Fostering the careers of physician–investigators has been a collective objective for many of the major forces in academic medicine, including schools of medicine, the National Institutes of Health (NIH), the National Academies, and the Association of American Medical Colleges.1 There is the general perception that well-trained and ambitious physician–investigators leave the field because of a variety of forces, like the prolonged training period required, the lack of institutional understanding of the physician–investigator phenotype, competition for funding from full-time PhD investigators, the lure of clinical work, and the fact that clinical salaries generally exceed research salaries, among other reasons. While these reasons probably all contribute to physician–investigators leaving the field to some degree, there are other important issues that we must resolve first to characterize and solve this problem.
One of the first issues to resolve is the need to better define the expectations for individuals with an MD degree who plan to do research. For discussion purposes, I will define three paradigmatic types of physician–investigators, recognizing that there are actually more variations and that within these categories there is considerable heterogeneity. First is the individual with an MD degree only who has clinical duties but expects to do clinical, patient-centered research. For definition purposes, I will label this category as the “clinical researcher.” The second category includes MDs and MD–PhDs who have a clinical role but perform their research in a laboratory or using computational tools. I will label this category as the “clinician–scientist.” The last category is the MD or MD–PhD who is focused on research and has little or no clinical activity. I will label this category as the “physician–scientist.” Overall, I will use the term “physician–investigator” to describe all three of these types of physicians who expect to do research.
The Challenges of Funding a Research Career
All physician–investigators face common problems, but there are some issues that vary by the focus of the investigator. One primary shared concern is funding. The NIH paylines have been hovering around 10% to 20%, and it is generally agreed that such low thresholds make it extremely difficult for investigators to build a career that will ultimately depend on predictable, sustained funding. The arbitrary qualities used to select which good-to-excellent science is funded make the use of grants as promotion criteria cruelly unfair. The high probability of having difficulty maintaining funding also encourages investigators to cut corners and exaggerate the significance of marginal findings, making the issue of reproducibility more of a concern. Looking at the prospect of trying to fund a career while depending on such an arbitrary process is very unappealing and one cause for physician–investigators to leave the field. This challenge will affect clinician–scientists in particular, as they may not be able to fund their labs and thus often transition to a more clinical role. I believe that the perceived competitiveness of obtaining funds is less of a problem for physician–investigators than the arbitrariness of the process, since physicians generally have strongly competitive instincts, but they will not want to compete when the prizes appear to be given out by a lottery.
For the clinical researcher and clinician–scientist, there is also the high probability of an imbalance between the funding she or he can receive from research activities and from the clinic. In general, clinical work is better compensated and calibrated to the volume of work (relative value units), while pay for research work is based on a percentage of professional effort and the salary proffered by the grant. For NIH grants, this is currently Executive Level II or $187,000/year. Most universities use institutional funds to fill the gap between the faculty member’s salary and the amount she or he receives in grants. For clinical researchers and clinician–scientists, the pressure may be to increase their clinical volume to fill this gap.
The Appointments, Promotion, and Tenure Problem
The second challenge facing all physician–investigators is an appointments, promotion, and tenure (APT) system that is not very responsive to the needs of hybrid faculty who work across clinical care and research. When investigators split their time between clinical activities and research, they are not likely to fully meet the target metrics of either full-time researchers or full-time clinicians. Too many schools do not place an explicit value on faculty members’ ability to take on both roles, a problem which particularly hinders clinician–scientists.
For clinical researchers another challenge is to demonstrate their productivity to the APT system. Clinical research studies often take longer to accrue and complete, with fewer opportunities to publish, than basic science work. In addition, clinical and translational research often involves study teams, and many institutions have not adjusted their promotion criteria for the inclusion of team science.
The Growing Time Burden of Clinical Work
A third pressure affecting both clinical researchers and clinician–scientists is the increasing burdens of clinical activity. The addition of the electronic medical record makes clinical work more time consuming, typically doubling the amount of time required per patient. This added time burden makes the clinical work–research balance even more difficult to maintain. Add the time pressure from needing to submit multiple grant applications to have even one succeed and the life of an MD–investigator can become an unhappy union of physician and keyboard.
The Need for Data
While we can identify a litany of challenges to maintaining a career as a physician–investigator, regardless of the specific category of investigator, we lack data regarding what is actually happening to research-focused physicians. The extent of these difficulties is not well measured. For example, what proportion drop out at various stages of their career: training, early faculty, midcareer, retirement? What motivates this change? While the issues that cause an individual to move in or out of research are likely specific to that individual, analyzing trends will be important to develop a strategy to ameliorate the actual problems.
Once data are obtained about physician–investigators, it also will be important to define expectations. For example, students who enter the MD–PhD pathway are often very young, typically in their early 20s. While a research career may be appealing early in life, with maturity it may lose its charm. What is a fair rate of attrition for MD–PhDs? At what stage will this career shift occur? If a laboratory-trained MD begins her or his career at the bench but finds the clinic more appealing, is that a failure or a result that can be anticipated? I would argue that it is an expected outcome. If the motivation for this change is purely financial, however, it is also a reason to consider how to make the research portion of an investigator’s career more satisfying.
New Proposals and Perspectives
This issue of Academic Medicine includes two articles addressing the career challenges faced by physician–investigators. Hall and colleagues2 addressed the issue from the point of view of the NIH. They offered three potential strategies for increasing the size and quality of the physician–investigator pool. Their first strategy is to support research in residency to shorten the period of postgraduate physician training and to allow residents to continue to conduct research during their clinical training. The American Board of Internal Medicine has demonstrated that their research residency program appears to work, so this is a promising avenue, particularly if the NIH can find ways to fund pilot studies to demonstrate its worth.
Hall and colleagues next suggested that there should be multiple on-ramps for a research career. Students should not have to decide at 21 if they want to spend their career conducting research. Strategies like encouraging fellows to obtain a PhD or master’s degree as part of their postresidency training could be a reasonable alternative to the combination of an MD–PhD program and a research residency. It would trade maturity and a high degree of motivation on the part of the trainee for a somewhat longer training process. Finally, Hall and colleagues believed that there should be more networking among physician–investigators so that best practices can be shared across institutions and so that this group does not feel so isolated.
In their article, Lingard and colleagues3 proposed two broad approaches—organizational and individual strategies—to supporting physician–investigators in faculty roles. Similar to Hall and colleagues, they stressed the importance of fostering community. They also argued that there should be better mentorship, that organizations should establish the value of physician–investigators, and that the financial penalties for pursuing a career in research should be minimized. At an individual level, they encouraged physician–investigators to develop their professional and research skills. They also identified key gaps that need to be addressed. For example, what is an appropriate amount of protected time for research? Do we have enough data to justify system changes to improve the lot of physician–investigators, or are we operating on hunches? And is our sense of the identity of a “physician–scientist” a clear and coherent one?
I believe that protecting the physician–investigator role is critical for medical progress, but we should be aware that we are dealing with quite different subpopulations. The clinical researcher may need less protected time for research and have fewer challenges with funding, given that clinical research and care can often be integrated. The clinician–scientist is in the most difficult situation, torn between two frequently hostile worlds. And the physician–scientist, who does not have the same clinical challenges as the others, still needs to compete with a pool of PhDs and other investigators, most of whom will have completed their training and begun their NIH-funded career at a younger age.
To solve the dilemma of how to best support our physician–investigators, as a community we should be encouraged by the notion that the NIH is willing to consider funding pilot programs that could yield pragmatically useful data. In general, we need more information about what is currently happening to physician–investigators, and then we need pilot studies that try to make this career path a little less daunting.