Medical school admissions committees and postgraduate education selection committees in Canada often assume that applicants with PhD degrees will pursue research careers after graduation with an MD degree or completion of specialty training. Such applicants may receive extra points in a competitive ranking system because of this assumption in institutions wishing to develop research-oriented graduates. Although many physicians with dual degrees (MD and PhD) pursue careers in research after specialty training, many do not. To our knowledge, little has been written about the career paths of these physicians.
In a survey of 7400 internal medicine faculty, 512 of these had both the MD and PhD degrees; 55% of the dual-degree physicians who responded (248/453) were engaged in significant research activity.1 This is compared with 35% of those with an MD degree only and 84% of those with a PhD degree only. An accelerated MD program for PhD students at the Washington University School of Medicine reviewed 15 years of the program and found that of those who had completed their training, only half (13/26) went into academic practice.2 A survey of family practitioners in a university department found that those with a PhD or a MD–MS were more productive than those with only an MD degree.3
Although having a PhD in addition to an MD seems to facilitate a physician’s research career over an MD alone, it also seems that the timing of earning these degrees may be a factor in later career plans. A U.S. Department of Health and Human Services, National Institutes of Health (NIH) report4 analyzed the Association of American Medical Colleges’ graduation questionnaire for MD graduates of 1980 and found that 95% of graduates from joint MD–PhD programs planned to do research. However, graduates who earned their PhDs before their MDs (instead earning them concurrently) reported plans for research careers at a lower rate of 53%. In a discussion paper, Gordon and Salmon5 suggest that in the British model, research training will lead to more success if it is done during or after specialist registrar training (residency) than earlier, but they have no data to support this contention. A survey of 262 doctors in the United Kingdom who were in research posts reported that 9% of the physicians decided on a research career before their MD program; 18% decided while they were in medical school, and 59% decided after qualifying as physicians.6
We are aware of only one study investigating the association between the research productivity of dual-degree physicians and the timing of their training. A survey of 43 academic dermatologists found that 12 (28%) respondents obtained their PhDs before medical school, 17 (40%) did so during medical school, and 14 (33%) did so after medical school. No differences were found in terms of research effort (percentage of time spent in clinical or laboratory research) among the groups.7
It is unclear whether the timing of the acquisition of the degrees is associated with a particular career path in dual-degree physicians. Do medical school admissions committees and postgraduate selection committees have realistic expectations of candidates with a PhD accepted for medical training? The purpose of our study is to determine whether the sequence of training to obtain both degrees is associated with a difference in the career profiles of these physicians. Furthermore, we explored the factors that encourage or dissuade Canadian dual-degree physicians in pursuing a research career.
Questionnaire development and distribution
After obtaining ethics review board approval from the University of Alberta, Edmonton, Canada, where the study was conducted, we identified Canadian physicians holding both a PhD and an MD through the database of the 2002 Canadian Medical Directory (MD Select). The medical directory provides demographic data such as year and place of MD graduation, gender, location, and type of current medical practice.
To obtain additional information about physicians’ practice profiles, we developed and distributed a questionnaire. The questionnaire asked participants to describe their academic work by choosing from answers provided about the sequence of obtaining their MD and PhD degrees; their current position held; the percentage of time spent on clinical, research, teaching, and administrative duties; their number of peer-reviewed publications in the previous three years; and the amount of grant support currently held. We did not enquire about the subject areas of participants’ PhDs. We also asked the physicians about their opinion regarding the significance of factors they considered disincentives and incentives to their own research productivity. Responses were indicated on a three-point scale of significance, where 3 = very significant, 2 = somewhat significant, and 1 = not significant. Finally, they were asked to provide written comments on factors affecting the paths of their careers.
We sent a pilot questionnaire to two physicians identified through the MD Select database. Based on their responses, no significant changes were made to the questionnaire. Questionnaires were then sent to all 734 remaining dual-degree physicians identified by MD Select in 2003 by either electronic mail or postal delivery according to information provided in the directory. We sent a second mailing to nonresponders to maximize the response. Each respondent was assigned an identification number. We analyzed the data from MD Select and from the questionnaires by number only, thereby maintaining confidentiality.
We used descriptive statistics to summarize the demographic characteristics of the responders. With respect to physicians’ practice profiles and factors affecting their research productivity, the respondents were grouped into one of three categories based on the timing of obtaining a PhD degree relative to obtaining an MD degree (PhD first followed by MD = early PhD; PhD concurrent with MD = concurrent PhD; MD first followed by PhD = late PhD). To determine whether differences existed between the three groups, we conducted an ANOVA using practice profile characteristics and research productivity factors as variables. Alpha was set 0.05 and Scheffé tests were used for multiple group comparisons.
We evaluated the open-ended responses using a directed approach to content analysis.8 The directed approach uses prior research to aid the development of the initial coding scheme before the data are analyzed; this coding scheme is refined with the addition of other themes that develop during analysis. The open-ended responses were coded separately by two of the authors (RK and SL) using predetermined themes from a literature review of factors influencing a research career; the most common of these were listed on the survey as disincentives and incentives (see Discussion). Additional themes (factors influencing career path) emerged during the analysis and were coded. Because of the breadth and depth of the survey comments obtained on this issue, we sought confirmation of our impressions with focus groups of respondents. Trustworthiness of the data was determined by debriefing between coders after the coding process and discussion of new themes, using rank order comparisons of the frequency of the themes to describe the data, and triangulating the data with focus groups. Finally, using a Pearson product moment correlation, we computed the agreement using a dichotomous approach between the authors for each derived themes. We used StatView version 5.0.1 statistical software for the analyses.
Respondents to the survey who were listed by MD Select as working in Alberta were invited to participate in one of two focus groups held at the University of Calgary and the University of Alberta in 2004. The focus groups were led by a trained focus-group leader and evaluation researcher and were attended by one of the authors (RK), who acted as recorder. The focus-group leader and this author reviewed the themes derived from the data to produce the questions used to direct the focus-group discussion. The focus-group discussions were tape recorded and transcribed. The qualitative data were coded, summarized, and analyzed; actual comments are included to further enrich and support the quantitative data.
A total of 473 (64.4%) physicians completed the questionnaire. The focus group at the University of Calgary had five participants, and the one at the University of Alberta had seven participants. The survey respondents’ demographic characteristics regarding the timing of obtaining their PhD degrees are shown in Table 1. Respondents are divided into three groups by sequence of degrees. Of these respondents, 62% (292) were in the late PhD group (58% received their MD degree from Canadian medical schools); 26% (121) were in the early PhD group (88% received their MD from Canada); and 12% (58) were in the concurrent PhD group (72% received their MD from Canada). Of the 62% in the late PhD group, 22% (64) are in general/family practice, 74% (216) practice in clinical specialties, and 4% (12) call themselves medical scientists.
Physicians in the late PhD group spend, on average, less time on clinical activities (43.5%) than those in either the concurrent PhD group (55.3%) or the early PhD group (70.8%). The late PhD group also reported spending more time on research (26.4%) than the concurrent PhD group (21.8%) or the early PhD group (10.5%). The late PhD group also spent more time on teaching (11.2%) than the early PhD group (6.7%).
With respect to research productivity, the physicians in the late PhD group had published more papers in the previous three years than either the early PhD group or the concurrent PhD group of physicians (α = 0.05) (Table 2). It is surprising to find that 36% of all respondents published no papers in the previous three years (60% of the early PhD group, 38% of the concurrent group, and 25% of the late PhD group). Similarly, 28% of all respondents published nine or more papers in that time (14% of the early PhD group, 26% of the concurrent group, and 34% of the late PhD group). Likewise, the late PhD group garnered more research funding than the early PhD group (α = 0.05) but not more than the concurrent group.
When asked about influences of various factors as incentives to do research, all three groups of physicians identified that their previous experience in research and the intellectual challenges posed by research were the most noteworthy influences. Availability of funding for research, having a mentor, being interested in teaching, and preferring an academic environment were also seen as somewhat important influences. There were no differences between the groups regarding factors identified as incentives. Whereas all groups identified lack of time and resources as the major disincentives to research careers, the only difference we found between groups was that physicians in the early PhD group were more likely to identify family concerns as a disincentive for a research career than were the physicians in the late PhD group. Other factors were seen as less noteworthy: financial considerations, dissatisfaction with academic life, and job security. In the responses to the survey’s open-ended questions concerning factors that affected their career paths, many physicians emphasized the incentives and disincentives listed earlier in the survey, but they identified other factors as well. The three most frequent disincentives, cited by more than 50 respondents, included lack of time for research, lack of resources for research, and financial concerns. Between 20 and 30 respondents cited each of the following disincentives: family responsibilities and preference for more clinical time. Other disincentives noted less frequently were poor job security, dissatisfaction with academic life, training (too long, lack of positions), problems related to being international medical graduates, and feeling unable to be competitive with PhDs for grants.
Between 20 and 25 respondents cited each of the following incentives to a research career: having mentors and collaborators, the intellectual challenge and outlet for creativity provided, and the advantages of dual degrees with respect to additional life skills gained. Approximately 10 respondents cited the ability to link clinical practice to research as an incentive for a research career. There was good agreement between the authors (RK and SL) on the identification of these factors (r ranged from 0.665 to 0.713).
The focus-team leader and one of the authors (RK) reviewed the data obtained from the surveys. The data were grouped into the following themes:
1. the great difficulty in developing a research career while being a clinician;
2. the lack of support and respect from pure PhD colleagues, leading to a lack of competitiveness for grant funding;
3. lack of support from clinical colleagues, deans, and chairs; and
4. despite challenges, the great intellectual satisfaction of research.
These themes were used to prepare the questions for the focus groups. In the physicians’ focus groups, participants were asked whether they agreed with the interpretations of their survey responses, and whether they could recognize themselves and their concerns in these interpretations. The following summarizes the common discussion points from two focus groups conducted with physicians who hold both an MD and PhD.
Participants in both focus groups agreed with the difficulty of developing a research career while being a clinician. They gave illustrations of the challenges of balancing the demands of research with clinical practice, teaching and administrative responsibilities, and family life. These were cited as significant factors in the survival of the research clinician. Commenting on the priority of clinical work, one focus-group attendee said:
But of course clinical work always supersedes anything that one has to do with research.
All respondents articulated barriers to research productivity, including the time to adequately prepare research grant applications for the competitive process against full-time researchers. Clinical scientists indicated that they felt disadvantaged compared with full-time researchers in the funding process itself, as illustrated by these comments:
I look very enviously upon my PhD colleagues, with whom I compete head to head in research grant competitions, who can spend two solid months polishing and doing their grants and carrying out their research.
I just don’t have the same time to devote to polishing [the grant proposals].
Both sets of focus-group participants provided a number of comments and suggestions to express the need for increased support from clinical colleagues, deans, and chairs. Salary support, protected time for research activities, and creating a critical mass of clinical researchers were most often identified. Mentors, support staff, lab facilities, child care support, and a more thorough orientation to research and clinical facilities were also identified as needed supports. Frustration with competing clinical and research responsibilities was stated by one attendee:
I mean some of my research time—supposedly I have 50% [research time], but half of it is after I’ve been on call for 24 hours, and I just can’t think then.
Although both intrinsic and extrinsic factors were cited as motivators of research, intellectual stimulation or intellectual satisfaction seemed to underlie these contributions to both sets of focus-group attendees. Although the income disparity between clinical work and research positions was repeatedly stated as a significant disincentive to pursue a research career, several respondents reported that research was “in their system.” One attendee described his intrinsic desire to pursue research as follows:
If you’re the sort of person who’s going to do to a PhD and do research and do that, you’re an inquisitive person. … I can’t leave something until I find out how it works … as a kid I took apart every watch in the house, including my dad’s Rolex … to find out how it worked and why one was better than another and so on. It’s an ingrained quality I guess.
The pursuit of this research passion was generally described as financially undervalued but intellectually rewarding and a powerful motivator for a research career.
In this study, we found that dual-degree Canadian physicians who obtained a PhD after an MD degree have more research-focused careers than those who either obtained their PhD concurrently with the MD or achieved a PhD before medical school. Our study supports the findings of the NIH study4 of the intentions of graduates (class of 1980) from MD–PhD programs to do research. It also supports the empirical findings of Wakeford et al,6 who found that most UK researchers became interested in a research career after obtaining their MD degree. Our findings differ from those of Prystowsky,7 who did not find a difference attributable to sequence of training among dermatologists, a finding that may be attributed to the selection of only academic-based physicians. Our study has the advantage of large sample size and very good response rate. MD Select is believed to be the most comprehensive listing of active Canadian physicians. In addition, the year of graduation of the sample spans five decades and is broad in terms of practice specialty and location of practice in Canada.
In Canada, MD–PhD programs are available at 10 out of 16 medical schools and produce markedly fewer graduates than in the United States. The only published review of these programs from 1997 found 51 students enrolled across Canada and 15 graduates from previous years. There is no equivalent to the medical scientist training program in Canada. A clinician investigator program organized through the Royal College of Physicians and Surgeons of Canada has approximately 100 postgraduates enrolled in any one year. These students complete at least two years of research embedded within their residency training, but not all proceed to obtain the PhD.9,10
Measures of research productivity have been addressed by several studies. For the family practitioners, the following criteria were used: publications, external presentations, and grants.3 Beaty et al1 determined that percentage of effort devoted to research and number of publications were crucial to the definition of research productivity but that it also included space for research and funding. For radiologists, the criteria were number of total publications, published abstracts, and presentations.11 The criteria we used were in line with these studies: the current position held (clinical, research, other); the proportion of time spent on clinical, research, teaching, and administrative duties; the number of peer-reviewed publications in the previous three years; and the amount of grant support currently held.
Whereas the timing of research training seems to be important in the development of a research career, we sought to determine other perceived barriers and incentives to such a career. Participants in the focus groups provided more detailed information that enriched the data collected by our survey. They supported the findings of the survey; together, both study components provide a thorough exploration of the factors that encourage or dissuade MD–PhDs in the pursuit of a research career in Canada.
The focus-group participants agreed with the difficulty in developing a research career while being a clinician and gave illustrations of the challenges of balancing the demands of research with clinical practice, teaching and administrative responsibilities, and family life. These were cited as significant factors in the survival of the research clinician. All respondents articulated barriers to research productivity and identified the lack of time and resources for research as well as personal financial concerns as chief barriers. Those having a PhD before medical school were more concerned about the effect of a research career on their families. These participants emphasized a lack of time to adequately prepare research grant applications for the competitive process compared with full-time researchers. Clinical scientists indicated that they felt disadvantaged in the funding process itself, because their clinical research interests and publication record are compared with those of full-time researchers who do not carry a clinical workload.
Finally, the focus-group participants provided a number of comments and suggestions to express the need for increased support from clinical colleagues, deans, and chairs. Salary support, protected time for research activities, and creating a critical mass of clinical researchers were most often identified. Mentors, support staff, lab facilities, child care support, and a more thorough orientation to research and clinical facilities were also identified as needed supports.
These obstacles are strikingly similar to those identified by other authors: inadequate protected time, maintaining research productivity, inadequate/ uncertain funding, difficulties finding knowledgeable colleagues, insufficient support services, personal and family obligations, pressures to teach, getting worthwhile research ideas, pressures to do clinical work, length of training, research hiatus for clinical training, debt, lure of greater salaries in private practice or pharmaceutical companies, hurdles of clinical trials versus basic science research, lack of mentors, lack of job security, and inadequate organizational infrastructure.12,13
Although both intrinsic and extrinsic factors were cited as motivators of research by our respondents, intellectual challenge and creativity and having previous research experience seemed to be most important. Although the income disparity between clinical work and research work was repeatedly stated by focus-group participants as a significant disincentive to pursuing a research career, several respondents reported that doing research was innate. The pursuit of this research passion was generally described as financially undervalued but intellectually rewarding. Other incentives to research noted in our study were working in the academic environment, the presence of mentors, the availability of funding, and having an interest in teaching. All these factors have been noted by others and were thought to influence an academic career.6,12,14
How might students be encouraged to pursue research careers in light of a formidable list of barriers? Many authors have proposed solutions.13,15 Even the trainees themselves have addressed this issue at a recent meeting of the Royal College of Physicians and Surgeons of Canada and the Canadian Society for Clinical Investigation.16 Although a discussion of this is beyond the scope of this article, solutions are directed primarily to multiple models of MD–PhD training programs, diminishing student debt, and mentoring.
There are several limitations to this study. Although the producers of MD Select attempt to include all practicing physicians in Canada and to keep their information up to date, they do note that the accuracy of their information depends on the cooperation of the physicians listed. Our survey did identify physicians who had retired or moved out of the country. The database, therefore, may not have included all physicians in active practice in Canada. In addition, there may be inherent biases in the listing of physicians. Although the response rate was good, biases may exist in nonrespondents, such as disillusionment with research and academic environments, or simply being too busy to respond. The focus groups were held at two Alberta universities, therefore selecting for physicians in academic practice in the two major urban centers in Alberta. The experiences of these physicians may differ from those of their rural counterparts and from the rest of the country. We recorded data regarding research productivity by way of ranges of papers produced and ranges of grant amounts held. This approach did not allow for the determination of means of the actual numbers. The magnitude of the true differences between groups may be greater than that reported for papers published, and a real difference between groups regarding grants held may exist.
Dual-degree physicians who obtain their PhD late in training, after their MD degree, have a more research-focused career in Canada than those who obtained their PhD either concurrently with the MD or before the MD. Despite the commonly held assumption that PhDs entering medical school will continue a research career, it seems that most of these applicants plan career changes from PhD researchers to medical clinicians because they desire a more clinical career. Selection committees for medical school admissions and residency positions should take note and should adjust their admission criteria accordingly. The qualitative research provided a greater understanding of the factors that encourage or dissuade dual-degree physicians in the pursuit of a research career in Canada. Participants in the physicians’ focus groups provided considerable agreement to the interpretations of the survey data. Focus-group participants made numerous insightful comments and suggestions, which could have important implications for researchers, funding agencies, and institutions.
The authors wish to acknowledge the contribution of Myrna Sears, MA, MDE, who was the focus-group leader for this study, and to thank those physicians who participated in the focus groups. The study was funded by the office of the dean, Faculty of Medicine and Dentistry, University of Alberta, and by the Alberta Heritage Foundation for Medical Research.