Retaining Clinician-Scientists: Nature Versus Nurture

Abstract

In their IOVS article “Rejuvenating Clinician-Scientist Training” (published March 28, 2014), Balamurali Ambati and Judd Cahoon rightly point out the dearth of new clinician-scientists in ophthalmology. Within the context of their suggestions for increasing the number of successful clinician-scientists, they claim that the traditional MD-PhD training programs and K awards have failed to produce individuals who will carry on the important work of clinically relevant research that will improve patients' lives and sight. In this response we present data, including information on the career paths of graduates of the Washington University ophthalmology residency, that call into question the presumed failure of MD-PhD and K award programs and show that, in fact, graduates of these programs are more likely to succeed as clinician-scientists than are their peers who have not trained in such scientifically rigorous environments. We propose that, rather than a failure of early training programs, it may be obstacles that arise later in training and among junior faculty that prevent promising careers from reaching maturity. Funding, one rather large obstacle, takes the form of imbalanced start-up monies, less National Institutes of Health (NIH) funding awarded to young investigators, and study section composition that may work against those with clinically driven questions. We also explore the challenges faced in the culture surrounding residency and fellowship training. We agree with Ambati and Cahoon that there needs to be more innovation in the way training programs are structured, but we believe that the evidence supports supplementing the current model rather than scrapping it and starting over with unproven initiatives. The data on training programs supports the contention that those who have already made substantial investment and commitment to the clinician-scientist pathway through participation in MSTP or K training programs are the most likely to succeed on this career trajectory. To muffle the siren song of private practice and retain those best prepared for the clinician-scientist pathway requires additional investment as their careers mature through protected research time, mentorship, and advocacy.

The recent Perspectives article by Balamurali Ambati and Judd Cahoon, “Rejuvenating Clinician-Scientist Training” (IOVS, March 28, 2014), ¹ proposes some interesting hypotheses for the reasons behind a decline in the number of clinician-scientists in ophthalmology in recent decades. In particular, they suggest that the programs created by the National Institutes of Health (NIH) to foster clinician-scientists (the Medical Scientist Training Program [MSTP] and K Career Development Awards) have failed their mission, that the decision to pursue academics is financial, and that the key to creating clinician-scientists is to make the experience financially competitive with private practice. They present alternative learning plans for both residency and fellowship programs that would address the ills they delineate by eliminating the expensive training programs and instead creating shorter, cheaper routes to research success. Here we argue that it is not that the arduous nature of the career development has failed to create clinician-scientists, but rather that systemic problems (practical and cultural) inherent in the transition from trainee to independent investigator serve to deter those who are the best prepared to take on that role. 

Training Programs Work

As is the case with any logical argument, the truth of the conclusion rests on the validity of the premises. Ambati and Cahoon's assertion that the MSTP and K award programs are failures is not substantiated by the primary references they cite to support it. ^2–4  

MD/PhDs Are Still More Likely Than MDs to Go Into Academic Medicine

Logically, this makes sense. MD/PhD graduates have already made a substantial investment in this career path. Forty percent of National Eye Institute (NEI) K awardees have both the MD and PhD degrees, ³ despite MD/PhDs representing only 3.5% of all ophthalmology residents. ² These data clearly show that despite Ambati and Cahoon's contention that the MSTP has failed to create clinician-scientists, MD/PhD graduates are disproportionately fulfilling that mission. But, to be fair, all of the data are not that rosy. For example, while 19% of all graduates of MD/PhD programs choose private practice after residency training, 44% of MD/PhD graduates who complete ophthalmology residency go into private practice. These numbers suggest that it is not that the MSTP fails to produce clinician-scientists, but rather that the ophthalmology residency training environment fails to sustain MD/PhDs on the academic career trajectory. 

K Awards Increase the Probability of Subsequent R01 Funding

Ambati and Cahoon's contention that the K Career Development Award program has failed to produce clinician-scientists is also suspect. The data are very clear in showing that K awardees are nearly four times as likely as MD nonawardees to obtain subsequent R01 funding (28% vs. 7.2%). ⁵ In this regard, ophthalmology is not performing as dismally as Ambati and Cahoon would have us believe. While the K to R conversion for all comers at NIH was 28% from 1990 to 2005, ⁵ those with K awards at NEI had better success at 36.4%. ³ The K08 program in particular is of critical importance for MD trainees who desire a career in research and has fared better than the K23 program in subsequent R01 success (33% vs. 20% for MD candidates, respectively). The K pathway provides research training opportunities that are absent from a medical education. 

Formal Instruction in Research Is Critical to Success

It is important to realize that there is a fundamental cultural difference between MDs and MD/PhDs in terms of how they identify, approach, and solve scientific problems. MDs are never formally trained in the ways of science, whereas the PhD candidate is. This, in our opinion, is the biggest barrier to R01 funding to clinician-scientists and is why MD/PhDs continue to be more competitive for R01 funding. They are formally trained in how to conduct science, write grants, accept delayed gratification, and deal with the scientific failure that will occur in a research career. The K Career Development grants require such training as a key element of the award proposal and completion—training that may, or may not, be incorporated into ad hoc post-MD research programs like research fellowships. 

But If This Is What We Call Success, We Should Be Doing Better

Of course, the interpretation of whether the MSTP and K award programs are a success or a failure depends upon what you want to achieve. Do we focus on the 4-fold increase in the probability of getting R01 funding that having had a K award confers, or do we succumb to the negativity that a K award confers only a modest advantage to clinician-scientists in obtaining R01 funding (28% vs. 18% for all comers, including PhD scientists) ⁶ ? And, as Ambati and Cahoon state, the shortage appears to be worsening with time. ⁴ They highlight the 2006 to 2010 NEI cohort that has thus far had only one of 28 awardees obtain R01 funding ³ as evidence that the K system is failing. We propose alternate explanations for the significant dearth of R01 funding in this cohort. First, this dataset is likely incomplete, given that the average time from K to R01 has been 6.0 (K08) and 4.4 (K23) years. ⁵ Thus many in the group have not had adequate time to obtain an R01 given the long timeline for K to R01 conversion. The second and most obvious explanation is the precipitous decline in successful new R01 funding NIH-wide, from a high of 25% in 2000 to the current rate of 14% in 2013. ⁷ Lastly, the competition with established investigators who have extended their careers has also impacted the ability of young scientists to obtain funding. Between 1980 and 2010, the percentage of R01 grants held by investigators under 36 years of age dropped by more than 4-fold (>16% to 3%) while at the same time the proportion of R01s held by those over 65 increased ∼12-fold (<1% to 6%), such that retirement-age investigators now hold twice as many grants as those under 36. ⁸ This protracted timeline for obtaining R01 funding jeopardizes young investigators in a specialty where there are substantial financial pressures, on both the investigators and the institutions that employ them, to shift their efforts into more lucrative endeavors. 

Is It Failure to Launch or Aborted Flight?

Ambati and Cahoon have focused on ideas about how to increase the number of non-PhD–trained clinicians who participate in research in order to increase the number of future clinician-scientists. They suggest that replacing existing, validated programs with these approaches will increase the numbers of clinician-scientists in ophthalmology. When done correctly, this may be true. The Specialty Training and Advanced Research program at the University of California Los Angeles (UCLA-STAR) is designed similarly to what Ambati and Cahoon propose except that it endorses attaining an advanced research degree (MS in clinical research or PhD), which is earned in concert with extended residency or fellowship training. The outcomes have been excellent, with 77% of the graduates staying in academics (similar to outcomes for MSTP). However, other specialties that have tried to lure clinician-scientists with extended research tracks have found this route to be disappointing. First, they have difficulty filling their slots as there is simply not enough interest in this training path. Second, because of the lack of interest, the applicants who do apply are less competitive than in traditional residency training. Ambati and Cahoon would suggest that this is because this option is not financially attractive, and that is possible. We are skeptical that short-term financial incentives are sufficient to offset the chronic systemic barriers to long-term career success. 

We Fail to Keep Potential Clinician-Scientists Engaged

We propose that it is not the system of enticing ophthalmologists into careers as clinician-scientists that is broken; rather the problem is the way in which the existing system fails to support those who have demonstrated a commitment to that career path. Data from both the MSTP and K Career Development Award programs suggest that retention is a problem. MSTPs in other subspecialties stay in academics at much higher rates than in ophthalmology. What is it that we do to dissuade similarly trained candidates from pursuing a career that they are so well trained and suited for? From experience in recruiting potential clinician-scientists into residency at our program, we find that applicants are dissuaded from expressing an interest in doing research during residency, as this would in some way compromise their ability to learn the clinical material. They hear this advice from both mentors at their home institutions and faculty interviewing at various residency programs. Our philosophy, which we presume is shared by Ambati and Cahoon, is that the integration of research into residency is critical to paving the way to a future as a clinician-scientist. If the ultimate goal is to balance these elements in a future career, why not start honing those skills during residency where there are mentors to help guide trainees through the often difficult process? 

The First Step Is Admitting There Is a Problem

We will present data from our own institution, where we endorse the spirit of some of the methods that Ambati and Cahoon propose. Commitment to these principles has helped us to mentor a healthy proportion of our graduates, MDs and MD/PhDs alike, with and without K awards, into clinician-scientists. 

When a Commitment to Clinician-Scientists Is Espoused in a Residency Training Program, More Trainees Will Continue Along the Chosen Path

We have analyzed data from the residency training program at Washington University, looking at graduates from 1991 to 2009. In that time 100 residents completed the program. Their characteristics are outlined in the Table. Our data demonstrate a number of points. First, a healthy proportion of trainees (29%) will choose an academic career if supported along that pathway. Second, the vast majority of the graduates who have chosen academic careers have recently published in a peer-reviewed journal (93.1%), indicating an active participation in academic inquiry. Third, approximately half of the physicians in our program who chose academic medicine as a career also consider themselves “clinician-scientists” who are actively pursuing grant funding to support their research enterprise (14/29, 48.3%). Four have been successful in procuring R01 funding while the other 10 are actively pursuing NIH and other external funding to support their research programs. This number is only slightly higher than the funding rate at NEI, suggesting that the limiting factor has less to do with motivation and more to do with the current grant review process and funding crisis. Additionally, two MD/PhD graduates have “clinician-scientist” careers (balancing research with patient care) outside the academic environment. One works in industry with a primary focus on research but with some patient care responsibilities. The other has obtained grant support to conduct outcomes research within the Kaiser Permanente system that would be nearly impossible to conduct in a traditional academic practice because of his ability to access the vast clinical dataset that the Kaiser system supports. While these cases would be considered “failures” using Ambati and Cahoon's definition of obtaining R01 funding, we would argue that success defined in this narrow way dismisses legitimate research endeavors that have potential to advance clinical medicine in direct and substantive ways, and possibly in areas that are untenable in the traditional academic route (i.e., outcomes research). 

Table

 

Career Choices of Washington University Resident Graduates 1991–2009

	MD Only	MD/PhD	Total
No. graduating	78, 78%	22, 22%	100, 100%
No. in private practice	56, 71.8%	11, 50%*	67, 67%
No. in academics	19, 24.4%	10, 45.5%	29, 29%
No. in industry	0, 0%	1, 4.5%*	1, 1%
No. with publications since 2012	16/19, 84.2%	10/10, 100%	27/29, 93.1%
No. with R01	0/19, 0%	4/10, 40.0%	4/29, 13.8%
No. “clinician-scientists”	4/19, 21.1%	10/10, 100%	14/29, 48.3%

*  One each in private practice and industry are active in both clinical research and patient care.

We Need to Rethink How We Nurture Clinician-Scientists in Ophthalmology

The current trend in science and medicine, and espoused by Ambati and Cahoon, is to increase the numbers of trainees who enter the clinician-scientist path. Our colleague Steve Petersen at Washington University, who has mentored many trainees in his career including clinician-scientists, cautions that cheap and short programs aimed at increasing numbers of potential clinician-scientists in the pipeline are analogous to insect reproduction. You make many, many offspring knowing that most will not survive. A number of them will. But each offspring is nearly identical. Individuality is not part of this process. It is all about volume. Period. What the MSTP and K Career Development Awards set out to do is more similar to primate reproduction. You have only a few offspring but invest heavily in their development. Each has unique qualities, strengths, and ideas. These individuals are then fostered and developed to make them the best suited not just to survive, but to change the world when they reach independence. The problem is that the system stops short of maturity. Clinician-scientists receive less start-up monies than PhD scientists at the beginning of their careers. ⁹ So from the beginning they are handicapped both financially and by limitations on their time owing to the nature of their work. And their work is demanding. Clinical departments generate revenue by seeing patients and doing surgery. So the pressure on clinician-scientists to add clinical time to their schedules is immense. The NEI does not participate in the K02 award program that is available through other institutes, including the National Institute of Neurological Disorders and Stroke and the National Institute of Mental Health. These are “salary only” awards that provide no research support but protect scientist's research time by providing salary. The institutes see this as critical to protect physicians who would otherwise be pulled to the clinic to cover their paycheck during a critical time when they need to collect preliminary data and begin a new line of investigation different from that of their K award mentors. This is one mechanism to continue to foster the best-prepared and most well-trained candidates along the path to independence and R01 funding. 

Another issue is to address the cultural divide that exists currently in resident education. While we have had success in supporting clinician-scientists in training at Washington University, some large credit must be given to appropriate selection of resident applicants at the forefront. In this process we participate in what we call “re-education” of applicants with clinician-scientist aspirations. Many applicants readily admit to having been coached and cautioned to avoid any talk of research commitment during residency training and to accept the view that residency is the time to focus on clinical medicine and not be distracted by science. This is particularly true of the MSTP applicants, who are often visibly relieved when we explain that the philosophy in our program is that striking a balance between clinical responsibilities and research demands is lifelong and the place to start that process is in residency. In residency there are mentors and systems in place designed to help, not hinder, attaining that goal. Just as Ambati and Cahoon suggest, a restructuring of the residency training environment is needed to interweave research and clinical training. To promote creativity in resident education, the Accreditation Council for Graduate Medical Education (ACGME) and Ophthalmology Residency Review Committee (RRC) should support an “innovative program” such as that existing for anesthesiology ¹⁰ to validate and encourage this career trajectory. 

A commitment to including research as a fundamental element of ophthalmology residency will also help to balance the exposure that residents have to clinician-scientists more generally. Even at our institution where residents have a dedicated research block, only 6% of the 36 months of training is devoted to demonstrating the importance of an academic research career. In contrast, 66% of resident training includes experiences with busy surgical mentors who have well-remunerated lifestyles and high job satisfaction, with the instant gratification that surgical outcomes often produce. It is a tough battle to convince trainees that the clinician-scientist pathway is a good and wise choice when residents spend more time with surgical mentors, some of whom are known to question the clinical competency of “lab rat” clinician-scientists, belittle the “monkhood” of their geeky academic counterparts, or render accusations of laziness or a life of leisure for clinician-scientists who spend time in the lab as opposed to the clinic. These seemingly offhand comments come in a steady stream to trainees and undermine the cause that we as a profession are trying to promote. It is a major part of the cultural problem that needs to be addressed. 

We Don't Have to Throw the Baby out With the Bathwater

Here we argue that it is not that NIH training programs fostering clinician-scientist training have failed, but rather that systemic attitudes about how to conduct science as a clinician, ineffectual processes by which the NIH reviews and awards grants, and a cultural divide within ophthalmology have conspired to disillusion a generation of prospective clinician-scientists. Despite our disagreement with the validity of Ambati and Cahoon's premises, we agree in part with their vision for how to improve a failing system. Finding creative and sustainable ways to incorporate serious research experiences in residency and fellowship training, we believe, is critical to build on the initiative of motivated trainees. But we would caution against a strategy that supplants proven effective (albeit expensive) strategies with unproven methods that may serve only to exacerbate the problem. Instead, we should continue to nurture the trainees who choose traditional educational programs (MSTP and K08 programs) while at the same time exploring alternative pathways such as innovation in residency and fellowship programs that will benefit both traditional and nontraditional enrollees. Programs that are less effective in producing successful clinician-scientists (K23) should be scrutinized and perhaps adapted or eliminated. Support for new faculty in terms of mentorship, protected research time with salary support, and adequate start-up monies will make the clinician-scientist competitive with his or her PhD colleagues. Furthermore, the lack of representation of MD researchers on study sections must be remedied: It underscores how “peer review” is failing the clinician-scientist whose grant will likely be evaluated by nonclinician reviewers who may or may not have the critical insight to adequately assess the impact of the science. Taken together, interventions such as these could help sustain the clinician-scientists in a time when the application of their expertise is of critical importance. 

Acknowledgments

The authors thank Gregory Wu and Morton Smith for input and helpful commentary. 

Supported by an unrestricted grant from Research to Prevent Blindness, Inc. to the Department of Ophthalmology and Visual Sciences at Washington University and by the Alan A. and Edith L. Wolff Distinguished Chair Fund to TM. 

Disclosure: S.M. Culican, None; J.D. Rupp, None; T.P. Margolis, None