To examine the perceptions of rheumatology fellows regarding their research training, mentoring, and interest in a career in academia.
To examine the perceptions of rheumatology fellows regarding their research training, mentoring, and interest in a career in academia.
We solicited by e-mail 386 fellows in the American College of Rheumatology 2005–2006 fellow database to take an anonymous Internet-based survey addressing the topics of research training, mentoring, and interest in an academic career.
We received 176 responses (50% response rate after excluding invalid contacts) to the survey. During their training, 58% of fellows reported an interest in academia and 21% in research. There was great satisfaction with mentoring. However, there were concerns about academic salaries, with 50% of respondents stating a preference for a higher paying community position. Furthermore, there were substantial concerns about the difficulty of generating funds to cover salaries. In addition, several respondents viewed an academic career as incompatible with starting a family. Compared with male fellows, female fellows were more likely to want a career in academics, were less concerned about academic salaries, and were more concerned about funding and family life.
Despite an interest in academia and satisfaction with current mentoring, several barriers to academia were identified among rheumatology fellows. The concern that academia and family life are incompatible needs further attention. University deans should consider reevaluating promotion programs to make allowances for family and parenting demands. Rheumatology division chairs should better promote the nonfinancial rewards of a career in academia. Programs such as the National Institutes of Health Loan Repayment Program should be strongly advertised to interested applicants with financial concerns.
The decline in the number of physician researchers across medicine has been well documented. Nearly 30 years ago, the clinician investigator was described as “an endangered species” by then National Institutes of Health (NIH) director James Wyngaarden (1). Ten years ago, Leon Rosenberg, Professor of Microbiology at Princeton and a well-published authority on academic medicine, deemed physician scientists to be “endangered and essential,” suggesting a collaborative national effort through the creation of a “broad-based national panel of leaders” to develop recommendations (2).
The number of physicians choosing a career in research has declined from 23,268 in 1985 (4.6% of the physician work force) to 14,340 in 2003 (1.8% of the work force) (3). The physician-scientist work force is also aging, with ∼50% of NIH award recipients age ≥50 years (3). Meanwhile, success for first-time physician R01 applicants has been falling (4).
In response to these concerns, career development awards have been expanded and the NIH Loan Repayment Program was created in 2001. However, additional factors such as the current difficult funding climate at the NIH and private foundations and the associated faculty insecurity may inhibit correction of recent trends. Examining the concerns from the standpoint of fellows may provide insight into genuine barriers to pursuing an academic career.
The National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) recently examined the success of their training grant (T32, F32, K01, K08) recipients between 1993 and 1997 (5). Ten years after the initial training award, 17% of 271 T32 recipients had obtained R01 funding, 78% had published within the last 10 years, and 50% within the past 2 years. Among 58 K08 award recipients, 55% had obtained R01 funding, 98% had published within the last 10 years, and 85% within the last 2 years. Therefore, the K08 mechanism was successful for ∼50% of its recipients.
One of the working group's final assessments was that “the most significant impediment to attracting and retaining qualified individuals for careers in NIAMS-related fields is the (accurate) perception that this is a high-risk career path. All NIAMS-supported trainees eventually face the increasing daunting challenge of achieving and maintaining independent R01 support” (5). In addition to recommending efforts to increase funding for training and R01 grants, the working group recommended reinforcing the value of grant writing and management, and the value of mentorship.
Rheumatology as a field is not exempt from these concerns. The American College of Rheumatology (ACR) recently examined these issues in its 2005 work force study of US rheumatologists (6). This study projected a shortage of clinical rheumatologists by 2025, attributable to the increasing prevalence of musculoskeletal disease in the aging US population and a lack of projected growth in the supply of rheumatologists. The study went on to examine the health of academic rheumatologists, reporting that the time to independent research status averaged 7–8 years after faculty appointment, with 35% of academic respondents reporting more than 9 years required to establish independent status. The proportion of rheumatologists receiving independent grant support (defined by an R01 award) declined from 41% to 22%.
Of those respondents who left academics within the last 5 years, the reasons cited for leaving included lack of support (50%), interest in higher paying jobs (50%), difficulty funding research (32%), shorter work hours (27%), and inadequate mentoring (13%). The majority of those leaving academics (61%) entered private practice.
In response to this information, the ACR and the ACR Research and Education Foundation have committed to expanding support of rheumatology training programs through awards, recruitment initiatives for medical students and residents, and increased number of young investigator awards to “preserve our current academic programs” (7). The ACR Committee on Research also created and charged the Young Investigator Subcommittee to explore how the ACR could best serve the needs of its younger research members. To address these concerns, the subcommittee developed a survey to assess the 2005–2006 rheumatology fellows' interests in pursuing a career in academics and research, and perceived barriers to a career in academics. The results of this study are reported herein.
The Young Investigator Subcommittee designed the domains and questions for the survey (for a copy of the survey see supplemental Appendix A, available in the online version of this article at http://www3.interscience.wiley.com/journal/77005015/home), which was then developed by the ACR staff for Web-based implementation. All rheumatology fellows in the ACR fellows' database were solicited by e-mail in May 2006 to complete the online survey. Two reminder solicitations were sent in the subsequent 2 months. Participants completing the survey entered a drawing for a free registration to the 2006 ACR Annual Scientific Meeting and free textbooks (adult or pediatric) as incentive to complete the survey. Responses to the survey were collected without personal identifiable data. Because the subcommittee had interest in disseminating results of the survey, the University of California, Los Angeles Office for Protection of Research Subjects reviewed the protocol and designated the project exempt from further Institutional Review Board review.
The primary goal of the survey was to assess fellows' interests in academic medicine, their perceptions of the quality of their mentoring, and perceived barriers to a career in academic medicine, including concerns about funding, salary compensation, and family life.
Respondents were asked to describe their type of rheumatology program (adult, pediatric, or combined adult/pediatric programs), their year in fellowship, sex, age, race, and ethnicity. Respondents were asked to describe their fellowship program's hospital affiliation (university, veteran, other government, or private) and the number of fellows enrolled in the program.
The mean, median, and percent responses were calculated. Bivariate analyses were performed using Student's t-test, Wilcoxon's pairwise comparisons, chi-square test, or Fisher's exact tests as indicated. Analyses were weighted to adjust for potential duplicate survey responses. Owing to the structure of the Web-based survey, missing responses to individual questions were uncommon. Those records with missing responses were dropped from the related analysis. All analyses were conducted using SAS statistical software (8).
The ACR maintains a database for US and Canadian fellows for the purpose of communicating with rheumatology fellows. During the 2005–2006 academic year, fellows' names and program year (e.g., first year, second year, or later) were available in the database for 363 US fellows and 46 Canadian fellows (n = 409 fellows). For the same year, the Accreditation Council for Graduate Medical Education recorded 380 US rheumatology fellows at 108 programs, suggesting the ACR database was fairly complete (9).
For the 409 fellows in the database, there were 355 valid e-mail addresses. Through 3 separate e-mail solicitations toward the end of the academic year, we received a total of 185 responses from 128 unique user Internet provider (IP) addresses. Six responses had little to no data (with less than 1 minute spent at the Web site by the user) and were dropped from the database.
Since many fellows may have completed these surveys at work on shared computers, it is not surprising that we received duplicate user IP addresses. Upon examining the 69 responses that did not have a unique IP address, we used self-reported age, sex, race/ethnicity, program type, and year of training to identify potential duplicate responses. We found 4 sets of duplicate user IP addresses with indistinguishable self-reported descriptors for a total of 10 responses. (Figure 1). We then looked at self-reported research project titles and other fields. For 1 of the 4 sets, the research project titles and other text fields were so disparate that we determined these to be 4 unique individual responses. For 2 of the remaining sets (2 responses each), the research titles or start-stop times completing the webform indicated that they were duplicate responses. There was 1 remaining set (2 responses) with an identical user IP address and descriptors that we could not determine were unique, and therefore identified them as potential duplicates. This resulted in a final analytic sample of 173 unique respondents with 3 sets of likely duplicate respondents (6 responses). For all analyses, the potential duplicate responses were weighted by 50%, resulting in a final weighted analytic sample of 176 fellows (50% response rate of those with valid e-mail addresses). From the analytic sample, fellows spent a median of 10 minutes completing the survey (25th, 75th percentiles 8, 14 minutes, respectively).
Respondents had a mean ± SD age of 33.3 ± 3.8 years, 56% were women, and 57% identified themselves as white (non-Hispanic), 36% Asian, 4% Hispanic, 2% African American, and 2% other (Table 1). The majority of fellows were in adult rheumatology training programs (81%) with a similar number of first-year and second-year respondents (44% and 43%, respectively), and 13% of respondents were in their third year or later. Most fellows were affiliated with university hospitals (90%), followed by programs that had affiliations with veterans (47%), other government (23%), and private hospitals (15%). The total tallies to >100%, since respondents could check all that apply. Median fellowship program size was 4 fellows (25th, 75th percentiles 3, 6, respectively).
|Respondents (n = 176)|
|Age, mean ± SD years||33.3 ± 3.8|
|Fellowship program size, median (25th, 75th percentiles)||4 (3, 6)|
When asked about the numbers of months available for research, fellows reported a median 1 month of research during their first year (25th, 75th percentiles 0, 2 months, respectively), and a median 8 research months during their second year (25th, 75th percentiles 3, 11 months, respectively). Fellows identifying an interest in academics or research reported significantly more second-year research months than fellows with a nonacademic interest (10 versus 4.5; P = 0.0009), as did fellows with interest in research versus fellows with interest in patient care (10 versus 6.5; P = 0.0004).
Fellows were further asked to describe how much of their week was protected for research during their typical research month. An equal number of fellows, ∼1 in 6, described their protected time for research as either <60% or >90% during their research month. First-year fellows reported having much less protected time than second-year fellows. Again, fellows reporting an interest in academics or research described a greater amount of protected time than their counterparts (P = 0.02 and 0.03, respectively). During their research month, 59% of all fellows reported taking call and described their clinical workload as moderately busy.
In addition to protected time, the fellows were queried about available formal research training. Two-thirds of fellows reported that their institution had specific research training programs. Of those reporting structured training programs, the described programs included research symposiums or classes (78%), Masters programs in research (86%), or Doctoral programs (48%). However, 37% of these respondents reported that they would need to find funding if they wanted to take advantage of these programs. In a separate question, only 54% of fellows reported being aware of any grant-writing classes.
Eighty-five percent of fellows stated that they had a research project and all but 1 fellow identified a research project mentor. Fellows favorably described their research mentors, with 90% describing them as role models with high ratings regarding their mentor's reputation and track record with prior mentees. Mentor involvement included providing the project (57%), helping design the fellow's own project (65%), providing funding (33%), guaranteeing to provide future funds (11%), or being instrumental in obtaining funds (28%). More than 50% of fellows reported being either very or extremely satisfied with both the availability and quality of their mentoring to date (52% and 53%, respectively). However, 59% of the fellows were unaware of any formal research mentoring programs at their institution. One of the fellows responded with the following qualitative comment: “At my particular institution, we could not ask for better mentoring. It was the strong role modeling that led me to decide to take an academic position despite the loss of income, increased work load, and stress that are inherent to academia.”
Fellows were asked to identify their principal career interest; 36 fellows (21% of respondents) identified research and 137 fellows identified patient care (3 missing responses) (Table 2). Of the 36 fellows expressing a primary interest in research, 21 fellows expressed an interest in basic science, 21 for non–industry-based clinical or translational research, 3 for clinical trials, and 3 for health service research. The total sums to >36 because fellows could identify multiple areas of interest. There were no differences in research interest by sex or year of fellowship. As anticipated due to the structure of pediatric fellowships, fellows enrolled in pediatric training programs were more likely to express interest in research than were adult rheumatology fellows (P = 0.03). All pediatric rheumatology fellowships are a minimum of 3 years, with greater time devoted to research.
|Entire sample (n = 173)†||Interest in research||Interest in academics||Sex|
|Patient care (n = 137)||Research (n = 36)||Nonacademic (n = 73)||Academic (n = 100)||Men (n = 75)||Women (n = 97)|
|Research is stated as principal career interest.||36 (21)‡||N/A||N/A||4||32§||19||23|
|I would like to pursue a career in academia.||100 (58)‡||49||92§||N/A||N/A||50||65§|
|Potential barriers to academics|
|If I wanted to pursue a career in academics, I would be able to get a position.||62||62||58||45||73§||64||62|
|Obtaining funding is or will be difficult.||78||76||86||73||82||70||84§|
|I am concerned about a lack of strong divisional and/or institutional support.||43||41||48||44||42||35||49|
|I would rather seek a higher paying job than an academic position can offer.||48||55||22§||62||38§||58||40§|
|Academic hours interfere too much with home life.||25||26||22||30||22||21||29|
|I have had inadequate mentoring to date.||22||22||23||19||24||18||26|
|I have never desired an academic position.¶||28||34||6§||64||4§||33||25|
|Research is very important to me.||53||41||100§||29||71§||51||56|
One hundred (58%) fellows stated that they wanted a career in academics, whereas 73 preferred nonacademic careers (3 missing responses). Fellows' interest in academics was strongly correlated with their stated interest in research (P < 0.0001), with 92% of fellows expressing an interest in research stating that they wanted a career in academics. By comparison, fellows expressing primary interest in patient care were evenly divided between academic and nonacademic careers (49% versus 51%). A greater proportion of women versus men (65% versus 50%; P < 0.05) stated they were interested in academia.
We examined several potential barriers to a career in academics, including concerns about funding, mentoring, salary, work hour concerns, and baseline interest in academics and research. In addition to the single-item Likert questions (for a copy of the questionnaire see supplemental Appendix A, available in the online version of this article at http://www3.interscience. wiley.com/journal/77005015/home), qualitative responses were also solicited. A total of 39 qualitative responses (22.5% of respondents) were obtained and broken down into the following categories by frequency of comment: concerns about funding (17), concerns about salary (12), comments on training (9), comments about mentoring (9), and concerns about family life (6). Representative comments are included below to supplement the quantitative responses from the survey.
Neither lack of baseline interest in academics, interest in research, nor inadequate mentoring was cited as a barrier to academics (Table 2). More than three-quarters (78%) of fellows reported being satisfied with their mentoring to date. However, several respondents emphasized the importance of early mentoring: “The critical point is that mentoring is crucial for fellows at the beginning of their training,” and “We need time to think and plan and meet with potential mentors early in the first year of fellowship. We need help coming up with meaningful projects that can [be] finished during fellowship.”
Although two-thirds of fellows (62%) believed that positions would be available to them if they were interested in pursuing a career in academia, nearly half of the fellows (43%) expressed concerns about the potential lack of strong divisional or institutional support. Funding was the primary concern about pursuing an academic career, with 78% of fellows stating obtaining funding for their salary would be difficult (Table 2). When specifically asked about the availability of funds for a university position and the effort needed to obtain those funds, only 9% of fellows responded that funding was readily available for a research position, whereas 33% of fellows reported funding available for a university clinical position. Two-thirds of fellows reported that with large effort on their part, funding for a research or clinical position would be available. Some of the written comments received follow here: “From personal observation of junior faculty, funding for clinical research and even translational research is extremely difficult,” “While young researcher funding sources are available competitively, I am concerned about the overall funding environment through the NIH, etc.,” and “Academics seems like a fairly scary proposition. Unsecured funding is getting more and more difficult to obtain.” Funding concerns were more prevalent among female fellows than male fellows (84% versus 70%; P < 0.05).
Almost half of respondents (48%) stated that they would rather pursue a higher paying job than an academic position would pay, with male respondents more likely than female respondents (58% versus 40%; P < 0.05) to express interest in a higher paying job (Table 2). Fellows were separately asked how they thought an academic research or clinical salary compares with a community salary. More than half of the respondents reported that a research salary would pay <60% of a community salary. Fellows had the impression that pay for clinical positions was better than research positions but still sharply below community incomes (Figure 2). These perceptions accurately reflect the results of self-reported incomes from the ACR work force study (10). Again, the following written comments were reported: “Until the academic centers can come up with salaries that can be competitive, it is my opinion that the academic rheumatologist will become a dinosaur,” and “My student loans exceed $225,000, simple mathematics drive the need for a higher paying position with a guaranteed income.”
In today's society, economic pressures and division of labor for parenting among 2 working parents is impacting the more traditional academic working model. Twenty-five percent of residents stated that they thought academic hours would interfere with home life (Table 2). There were several strongly worded written comments describing the potential difficulties of balancing family and academia: “Medical training to date has taken a toll on my family; they deserve more of my time. This (more than money) is my primary motivating factor in making a long-term career decision” (male respondent), and “As a woman who wants to have a family, I think that it's difficult to be on a strict tenure-track position without having the ability to put promotional evaluation on hold for one year [per] new child.”
Results from this survey provide several interesting insights. In contrast to our a priori concerns about the adequacy of mentoring, fellows reported being satisfied with their mentoring. This was an encouraging finding, although this was a self-reported assessment and fellows in training may not be the best judges of whether their mentoring is adequate. During fellowship, academic mentoring is focused on learning skills and obtaining early funding. Later, mentorship must include attention to career development, academic culture, long-term strategies for independence, and continued funding.
Despite current challenges, fellows still express interest in academia. Training programs have an excellent opportunity during the fellow's training to either favorably impress or discourage fellows about the potential rewards from a career in academia, depending on how faculty portray their job satisfaction.
Fellow respondents have identified several important potential barriers to academia. With medical school debt averaging $129,943 (11) in 2006 and 85% of students carrying more than $100,000 of debt at graduation (3), it is not surprising that debt obligation was cited as a barrier to academia. However, the NIH Loan Repayment Program (http://www.lrp.nih.gov) offers an excellent means to largely reduce this potential barrier. The program repays the applicant's educational debt (minus a small participant obligation) at the rate of 25% annually, up to a maximum of $35,000 per year for work in 1 of 5 areas (clinical research, pediatric research, health disparities research, clinical research for individuals from disadvantaged backgrounds, and contraception and infertility).
In 2007, the NIH Loan Repayment Program granted 1,646 awards totaling almost $74 million in payments. Success rates for new and renewal applicants were 42% and 72%, respectively (12). The NIH does not collect data on applicants by specialty; however, rheumatology fellows or recent graduates would be primarily applying to NIAMS. In 2007, NIAMS ranked twelfth out of the 21 institutes for a number of awards (42 awards), eleventh for total dollars awarded ($1.8 million), and fifth as a proportion of Loan Repayment Program dollars per total appropriations by institute.
Despite medical school debt and an accurate perception of lower remuneration for academic positions, respondents expressed tolerance about lower salaries but significant anxiety about securing funds to cover these salaries. Female fellows more frequently expressed these concerns, perhaps a reflection of their increased interest in academia. The concerns expressed in this survey echo the issues identified in the recent ACR work force survey (10). Specifically, problems funding research and higher pay ranked as the 2 highest reasons for leaving academia. A preference for providing community clinical care was the least cited reason for leaving academia.
As the time interval between postgraduate work and K-award, as well as the interval between K-award and R-award are increasing, young faculty need to find other means of salary support or programs will have to provide a greater share of young faculties' salaries. The increased funding pressure will inherently lead to attrition. The cost of interrupting research support to an established laboratory can be enormous. With NIH paylines at ∼10% of all submitted grant proposals, a new investigator who misses this cutoff after a few attempts faces the demise of a potentially promising career (13). Although this survey did not directly address the psychological stressors associated with the requirement to obtain research funding in academia, this is of particular concern in the current financial state of the NIH.
Concerns about starting family life and a career in academia ranked as the second-highest concern among respondents. Although 65% of women and 50% of men in our survey would like to pursue a career in academia, 29% of women and 21% of men stated that “academic hours interfere too much with home life.”
Attributable to the demands of medical education and training, the decision to start a family often occurs about the same time that an academic career is starting. Academic demands and family obligations compete for a finite number of hours. The most intense academic demands are made at the same time that parents need flexibility to care for their young children.
The traditional academic tenure–track timeline does little to address the realities of family life. With greater numbers of women entering rheumatology and the observation that some will opt to work part time, the projected rheumatology clinical work force shortage may be even worse for academia unless academic work schedules can become more flexible. With co-parenting becoming the model of family care, these concerns are not limited to female rheumatologists and are likely to be more prevalent in rheumatology today than they were in preceding decades, unless universities respond to demographic and cultural changes.
Along these lines, many institutions have begun to address these issues by giving candidates an additional year before promotion evaluation for each new child. It is vital to make academic institutions aware of these trends so that such family accommodations can be incorporated more universally.
The demands of private practice, in comparison with academia, should not be underestimated by young rheumatologists. Some private practices have minimum billing requirements or inflexible vacation and on-call schedules for their most junior partners. Industry jobs may require significant travel and can also have tight time demands. These realities should be weighed against the perceived challenges of academics.
Since academic careers are often not as lucrative as private practice, rheumatology along with other similarly structured divisions should stimulate their departments, universities, and hospitals to help make their positions more attractive to candidates, for example, by supporting subsidized on-site day care, job sharing, or flexible job hours. To retain the best and brightest physicians as investigators and teachers, it is vital to address the practical needs of parents of young children.
Based on the results of this survey, approximately half of rheumatology fellows are interested in academic medicine and the majority is satisfied with their mentoring. However, major perceived barriers to a career in academic medicine include concerns about funding and family life.
These findings do not bode well for the pipeline of young academicians. There is clear data that funding is harder to obtain, and it takes longer to develop research independence. Meanwhile, greater pressures are being placed on young graduates through greater family demands and higher indebtedness at culmination of training. Lower remuneration does not appear to be as significant a problem as does the anxiety about securing long-term funding for academic salaries. Certain actions may be taken at the division, department, or university level to reduce barriers to academia and foster a better perception about a career in academics. However, encouraging young rheumatologists to consider academia would be unsound unless advocacy work can be done at the national level to bring about improvements in the availability of funding for researchers and clinician educators.
The Association of Professors of Medicine recently made recommendations for revitalizing the nation's physician-scientist work force (3). Those recommendations included focusing on physician-scientist retention and mentoring, with emphasis on advancement and retention of female physician-scientists.
Dr. FitzGerald had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. FitzGerald, Mandl, Cron, Lacaille, Deane, Franchin, Lim, Park.
Acquisition of data. FitzGerald, Mandl, Cron, Lacaille, Deane, Franchin, Lim, Park.
Analysis and interpretation of data. FitzGerald, Mandl, Cron, Lacaille, Deane, Franchin, Lim, Park.
Manuscript preparation. FitzGerald, Mandl, Cron, Lacaille, Deane, Franchin, Lim, Park.
Statistical analysis. FitzGerald.
We would like to acknowledge the hard work and support of the ACR staff who implemented the survey and collected the data, and specifically thank LaTanya Batts, the ACR representative to the Young Investigator Subcommittee. We also wish to thank Dr. Jane Salmon, our respected mentor and the ACR Committee on Research liaison representative, and Dr. Bevra Hahn for their review and comments on the manuscript, as well as the respondents who participated in the study.