Despite the many financial pressures facing the nation's academic medical centers, future generations of physicians still need to be educated, and advances in research and their application to clinical care still need venues. The need to nurture a cadre of leaders in the academic disciplines also remains. The current underrepresentation of women and racial and ethnic minorities in leadership roles in academic medicine presents a challenge to identify and eliminate any potential barriers to their participation. There is ample evidence of harassment, salary inequities, and a sense of isolation based on gender or racial or ethnic background occurring within the academic setting that impede the initiation or continuation of careers in academic medicine by women and minorities.1,2 In a barrier-free world it would be reasonable to expect that the increasing diversity of the general population would be reflected in the academic ranks. A present, this is not the case.
We know little about how residents choose to pursue careers in academic medicine, but we hypothesized that their perceptions about academic medicine as a career and mentoring during training and beyond are critical factors residents use to determine their career paths. Residents gather their perceptions about the desirability of an academic career path from experiences with and observations of the difficulties and personal satisfactions their mentors face at work. A better understanding of the perceptions of residents as they proceed toward their career decisions, as well as those of individuals (Fellows) who have made career path decisions, may help frame the development of educational environments that will nurture a strong and diverse cadre of academicians for the future.
To gather this information, and to formulate strategies to foster diversity among future academic educators and researchers, we surveyed all residents in training in obstetrics and gynecology in 1998 and performed a linked survey of Fellows of the American College of Obstetricians and Gynecologists (ACOG).
We administered two separate scaled surveys that focused on attitudes and opinions toward academic medicine: one to a group of 2,000 Fellows of the ACOG, and the other to the 4,814 obstetrics and gynecology residents taking the 1998 Council on Resident Education in Obstetrics and Gynecology examination. The Fellows were 240 members of the Collaborative Ambulatory Research Network (CARN), a representative group of Fellows who voluntarily participate in ACOG research department surveys, and 1,760 Fellows (excluding CARN Fellows) selected at random.
The questionnaires asked about demography, attitudes, and opinions regarding academic medicine, and the residents' and Fellows' experiences with having and being a mentor. To ensure the validity of content, the questionnaires were reviewed for content by a panel of resident—physicians and physicians in practices in and out of academic settings.
We analyzed the data with a standard statistical software package. We computed descriptive statistics for the measures used in the analyses, which are reported as means. We used two-tailed t-tests to compare group means of continuous variables. We used chi-square analysis to assess differences on categorical measures, and the Mann—Whitney U and Kruskal—Wallis tests to assess group differences on ordinal measures. We tested all analyses for significance defined at an alpha of .05.
A total of 4,659 residents (96.8%) and 811 Fellows (40.6%) responded to the survey; 166 (69.2%) of the Fellows were CARN Fellows and 645 (36.6%) were not.
Among the Fellows who responded, 63.9% were male, 34.8% were female, and 1.2% declined to state their gender. These respondents were representative of the ACOG Fellowship, of which 64.4% were men. Among residents who responded, the gender proportions were reversed: 35.3% were men, 64.3% were women, and 0.4% declined to state their gender. Those that declined to state their gender were removed from the database prior to the statistical analyses. The majority of all respondents reported their race as white. Table 1 shows the characteristics of the respondents.
Women Fellows were younger than were men Fellows. There was no age difference between white and nonwhite Fellows, and the median age of Fellows who responded was 46, the same median age as that of the Fellowship at large. Thus, the responding Fellows represented the same demographics as the ACOG Fellowship. The number of minority Fellows responding was too small for separate statistical analysis. The majority of the Fellows (50.7% of men and 52.8% of women) had never been in academia; only 23% (123) of the men Fellows and 22.5% (64) of women Fellows in this survey were in academic medicine.
Choosing Academic Careers
The residents' responses are described by year of residency in Table 2. Among residents, 26.1% agreed or strongly agreed with the statement “I would never consider a career in academic medicine.” For women, fourth-year residents were more likely to agree or strongly agree (33.7%) than were first-year residents (20.3%). First-year women residents were more favorably disposed toward a career in academic medicine than were first-year men residents (p = .042). There was no significant difference between the genders in other years, culminating in nearly equivalent views in the fourth year of residency.
In general, Fellows who had left academia chose “too bureaucratic” and “low salary” as their reasons for leaving. Men tended to say that “wanted more time for personal/family” had been a factor (64.3%), and women tended to say this had not been a factor (60%). However, the men's responses were spread out over the range of 1 to 4 (from 1 = not at all a factor to 4 = determining factor), with a modal response of 3 (37.5%), while the women were bimodal or nearly equally split between 1 (40%) and 4 (36%), which suggests that the women respondents formed two different subgroups (see Figure 1).
Both men and women Fellows who chose not to pursue a career in academic medicine chose “not interested in writing grants,” “doing research,” and “writing publications” as reasons. The women were more likely than were the men to give as reasons that “academia is too competitive” and that they “do not like to teach medical students,” but neither of these was selected by a majority of women. The men were more likely than were the women to give as a reason “excessive administrative duties,” but a majority of the women also agreed with that statement. As mentioned above, further subdivision of the responses by race and ethnicity of the Fellows was not possible due to the small numbers in these subgroups.
Among the residents, there were significant gender differences in the perceptions of which gender was more actively mentored and recruited for faculty positions (see Table 3), which gender was more likely to receive helpful career advice, and towards which gender supervisors were likely to exhibit a condescending attitude. The women residents were more likely than were the men residents to believe that men were more actively recruited for faculty positions. Proportionately, the nonwhite women were the most likely to believe that men were more actively recruited.
Among the Fellows, roughly equivalent proportions of men and women (32.6% and 36.5%, respectively) reported being actively recruited for faculty positions. The women Fellows generally agreed with statements that defined the determining factors in women's underrepresentation in academic medicine as: “males are more heavily recruited for faculty positions” (mean = 2.92), and “males are more likely to receive helpful career advice from supervisors” (mean = 2.84). The men Fellows, however, generally did not agree with these statements (means = 1.82 and = 1.69, respectively, p < .001). Just under half of the women Fellows agreed that “supervisors are more condescending/hostile toward women” was a factor (mean = 2.49), as opposed to a mean of 1.72 from the men Fellows on the same statement (p < .001). Both men and women Fellows generally agreed that “balancing academics and home life is difficult, and women tend to have more family responsibilities” (mean = 3.25) and “salary and benefits are better outside academia” (mean = 2.91) are factors.
Race and Ethnicity Issues
The majority of the white residents felt that there was no racial bias in recruitment for faculty careers, while the majority of the nonwhite residents believed that white residents were preferentially recruited (see Table 4). When we combined responses to questions about bias based on race, ethnicity, and gender in the recruitment of residents for faculty careers, 27.3% of the residents (the largest group) said there was no bias (Table 4). The next-largest group (22.3% of the residents) indicated that white men were favored in such recruitment. However, there were differences in responses between white and non-white residents (p < .001) and between genders for both white and nonwhite residents (p < .001) with respect to whether they felt there was bias in recruitment of residents for faculty careers. The group of residents that indicated there was neither gender nor racial or ethnic bias contained a higher proportion of whites than did the total sample (79.6% as opposed to 64.1%), but this group of residents was identical in its gender makeup to that of the total sample (64.5% women, 35.5% men). The group that indicated that white men were preferentially recruited was largely composed of women (82.3%), and the majority of the women were nonwhite (55.0%).
The vast majority of white residents believed there was no racial bias in career advice from supervisors, while a plurality of nonwhite residents believed that whites were more likely to get helpful advice (Table 4). Both white and nonwhite residents rarely thought that nonwhites were more likely to get helpful advice.
A majority of all residents believed that supervisors displayed equally condescending attitudes toward all individuals regardless of racial or ethnic background. Of residents who noted a difference, white residents and residents from minority subgroups also believed that of all racial or ethnic groups, nonwhites were more likely to be the recipients of condescension. However, comparing these results with an analogous question concerning gender indicates that gender was perceived to be a more important source of bias than was race. Only 6.9% of residents believed there were condescending attitudes by racial or ethnic status but not by gender, while 25.1% of residents believed there was gender bias but not racial or ethnic bias. The results also indicated that all of the respondents thought that nonwhite women were most likely to face such attitudes. Furthermore, almost one third (33.2%) of the nonwhite women residents, as a subgroup, responded that supervisors were more likely to display a condescending attitude towards both women and nonwhite individuals. A total of 45.7% of the nonwhite women residents believed that supervisors either did not display condescending attitudes or displayed them without gender or racial or ethnic bias.
As stated above, the number of minority Fellows responding was too small for separate statistical analysis.
The legacy of medicine will depend on those who lead and educate the next generations. Assuring a leadership that is diverse in race, ethnicity, and gender requires investment in a broad strategy of support and encouragement for potential academic physicians, from their pre-college years through their advancement to faculty positions in academic medicine or in leadership roles.3–6 This study focused on the perceptions and issues that impinge on residents' making the decision to enter academic medicine as a career at graduation. It is at this point in medical education that all the positive aspects of training (mentoring, role modeling, early encouragement) are balanced against negative experiences, particularly the cumulative effects of isolation and perceptions of discrimination in advice and support.
Our surveys found disturbing trends that may inhibit attracting a diverse group of men and women into academic medicine. Both men and women residents felt that the other gender received more or better counseling and mentoring—suggesting that neither received the individualized and specific support they needed to sustain interest in academic medicine. This finding adds to a foundation of perceptions of biased mentoring noted by fourth-year women and minority medical students in an earlier study.7 In obstetrics and gynecology, there is a perception that market pressures favor the selection of women residents; the potential for discouraging men students and residents from entering into academic careers in this discipline exists. The finding that men and women residents in obstetrics and gynecology perceived a lack of support from their faculty could be a warning that biased treatment will be as destructive to recruitment of men into academic medicine as it has been to women.
A total of 49.5% of women residents and 27.4% of men residents felt that supervisors displayed a more condescending attitude toward women physicians and scientists, while neither group (4.3% or 9.5%) saw any condescending attitude towards men physicians and scientists. Our finding of expressed bias against women physicians and scientists is not a phenomenon limited to colleges of medicine; it recapitulates experiences students and faculty have in undergraduate or other settings.8 Of particular concern is that this experience is alive and well among women residents in obstetrics and gynecology, a discipline that now has a preponderance of women residents. Less blatant forms of bias such as stereotypes about “good” career choices or acceptable patterns of interaction that differ by gender also exist and, given the experiences of bias we found, they continue to occur. The greater rate of decline in the women residents' interest in careers in academic medicine from PGY-1 to PGY-4 compared with the rate of declining interest in men residents suggests that women receive a more negative message about this career choice. All faculty need to carefully examine their behaviors and make the necessary changes to provide unbiased, effective support, and to encourage a full range of career options for all our young trainees.
The perception of nonwhite residents that they received less positive and even biased mentoring during their training creates a barrier to their consideration of academic careers. Most of the nonwhite residents (60%) felt that white residents were more actively recruited for careers as faculty, while 56.2% of the white residents felt that no racial or ethnic bias existed in career recruitment. The differences we found in these residents' perceptions about mentoring along racial and ethnic lines are not unusual, and the dissonance in these beliefs that nonwhite trainees perceive reduces the likelihood that they will choose careers in academic medicine. If a nonwhite resident perceives that his or her experience is “invisible” to the majority of peers and the present academic world, then the problem cannot be addressed. It is particularly difficult for medicine, with its own cultural assumption of immunity from prejudices, to recognize that the same harmful biases and behaviors that exist within the community at large also exist within the community of medicine. It is difficult for individuals who experience prejudice to have the trust to make their experiences apparent in any forum other than these highly confidential surveys, “Because many of the stories are painful, and revealing one's pain involves an element of trust.”9 The experience of bias in mentorship must be addressed at all levels of medical training to reveal and overcome the unexamined stereotypes and beliefs that produce these perceptions.
The tripartite mission of academic medicine (research, education, and clinical care) and the perception that academicians must be strong in each area may also prevent residents from considering academic medicine at the critical juncture between residency and practice. Fellows who had chosen to leave or had never entered academic medicine identified the requirements for research and for teaching and the burden of overly complex administrative requirements as issues that had influenced their decisions. Administrators at academic medical centers can help to resolve these issues through support for tenure and advancement programs that value excellence in the many skills that contribute to the three areas of academic medicine. Our residents must be able to see a productive and valued role for themselves in the academic family, even if their scholarship focuses on only one or two of these areas.
Residents have a close working view of life in academic settings, yet they show a diminishing interest in joining that life over their four years. Greater familiarity with an academic life, then, breeds “contempt” for it rather than attraction. While we were concerned with the influence of mentoring on residents' career decisions in these surveys, it is also possible that other factors, such as perceived disparity in incomes between academic and “private” practice models, particularly for residents with large debts, also influence residents' decreasing interest. Such factors, when combined with what residents assume about expectations of productivity in academic administration as well as teaching, research, and clinical care, and their perceptions of biased or inadequate mentorship, are all potential negative influences on the development of future faculty. These concerns can be addressed by valuing residents' disparate sets of skills and challenging mentors and faculty to overcome stereotypes that lead to biased treatment of individuals.
An area of particular concern for the discipline of obstetrics and gynecology is whether women, now entering the field in greater proportions, can develop their interest in academic medicine or leadership roles as easily as can their colleagues who are men. One issue women face is the desire to, and the social expectation that they will, spend more time with family. Our unexpected bimodal finding in the responses of the women Fellows about the importance of family issues in making their choices for academic medicine indicates that perceptions of its importance vary widely. The Fellows who were men placed great importance on family responsibilities in making their career choices as well. Accommodations for family responsibilities could increase the attractiveness of an academic career for both men and women.
In sum, our findings do not paint an encouraging picture of the ability of obstetrics and gynecology to use mentoring to attract and retain women and minorities to academic faculty positions and positions of leadership. There is also an indication that men as well may be experiencing discrimination in the mentorship they receive. Dr. Cohen has noted “A racially and ethnically diverse faculty, fully empowered by the equitable presence of minorities within all ranks of the academy, is the only conceivable bridge to the diverse physician workforce and the culturally competent health care system that the full spectrum of the American public deserves.”3 A bridge to race, ethnicity, and gender diversity requires current leaders to be role models of positive and unbiased mentorship. It also requires that faculty be given and give to others, particularly their trainees, broad encouragement to take advantage of all opportunities for advancement to leadership roles and for experience with all types of academic scholarship. And, it requires affirming the need for gender and racial and ethnic diversity through strategies for addressing the issues outlined here and attracting and mentoring diverse student, faculty, and leadership groups. It is time to assure that the legacy of this decade will be the opening of the opportunities of academic medicine and leadership in all disciplines to all the individuals who could choose academic medicine as their life's work.