Background For 2 decades, the number of physician-scientists has not kept pace with the overall growth of the medical research community. Concomitantly, the number of women entering medical schools has increased markedly. We have explored the effect of the changing gender composition of medical schools on the present and future pipeline of young physician-scientists.
Methods We analyzed data obtained from the Association of American Medical Colleges, the National Institutes of Health, and the Howard Hughes Medical Institute pertaining to the expressed research intentions or research participation of male and female medical students in the United States.
Results A statistically significant decline in the percentage of matriculating and graduating medical students\Mboth men and women\Mwho expressed strong research career intentions occurred during the decade between 1987 and 1997. Moreover, matriculating and graduating women were significantly less likely than men to indicate strong research career intentions. Each of these trends has been observed for medical schools overall and for research-intensive ones. Cohort data obtained by tracking individuals from matriculation to graduation revealed that women who expressed strong research career intentions upon matriculation were more likely than men to decrease their research career intentions during medical school. Medical student participation in research supported the gender gap identified by assessing research intentions. Female medical student participation in the Medical Scientist Training Program and the Howard Hughes Medical Institute/National Institutes of Health-sponsored Cloisters Program has increased but lags far behind the growth in the female population in medical schools.
Conclusion Three worrisome trends in the research career intentions and participation of the nation\'s medical students (a decade-long decline for both men and women, a large and persistent gender gap, and a negative effect of the medical school experience for women) presage a further decline in the physician-scientist pipeline unless they are reversed promptly and decisively.
- medical students
- physician population
- research career intentions
- sex differences
Statistics from Altmetric.com
During the past 20 years, the physician-scientist has become a threatened or endangered species in the United States.1-4The number of U.S. physicians who reported research as their primary activity decreased by 6% between 1980 and 1997.4Between 1978 and 1998, there was no net growth in the number of first-time research grant applications submitted by MDs,4and from 1994 to 1998, the number of first-time National Institutes of Health (NIH) grant applications submitted by MDs actually declined by 25%.3,4At the same time, the average age of NIH-funded physician-scientists has increased steadily. In 1977, 44% of all NIH principal investigators with medical degrees were older than 45 years of age, and this percentage increased to 56% by 1997.4Such evidence indicates that the overall decline in physician-scientists is not due to increased attrition of established physician-scientists but rather represents a decline in MDs taking up research careers.5
The future decrease in the number of physician-scientists has been foreshadowed during their medical school experience. Between 1987 and 1996, the percentage of matriculating medical students with strong research intentions decreased from 13.7 to 10.5%.6During the same period, the percentage of graduating medical students with strong research intentions decreased from 14.7 to 11.7%.6
Perhaps the most significant change in the medical profession during the past 30 years has been the increasing proportion of physicians who are women. Between 1970 and 1998, the number of female first-year entrants to U.S. medical schools increased dramatically from 1,228 (11.0% of all students) to 7,162 (44.3%).7Previous studies have suggested that women are less interested in research careers than men are, citing different career motivations, humanitarian and/or idealistic considerations, and a preference for patient care as explanations.8,9Another reason why women may tend to be less interested in a research career than men are is the institutional barriers that hinder their progress in research careers. Women who pursue medical research careers tend to receive fewer resources for research and lower salaries and are promoted more slowly than men, despite having achieved a similar level of academic training.10,11
These factors might cause women to drop out of research before, during, or after attending medical school. If this hypothesis is true, a picture of the medical profession emerges in which men are increasingly being replaced by women who tend to be less interested in research careers and are more likely to drop out of the physician-scientist training path. To test this hypothesis, we assessed women's research career intentions and participation during medical school and compared these findings with those for men.
Data related to medical student research intentions were obtained from a medical student database maintained by the Association of American Medical Colleges (AAMC). Each year, the AAMC administers a nationwide questionnaire to both matriculating and graduating medical students to obtain a demographic profile of the medical school classes and to assess their career intentions. Both the Matriculating Student Questionnaire (MSQ) and the Graduating Student Questionnaire (GSQ) include the following question that is asked to assess the career intentions of medical students with regard to research involvement:
How extensively do you expect to be involved in research during your medical career?
Significantly involved (several years set aside for full-time research or 25% or more of continuous career devoted to research pursuits)
Somewhat involved (1 year or less set aside for research or less than 25% of continuous career)
Involved in a limited way (e.g., occasional cooperation with clinical trials of new drugs or medical devices in my practice)
Of the five choices, we characterized the replies “exclusively” and “significantly involved” as strong research career intentions, “somewhat involved” as moderate research career intentions, and “involved in a limited way” and “not involved” as weak research career intentions. We obtained MSQ and GSQ data sorted by sex for this question for the interval between 1987 and 1996. (We focused on this decade because the GSQ did not include this question after 1996, whereas the MSQ did.)
First, we requested data for all U.S. medical schools. To assess the effect of a medical school's research emphasis on the research career intentions of its students, we also requested data for the top 40 and top 10 research-intensive medical schools (as determined on the basis of the amount of NIH research funding received during the 1999 fiscal year). For the MSQ, the response rate for these groups was initially 70-78% and increased to approximately 95% during the period studied. For the GSQ, the response rate was initially 59-66% and increased to approximately 80% during the period studied. Men and women had almost identical response rates.
To determine how research intentions changed during medical school, we requested AAMC cohort data for individual medical students at all U.S. medical schools as well as at the top 40 and top 10 research-oriented medical schools. The research career intentions of individual medical students were monitored from matriculation to graduation by tracing their responses to the question concerning expected research involvement on the MSQ and the GSQ. A cohort comprised students who began medical school in the year in which the MSQ was taken, with that year used to identify the cohort, and ended when these students graduated from medical school. Because of data limitations, we could obtain information for only six cohorts extending from 1987 to 1996. For the national data, an average of 4,015 women (65% response rate) and 6,153 men (62% response rate) were tracked during each cohort period. For the top 40 schools' data, an average of 1,327 women (59% response rate) and 2,021 men (57% response rate) were tracked during each cohort period. For the top 10 schools' data, an average of 310 women (59% response rate) and 448 men (55% response rate) were tracked during each cohort period.
To assess research participation during medical school, we collected data from the Medical Scientist Training Program (MSTP) and the Cloisters Program. The MSTP provides funding for students to obtain both the MD and PhD degrees at 1 of 39 medical schools throughout the United States. The Cloisters Program, offered jointly by the Howard Hughes Medical Institute and the NIH, provides funding for medical students to take a year off from medical school to conduct research at the NIH campus in Bethesda, Md. We obtained a gender breakdown of MSTP appointees and Cloisters Program applicants over time.
The percentages of matriculating and graduating male and female medical students who expressed strong research career intentions were tabulated. Confidence interval (CI) limits at the 95% level with regard to these percentages were calculated with the use of Stata software (version 6.0, CI, which computes percentages and CI limits; Stata Corp., College Station, Tex) and are displayed in Figure 1 and Figure 2. Linear regression analysis lines were fit to the percentages shown in Figure 1 and Figure 2 with the use of Stata software (version 6.0, regress), and the resulting slope coefficients, standard errors of the mean, and significance values are listed in Table 1. The percentages of male and female students who expressed strong research career intentions upon matriculation but whose research intentions decreased during medical school are shown, together with their 95% CI limits, in Table 2. Calculations were performed with the use of Stata software (version 6.0, CI).
Matriculating Medical Students
Figure 1 A-C shows the percentage of matriculating students with strong research career intentions at all U.S. medical schools, the top 40 research-oriented medical schools, and the top 10 research-oriented medical schools between 1987 and 1997. The figures show that a consistently smaller percentage of women than men expressed strong research intentions. The male-female differences are statistically significant in every year except 1988, 1989, and 1991 for the all-school and top 40 school pools. The differences were also significant for the top 10 schools except in 1988, 1989, and 1991. A comparison of matriculating medical student research career intentions in the national, top 40, and top 10 pools of medical schools reveals an apparent correlation of a medical school's research intensity with the percentage of its matriculating medical students who express strong research intentions. Students at the top 10 research-oriented medical schools were most likely to matriculate with strong research career intentions, followed by students at the top 40 research-oriented medical schools, and finally students in the national pool of medical schools. Linear regression analysis (Table 1) performed with the percentages shown in Figure 1 A-C also revealed a downward trend in the percentage of matriculating male and female medical students expressing strong research career intentions. The slopes were statistically significant for the national pool and the top 40 schools and close to significant for the top 10 schools.
Graduating Medical Students
Figure 2 A-C shows the percentage of graduating medical students with strong research career intentions at all U.S. medical schools, at the top 40 research-oriented medical schools, and at the top 10 research-oriented medical schools during the 1987-1996 period. The male-female differences are statistically significant in all years in the national pool and among the top 40 schools and in all years except 1991 in the top 10 schools. In comparing Figure 1 and Figure 2, it is apparent that matriculating and graduating medical students showed the same trends. Fewer female than male medical students indicated research career intentions. A comparison of the different groupings of medical schools revealed that graduating medical students at the top 10 research-oriented medical schools were more likely to have strong research career intentions than those in the top 40 research-oriented medical schools or in the all-schools pool. As with matriculating medical students, the percentage of graduating medical students with strong research intentions declined over time among both men and women in a generally statistically significant way (Table 1).
Table 2 shows the change in the research career intentions of several cohorts of medical students who had expressed strong research career intentions upon matriculation. We focus on whether the majority (i.e., 50% or more) of those who had expressed such research intentions at matriculation maintained them throughout medical school. In the top 10 schools, the upper confidence percentile for women was less than 50 for the 1987 and 1992 cohorts, which implies that the majority of women in these cohorts who started medical school with strong research career intentions maintained them throughout medical school. In the intermediate years, the upper percentile exceeded 50, indicating that there was a substantial decrease in research career intentions among these students while in medical school. For men in the top 10 schools, the corresponding upper percentile was below 50 for all cohorts except 1992. The point estimate, even for the 1992 cohort, is 29.4%, suggesting that the high upper end of the 95% CI is a consequence of the small sample size (N=82).
Among the top 40 schools, higher percentages of women who initially expressed strong research career intentions no longer did so at graduation. Indeed, based on the upper 95% CI percentile, the majority of these women showed decreased research career intentions during medical school. This finding was not true for men, among whom the majority who started out with strong research career intentions maintained them throughout medical school. This finding was true for all cohorts except the one that matriculated in 1991. When results from all U.S. medical schools were analyzed this way (Table 2), the 95% CI percentile for both men and women in all cohorts exceeded 50, meaning that a majority of students of either sex showed a decrease in research career intentions during medical school.
The participation rates of male and female medical students in the NIH-sponsored MSTP and in the Cloisters Program are shown in Figure 3 A and B, respectively. In both programs, most participants were and continue to be men, but female participation is increasing. Between 1980 and 1996, the number of female MSTP appointees increased from 115 to 256, representing an increase from 16.6 to 29.1% of total appointees. During the same interval, the percentage of female first-year entrants to medical schools increased from 28.7 to 42.7%. For the Cloisters Program, the number of female applicants increased from 19 (29.2%) in 1985 to 75 (32.9%) in 1999. Thus, the steady increase in female medical students has translated into an increase in female participation in research programs during medical school. The percentage of women participating in the research programs tallied has not increased nearly as fast, however, as the proportion of all female medical students.
This study traces the influence of gender on the research career intentions and participation of medical students, who represent the next generation of physician-scientists. A significant and persistent gap exists with regard to women's and men's research intentions during medical school. For both matriculating and graduating medical students, a consistently smaller percentage of women expressed such intentions. Although there is an imbalance in the research career intentions of male and female medical students, the same discouraging temporal trend is reflected among both sexes. During the decade between 1987 and 1997, male as well as female medical students steadily lost interest in careers with a research emphasis. The trend exists for both matriculating and graduating medical students, implying that students are losing interest in research careers before they even enter medical school. It is not known whether this decline is due to a true decline in research intentions of potential medical students or whether it is due to a medical school admissions process that discourages or discriminates against students who are interested in a research career. Should this bias be more powerful toward women than toward men, its impact would be even greater.
It is particularly worrisome that the decline in the percentage of matriculating and graduating students with strong research intentions, as judged by regression analysis, was the steepest at the 10 most research-intensive schools. These schools are responsible for educating and developing many current physician-scientists, and it is puzzling that they should show the most negative trends concerning medical student research career intentions. Reversing this alarming trend should be a high priority for institutional leaders at these schools—and at others as well.
The individual student cohort data assessing the effect of the medical school experience on research career intentions supported the trends identified for pools of male and female medical students. Women were consistently more likely than men to decrease their research career intentions during medical school. For both men and women, students who attended more research-intensive schools were more likely to retain an initial strong research career intention than were those who attended less research-intensive institutions.
After combining the data concerning medical student research intentions with the known shift occurring in the gender composition of medical schools, we estimated that the actual number of both male and female medical students with strong research career intentions has been declining. The numerical decline has been more significant among men because of the combination of two negative trends: a decline in the percentage of men with strong research intentions and a decline in the total population of male medical students. Women with strong research intentions also have shown a numerical decline, but the decline has been tempered by the increase in their proportion in the overall medical student population. Thus, women have influenced the decline simply through their presence in medical school. To stem the decline in the number of young physician-scientists, institutions must improve their efforts at all stages of the medical research pipeline to portray medical research as a rewarding and enjoyable career that can be combined successfully with family obligations. Although men must continue to be targeted by the new research grant programs offered by the NIH and the not-for-profit sector,12,13it is essential that the consistently lower level of expressed career intentions and participation of women in medical research be addressed.14Because women comprise an increasing majority of students and clinicians in medical subspecialties such as pediatrics and obstetrics/gynecology, efforts must be made to increase their interest in research so that these fields do not lag behind scientifically. Ultimately, failure to increase the number of women physician-scientists will put at risk the very patients who depend on the ever-larger female physician population.
Several people were extraordinarily helpful in providing the authors with the data analyzed in this report: Andrew Quon at the AAMC, Robert Moore and Burton Shapiro at the NIH, and Barbara Filner at the Howard Hughes Medical Institute. Patricia Fox provided expert secretarial and editorial assistance.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.