicant         If MD–PhD programs are to be viewed as
skill set needed to conduct well-designed   will actually do when he or she is 30 to 40     an experiment in training physician–
research in humans is simply not            years old. A retrospective analysis does        investigators, then what is the
addressed in most graduate school           suggest that some choices of a clinical         appropriate comparison group? Most
training programs.                          field for residency are more likely than        medical students are not planning
                                            others to lead to a career in private           research careers, so using all medical
Although there are no standard criteria     practice (Table 3), but given the manner        students as a control group is not
to identify a physician–scientist, given    of data collection, we cannot determine         especially helpful. In considering a
the complexity of modern research           whether graduates chose to enter private        smaller group, the subset of physicians
many of us advise our students and          practice as a consequence of their              who apply for NIH grants, Dickler and
prospective students that they will need    residency choice or whether the choice of       colleagues18 reported that MDs who
to spend 75% to 80% of their time on        a residency was a consequence of a              become investigators are less successful
research-related activities, leaving the    decision not to pursue a career in              than MD–PhDs and PhDs in obtaining a
remaining time for clinical care,           research. We note that graduates of three       first NIH research project grant (R01)
teaching, and other activities that will    of the fields that have shown recent gains      and, if funded once, are less likely to
hopefully integrate well with their         in popularity (dermatology,                     receive a subsequent R01. They are also
research interests. However, our study      ophthalmology, and surgery) had higher-         more likely to do clinical rather than
data show that if the designation           than-average rates of going into private        basic research. There has also been a
physician–scientist is limited to those     practice. But it is equally important to        decrease in MDs serving on NIH study
who spend at least 75% of their time on     note that at least some of the alumni who       sections.19
research-related activities, then many      chose those fields eschewed private
MD–PhD graduates fit this definition,       practice and reported performing funded         All of this makes perfect sense if
but many do not (Figure 2A).                research.                                       successful MD–PhD program candidates
                                                                                            are viewed as individuals who have gone
A third conclusion is that recent           A final conclusion from the data is that        through a rigorous vetting process that
concerns that the dropout rate from                                                         emphasizes early research experience and
                                            MD–PhD program graduates pursue a
MD–PhD programs is very high and                                                            commitment as well as academic
                                            broad range of types of research, perhaps
that many of those who stay in the                                                          excellence. It does not mean that
                                            reflecting their training in both science
program until the end enter private                                                         attending an MD–PhD program is the
                                            and medicine as well as their original
practice are not justified by the data.                                                     only way to become a physician–scientist.
                                            goals when choosing to apply to MD–
We found that most MD–PhD students                                                          From our experience as program
                                            PhD programs. Our survey shows that
complete the program, and most                                                              directors, we have found that MD–PhD
                                            instead of becoming basic scientists with
alumni are not in private practice. The                                                     programs are particularly well suited to
                                            only a distant memory of their medical
average attrition rate of students who                                                      individuals who decide early enough in
entered programs in AY1998 –2007 was        training, many MD–PhD program                   their college careers that they have a
10%, very similar to the 12% reported       graduates are conducting translational          commitment to discovery in medicine
by Fang and Meyer15 for MSTP-funded         and patient-oriented research as well as        and are fortunate enough to receive
trainees who matriculated between           basic research. In keeping with this trend,     guidance that includes information about
1980 and 1988, but considerably lower       discussions at the annual meetings of           physician–scientist training programs.
than the 29% that was reported in 2008      program directors and administrators            There will always be a need for alternative
by Andriole and colleagues.9 The            suggest that many MD–PhD programs               pathways to capture the “late bloomers”
attrition rate varied from 3% to 34%        are paying increased attention to the need      because there are not currently (and
among the schools in this study, a          for their students to be trained (or at least   likely never will be) enough MD–PhD
variation that deserves closer scrutiny     exposed) to the skills needed for               program graduates to maintain the ranks
to establish cause. Our study does not      translational and clinical research. Our        of active physician–scientists. However,
directly address gender differences         survey did not address the question of          in general, avoiding an MD–PhD
regarding attrition—whether women           whether MD–PhD trainees receive the             program is not a strategy that shortens
are less likely than men to persist in a    formal training in the design and               the time to an independent career, even
physician–scientist career.9,16,17 Gender   implementation of human studies that            taking into account the rising time to
was not included in the information         is increasingly being viewed as essential       graduation that we noted for such
supplied on each trainee. At the            for clinical investigators. Although the        programs. The average age at first R01
program level, however, we did not find     training they receive in the scientific         was the same (43 years) for both MDs
a correlation between the percentage of     method will be helpful, if MD–PhD               and MD–PhDs in 2007.20 This suggests
trainees who are women and the              graduates are to be successful in this          that if one intends to be a
attrition rate from the program.            realm as well as in the laboratory,             physician–scientist at the time of entry to
                                            additional thought should be given to           medical school, skipping graduate school
Overall, we found that 16% of the MD–       what they will need to know and when            will not save time—presumably because
PhD alumni who completed                    they should learn it. Timing is                 an independent research career requires
postgraduate training eventually entered    especially important given the need to          an extended period of mentored research
private practice, a number that is larger   resist making a lengthy training                training, whether it is completed during
than one might hope but is perhaps          program even longer.                            medical school or after residency.

Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17                                                                                 7
Career Choice

Research careers require research              Zemlo and colleagues.19 Nonetheless, a         requires continuing attention by
training, which is not usually part of the     recent survey of clinical department           individual program leaders, and efforts
medical school curriculum. One                 chairs still reported a large number of        should be directed toward both the
alternative path is offered by the Howard      vacant positions for physician–                selection and the nurturing of trainees.
Hughes Medical Institute-sponsored NIH         investigators able to do clinical
Cloister Program, which participants           research.22                                    Finally, the time required to complete an
typically complete after their third year of                                                  MD–PhD program is increasing, a trend
medical school. Fang and Meyer15 found         The distribution among clinical fields         that bodes ill for reversing the ever-
that participants in the Cloister Program      chosen by MD–PhD program graduates             increasing age at first faculty
were more likely to hold research-             has also gradually changed over the past       appointment and first R01.2 The average
oriented faculty positions at medical          50 years. Like Andriole and colleagues,9       time to graduation has risen from 6.6
schools than were unsuccessful applicants      we found that the proportion choosing          years4 to 7.8 years (weighted by program
to the program, but they were less likely      internal medicine, neurology, pathology,       size) since the early years of MD–PhD
to hold such positions than were the           and pediatrics has declined, whereas           training programs, which is an 18%
graduates of MSTP-funded programs.             those choosing fields such as                  increase (Figure 3E). This trend may in
Other alternative pathways include             dermatology, ophthalmology, radiation          part reflect the increasing demands of
obtaining mentored research training or        oncology, and surgery has increased            medical education. If so, some thought
even attending graduate school after           (Figure 3B). In other words, more of the       should be given to the education
completing postgraduate clinical training.     recent graduates of MD–PhD programs            requirements of physician–scientists
Our anecdotal impression is that fewer         are choosing clinical training in              versus those who will become full-time
are choosing these last two alternatives, at   disciplines outside those that have            clinicians. Some of the upward trend may
least as a means to move toward an             historically been the most willing to          also reflect increasing training
independent laboratory-based research          provide the large amounts of protected         requirements for PhD students25 and a
career. Currently, most of those whose         time required to do meaningful research.       diminished willingness to cross-count
goal is to focus exclusively on patient-       To the extent that residency choice is a       credits for work done toward each
oriented rather than basic research obtain     predictor of which department MD–PhD           degree. Data to substantiate these
the required clinical research training        program graduates eventually join, a           possibilities were not obtained for this
postresidency, usually in programs that        review performed 20 years from now is          study but need to be collected in the near
lead to a master’s degree rather than a        likely to find that a far broader range of     future. Because graduates of MD–PhD
PhD.                                           clinical departments have become the           program typically complete six or more
                                               “home” for MD–PhD program alumni. If           years of postgraduate training, they have
Trends over time                               so, then the critical question is, what will   a long additional training period before
We observed several notable trends in the      they be doing in those departments?            their first faculty appointment. If an
data on program alumni. The proportion                                                        unacceptable total duration of training is
of graduates that choose to forego             Although the reasons underlying this           to be avoided, more attention must be
residency training has always been             shift can be debated, the change may           paid to the requirements at each phase of
relatively small. It seems to be declining     result in either a continued positive          training that contribute to the whole. Too
further, as is the proportion of alumni        outcome (i.e., a wonderful opportunity         often, each step on this path to
working in academia with their primary         to extend inquiry into new fields) or a        independence is overseen by a different
appointment in a nonclinical department        very undesirable outcome (i.e., more           organization or certification group, and
(Figure 3C). The decline in primary basic      investigators leaking out of the pipeline).    these groups rarely communicate with
science appointments occurred during a         It is too soon to tell, but, as we already     each other. For this reason and others, it
period that saw a large growth in both the     noted, the data from this retrospective        is easy to argue for much better vertical
faculty size21 and the research portfolios     study clearly show that graduates of some      integration in the training of physician–
of clinical departments. One might             residency fields have been far more likely     scientists. A call to make this and other
speculate that clinical department chairs      than others to eventually become private       changes is part of a recent report from
view MD–PhD investigators as a safer bet       practitioners (Table 3). Private practice      the Association of Professors of
than investigators with a PhD but no           fits nobody’s definition of a desirable        Medicine.26
MD—individuals with an MD–PhD can              outcome for MD–PhD program alumni.
potentially generate clinical revenues;        In 2007, Ahn and colleagues23 reported         Limitations of this study
those with a PhD cannot. The increase in       on the attitudes and career intentions of      This study is the largest of its kind to
primary appointments in clinical               current trainees; Andriole and                 date. It is not, however, without
departments may also speak to the              colleagues’9 similar survey followed.          limitations arising from the method used
commitment of MD–PhD graduates to              These studies differ from ours in their        to collect the data, and we would like to
pursue disease-related research that finds     focus on career intentions rather than         point some of those out. The study
a more comfortable home in clinical            actual career choices, but it is worrisome     includes data from only 24 programs,
departments. This would be consistent          that, if true, their survey data raise         albeit ones that included 40% of the MD–
with the high percentage of alumni who         questions about some trainees’                 PhD trainees in the United States at the
report involvement in translational and        commitment to research careers.9,23,24 If      time of the survey. The programs vary
clinical research (Figure 2B). A shift of      MD–PhD programs continue to expand,            widely in size and location, but 20 of the
physician–scientists to clinical               admission of candidates who lack a             participating programs had NIGMS
departments was also noted in 2000 by          strong commitment is clearly an area that      MSTP grants at the time of data

8                                                                                 Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17
Career Choice

collection, and, as already noted, results     Dr. Brass is professor, Department of Medicine and          nigms.nih.gov/reports/mstpstudy/#10.
from the MSTP-supported programs may           Department of Pharmacology, University of                   Accessed November 25, 2009.
                                               Pennsylvania School of Medicine, Philadelphia,         5    Martin JB. Training physician–scientists for
prove to be different in meaningful ways                                                                   the 1990s. Acad Med. 1991;66:123–129.
                                               Pennsylvania, where he directs the MD–PhD
from the large number of programs that         program. He was 2007 chair, MD–PhD Section,            6    Bradford WD, Anthony D, Chu CT, Pizzo
have not yet benefited from NIH support.       Association of American Medical Colleges Group on           SV. Career characteristics of graduates of a
These differences may prove to be quite        Graduate Research, Education, and Training (AAMC            Medical Scientist Training Program, 1970 –
                                               GREAT Group).                                               1990. Acad Med. 1996;71:484 –487.
important when assessing the
                                                                                                      7    Rosenberg LE. MD/PhD programs—A call
performance of individual programs, but        Dr. Akabas is professor, Department of Physiology
                                                                                                           for an accounting. JAMA. 2008;300:1208 –
                                               and Biophysics, Albert Einstein College of Medicine,
they may not change the aggregate                                                                          1209.
                                               Yeshiva University, Bronx, New York, where he
analysis by much. MSTP-supported               directs the MD–PhD program. He is co-chair, Data       8    Whitcomb ME. The need to restructure MD–
programs tend to be larger on average          and Analysis Committee, MD–PhD Section, AAMC                PhD training. Acad Med. 2007;82:623–624.
than those without MSTP support. The           GREAT Group.                                           9    Andriole DA, Whelan AJ, Jeffe DB.
                                                                                                           Characteristics and career intentions of the
four smallest programs in this study           Ms. Burnley is director, Administration and                 emerging MD/PhD workforce. JAMA. 2008;
enrolled only 31 to 34 trainees and had 36     Finance, Harvard–Massachusetts Institute of                 300:1165–1173.
                                               Technology Combined MD–PhD Program, Harvard
to 71 alumni at the time of the study. In      Medical School, Boston, Massachusetts. She is co-
                                                                                                      10   Dorsey ER, Jarjoura D, Rutecki GW.
contrast, even the four smallest of the NIH-                                                               Influence of controllable lifestyle on recent
                                               chair, Data and Analysis Committee, MD–PhD
supported programs had more trainees           Section, AAMC GREAT Group.                                  trends in specialty choice by US medical
                                                                                                           students. JAMA. 2003;290:1173–1178.
(52–66) and more alumni (60 –133).             Dr. Engman is professor, Department of                 11   Newton DA, Grayson MS. Trends in career
                                               Pathology, Northwestern University, Feinberg School         choice by US medical school graduates.
It is also worth emphasizing that the          of Medicine, Chicago, Illinois, where he directs the        JAMA. 2003;290:1179 –1182.
                                               MD–PhD program. He was 2008 chair, MD–PhD              12   Garrison G. AAMC Data and MD–PhD
data on MD–PhD alumni were                     Section, AAMC GREAT Group.                                  Students. Available at: www.aamc.org/
provided to us by the programs and not                                                                     members/great/mdphd/presentations/
                                               Dr. Wiley is professor, Department of Pathology,
directly by the alumni. Some programs                                                                      garrisonhandout.pdf. Accessed December 13,
                                               University of Pittsburgh School of Medicine,
did not include all alumni, and because        Pittsburgh, Pennsylvania, where he directs the MD–          2009.
this was not a prospective survey, not         PhD program. He was 2009 chair, MD–PhD Section,        13   Frieden C, Fox BJ. Career choices of
every question was asked by every              AAMC GREAT Group.                                           graduates from Washington University’s
                                                                                                           Medical Scientist Training Program. Acad
program when they last surveyed their          Dr. Andersen is professor, Department of                    Med. 1991;66:162–164.
alumni. However, it is somewhat                Physiology and Biophysics, Weill Cornell Medical
                                                                                                      14   McClellan DA, Talalay P. M.D.–Ph.D.
                                               College, Cornell University, New York, New York,
reassuring that essentially identical          where he directs the MD–PhD program. He was
                                                                                                           training at the Johns Hopkins University
results were obtained when our analysis        2005 chair, MD–PhD Section, AAMC GREAT Group.
                                                                                                           School of Medicine, 1962–1991. Acad Med.
of outcomes was limited to just those                                                                      1992;67:36 –41.
                                               Acknowledgments: The authors thank the MD–             15   Fang D, Meyer RE. Effect of two Howard
programs that submitted data on more                                                                       Hughes Medical Institute research training
                                               PhD program directors and administrators who
than 98% of their alumni.                      provided data for this survey.                              programs for medical students on the
                                                                                                           likelihood of pursuing research careers. Acad
                                               Funding/Support: None.                                      Med. 2003;78:1271–1280.
Conclusions                                                                                           16   Watt CD, Greeley SA, Shea JA, Ahn J.
                                               Other disclosures: None.                                    Educational views and attitudes, and career
In summary, this study shows that many                                                                     goals of MD–PhD students at the University
                                               Ethical approval: Not applicable.
MD–PhD program graduates are staying                                                                       of Pennsylvania School of Medicine. Acad
on the physician–investigator career           Disclaimer: This study was performed with the               Med. 2005;80:193–198.
track. It provides an interim response to a    support of the leadership of the National              17   Andrews NC. The other physician–scientist
                                               Association of MD–PhD Programs and the                      problem: Where have all the young girls
recent call7 for evidence that MD–PhD                                                                      gone? Nat Med. 2002;8:439 –441.
programs accomplish their mission:             Association of American Medical Colleges Group
                                               on Graduate Research, Education, and Training          18   Dickler HB, Fang D, Heinig SJ, Johnson E,
Clearly, they can and do. However,             (GREAT Group) Section on MD–PhD Programs.                   Korn D. New physician–investigators
looking beyond the very positive                                                                           receiving National Institutes of Health
                                               However, the views expressed are those of the
conclusions driven by the aggregate                                                                        research project grants: A historical
                                               authors alone and do not necessarily reflect those
                                                                                                           perspective on the “endangered species.”
analysis, the data collected in this survey    of either organization.
                                                                                                           JAMA. 2007;297:2496 –501.
also show that the range of eventual                                                                  19   Zemlo TR, Garrison HH, Partridge NC, Ley TJ.
professional “phenotypes” of program                                                                       The physician–scientist: Career issues and
                                               References                                                  challenges at the year 2000. FASEB J. 2000;14:221–
graduates is very broad. It is therefore
                                                1 Goldstein JL, Brown MS. The clinical                     230.
entirely legitimate to ask whether all of
                                                  investigator: Bewitched, bothered, and              20   National Institutes of Health, Office of
these phenotypes are consistent with the          bewildered— but still beloved. J Clin Invest.            Extramural Research. Average Age of
currently espoused goals of MD–PhD                1997;99:2803–2812.                                       Principal Investigators. Available at: http://
training. Recognition of this point will        2 Ley TJ, Rosenberg LE. The physician–scientist            report.nih.gov/NIH_Investment/PDF_
hopefully prompt further debate.                  career pipeline in 2005: Build it, and they will         sectionwise/NIH_Extramural_DataBook_
                                                  come. JAMA. 2005;294:1343–1351.                          PDF/NEDB_SPECIAL_TOPIC-AVERAGE_
Whatever the outcome of those much-                                                                        AGE.pdf. Accessed January 5, 2010.
                                                3 Rosenberg LE. The physician–scientist: An
needed debates, continued attention will          essential—and fragile—link in the medical           21   Association of American Medical Colleges.
need to be paid to the selection of the           research chain. J Clin Invest. 1999;103:1621–            U.S. faculty roster reports. Available at:
most appropriate candidates, the                  1626.                                                    www.aamc.org/data/facultyroster/reports.htm.
mentoring and training of matriculants,         4 National Institute of General Medical                    Accessed December 13, 2009.
                                                  Sciences. MSTP study: The careers and               22   Fang D, Dickler H, Heinig SJ, Korn D.
and the care of program graduates as they         professional activities of graduates of the              Recruitment of New Physician Investigators in
traverse the challenging divide between           NIGMS Medical Scientist Training                         Clinical Research: Findings From a Survey of
graduation and independence.                      Program. Available at: http://publications.              Clinical Department Chairs at U.S. Medical

Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17                                                                                                      9
Career Choice

   Schools. Available at: https://services.aamc.org/      Attitudes, Goals, and Education) survey.        Va: National Science Foundation, Division
   publications/index.cfm?fuseaction⫽Product.             Acad Med. 2007;82:633–645.                      of Science Resources Statistics; 2006.
   displayForm&prd_id⫽211&cfid⫽1&cftoken⫽              24 Ahn J, Watt CD, Greeley SA, Bernstein J.     26 Association of Professors of Medicine.
   8A91BEF8-B9BB-BF89-888382FD9D06C5DD.                   MD–PhD students in a major training             Recommendations for Revitalizing the
   Accessed December 13, 2009.                            program show strong interest in becoming        Nation’s Physician–Scientist Workforce.
23 Ahn J, Watt CD, Man LX, Greeley SA, Shea               surgeon–scientists. Clin Orthop Relat Res.      Available at: www.im.org/PolicyAnd
   JA. Educating future leaders of medical                August 2004;258 –263.                           Advocacy/PolicyIssues/Research/PSI/
   research: Analysis of student opinions and          25 Thurgood L, Golladay MJ, Hill ST. U.S.          Pages/default.aspx. Accessed December 13,
   goals from the MD–PhD SAGE (Students’                  Doctorates in the 20th Century. Arlington,      2009.

10                                                                                          Academic Medicine, doi: 10.1097/ACM.0b013e3181d3ca17
You can also read