The physician-scientist: An endangered species?

After diving into a heapin’ helpin’ of sheer craziness over the last week or two (well, except for yesterday, when I deconstructed an acupuncture study, which, while not crazy, certainly was misguided), I think it’s time for a bit of self-absorbed navel gazing. After all, isn’t that what bloggin’s all about? Oh, wait, that’s what I do almost every day here. No, what I really mean is that I came across an article that struck rather close to home regarding my career trajectory. So, if you don’t mind, for one day I’ll leave behind the rabid anti-vaccine loons, the homepaths, the alt-med mavens, and other assorted cranks and look at a problem that I’ve written about before but feel in the mood to discuss again. The impetus for this is an article by Andrew Schafer that was published in Science Careers entitled Perspective: The Successful Physician-Scientist of the 21st Century:

Physician-scientists have always brought a unique perspective to biomedical research that is inspired by their personal experience in caring for patients. Indeed, throughout history, physicians have played a central role in advancing the science of medicine as the “translators” of medical research. Yet there has been growing concern over the past 3 decades that the workforce of physician-scientists, at least in the form we have come to know them in previous generations, may be vanishing.

We have a problem.


When I first read this paragraph, my first thought was that, ever since I entered medical school back in the mid-1980s with the intention of becoming a physician-scientist, I’ve been told “we have a problem,” that there aren’t “enough” physician-scientists. Don’t get me wrong. As much as it may annoy some of the basic science researchers who read this blog, I agree that it is true. Physicians do bring a unique perspective to biomedical research compared to basic scientists. That is not in any way to denigrate the contributions of basic scientists in the least, as (I hope) readers will see. The contributions of physicians and scientists to biomedical research, in an ideal world, should be complementary.

I’ll give you an example that just happened a while back. I was discussing research with a basic scientist, someone whose work is quite good and with whom I was thinking of collaborating. He was quite enamored of a model he had developed of early events in breast cancer cell metastasis. I thought it was a really cool model, too. Unfortunately, I knew it was also clinically mostly irrelevant. Fortunately, I was able to put him on the track of a more clinically relevant model.

Just the other day, something similar happened. I was the only clinician at a meeting of basic researchers, when a point came up about how interesting it was that so many tumors occurred in the upper outer quadrant of the breast. I pointed out that this observation is not really considered striking among surgeons and in fact that the issue has been studied. There happens to be more breast tissue in the upper outer quadrant, and the increased incidence of upper outer quadrant cancers is most likely consistent with that observation.

None of this is meant to be bragging or denigrating the knowledge base of basic scientists. Believe me, as an MD/PhD, I’ve seen it go the other way, where basic scientists (or I) have had to correct an overzealous clinician who thought he had made a new finding and figured out how to investigate it, not knowing that they were treading well trod ground. Collaboration is one way to overcome this. The problem is that, the way academic medicine has been going, there is a perception that there are fewer and fewer physicians who are carrying out translational research who can collaborate with these basic scientists. At least, there are some worrying trends:

There is ample evidence to support this worrisome trend. Although the numbers of National Institutes of Health (NIH) grant applications and applicants over the past 15 years has more than doubled, those numbers have been essentially flat for M.D.-only physician-scientist applicants. During the 5-year period from 1998 to 2003, during which the NIH budget doubled, there was a 43% increase in first-time R01 applicants with Ph.D.s as principal investigators (PIs) and a 104% increase in applications with M.D.-Ph.D. PIs — a very small percentage of the total pool of applicants. In contrast, applications from those with M.D. degrees declined by 4%.

Note that this is referring to physicians with MDs alone who do research of sufficiently high quality to garner NIH R01 funding. Worse, as Dr. Schafer points out, physicians are much less persistent in resubmitting grants after the inevitable rejection of the first application. Anecdotally, my experience bears this relative decrease in MDs doing research, particularly in surgery. Back when I was a medical student and resident, not only did I meet and get to know several surgeons without PhDs who had their own labs and did research, I worked with a couple of them. These days, I know only a couple, only one of whom can be described as continuing to be successful at it while still maintaining a clinical practice. All the other MD-researchers who have remained successful that I know have given up clinical practice. Meanwhile, at academic surgery conferences that I attend, a constant theme is the lament about how difficult it is to get a career in research, how few surgeons–and physicians in general–want to do research anymore.

Of course, as I’ve alluded to in the past, the forces arrayed against even the most dedicated physician wanting to do research are considerable. Reimbursements from third party payors and government sources continue to be ratcheted down. As a result, the pressure to do more and more clinical work just to carry one’s own weight continues to be a serious problem. As much as medical schools would seem to like to believe that clinicians will spend their nights and weekends doing nothing but extra work, such an expectation is not realistic.

One observation that Dr. Schafer makes that is particularly accurate to me is this is when he speculates as to how it might have happened that there are decreasing numbers of physician-scientists and that at every step in the early life cycle of NIH grant funding they tend to disappear from the pool:

At the core of it, I think, is the reality that the arenas of basic biomedical research (on one side) and the clinical practice of medicine (on the other) have progressively and dramatically separated. This widening chasm has created a rapidly increasing language barrier between basic biomedical scientists and practicing clinicians. It is a two-way barrier: Midcareer clinicians today are unable to understand even the basic vocabulary of molecular biology and genetics, and biomedical investigators (even those with M.D. degrees) are increasingly losing track of rapid advances in clinical medicine, which is always increasing in technologic complexity.

And it’s true. Physicians who don’t stay continually active doing research and interacting with basic scientists very rapidly fall behind, to the point where they no longer even understand the basic science. I must admit, however, that I hadn’t considered so much that this increasing distance might be due to movement on both sides, but now that I think about it it does seem to ring true. Most basic scientists aren’t aware of advances in how we clinicians treat, for example, cancer.

I’m not sure I entirely agree with Dr. Schafer, though, when he also blames this increasing chasm on the “he reductionism in medical research in the early years of the molecular biology and genetics revolution throughout much of the second half of the 20th century.” I do see his point, though. It is true that, at least during most of my medical career, science did move in an increasingly molecular direction. It’s also true that, from my perspective, in the 1980s to 1990s it seemed that very few basic scientists did anything resembling whole organ or whole organism physiological studies, preferring instead molecular biological approaches. This tended to leave studies of whole organ physiology mainly to physician-scientists. To put it simplisticially, basic scientists dealt with genes and proteins; physicians would look at macroscopic phenomena, like blood flow and grose measures of metabolism. What I don’t see happening is Dr. Schafer’s prediction that systems biology will help bridge the chasm between clinicians and basic scientists.

It may well be that systems biology may serve to nearrow this gap, but I wouldn’t count on it. The reason, of course, is that systems biology requires ever more advanced molecular and mathematical models in order to construct the signaling networks based on experimental data. If anything, it’s molecular biology on steroids, requiring specialized mathematical and statistical calculations, not to mention a whole lot of computing power.

One characterization of this chasm that I do (mostly) agree with comes from Barry Coller:

  • Clinicians are motivated by the need for immediate action (sometimes to even save a life), whereas scientists are conditioned to avoid rushing to judgment;
  • Clinicians are taught to adhere to standards and guidelines of practice, whereas scientists are encouraged to challenge existing paradigms;
  • Clinicians traditionally respect hierarchy and expert authority, whereas scientists tend to critique and challenge accepted wisdom;
  • For clinicians, errors are potentially mortal threats, whereas for scientists, errors are inevitable manifestations of the creative process;
  • Clinicians focus on the unique, whereas scientists look for generalizable principles.

The only part of this I tend to disagree with is the last principle. Clinicians don’t just focus on the unique; they are taught from the very beginning of medical school to look for patterns. The real difference is that this pattern recognition isn’t always systematic. Often it works on the level of an overal “gestalt.” If there’s one thing a good clinician can do, it’s to tell when a patient “looks sick.” The best clinicians can walk into a room and at a glance have a good idea of just how seriously ill a patient is, and the very best surgeons can often identify peritonitis before they even lay a hand on the patient. Of course, this is just as much a cultural difference between physicians and basic scientists, the latter of whom rely on experimentation, hypothesis testing, and observation to come to their conclusions. It’s also sometimes a serious problem in that physicians without training in the scientific method all too often let their pattern recognition skills lead them astray into confusing correlation with causation and confusing placebo effects and regression to the mean for real responses to therapy. I can’t help but speculate that this is one reason why physicians tend to be more prone to woo than they should be.

So what does this all mean?

A lot of this strikes me as a case of “everything old is new again.” As I mentioned at the beginning, I’ve been hearing about the imminent demise of the physician-scientists, well, ever since I started training to become one. Sometimes I have to wonder whether some of this rhetoric is, either consciously or unconsciously, designed to reinforce a sense of importance in the endeavor. On the other hand, it is important to be able to translate basic science findings into clinical practice. What we may well be observing is not so much the demise of the physician-scientist, but its evolution. There used to be a time when a person with an MD could pick up research skills by spending a couple of years in a laboratory under the tutelage of a research mentor and end up with the research skills necessary to be competitive for NIH funding. That time is no more. Consequently, what we are seeing now is the rise of the MD-PhD as the preferred physician-scientist, leaving the MDs to take care of the patients. The problem is that there aren’t a lot of MD-PhDs, and there likely never will be. Few people are dedicated or crazy enough to get both degrees, go through a demanding residency, and then try to compete with basic scientists for grant funding. Sometimes I don’t know if I’m crazy enough to do it anymore.

Perhaps the most relevant observation made by Dr. Schafer is this one:

Learn that medical research today is a team sport. During this generation, the breathtaking pace and scope of progress in both the science and the practice of medicine has vastly outstripped the capacity of any individual physician-scientist to maintain even a semblance of currency in both arenas. The key to success is your ability to thoughtfully surround yourself with partners, particularly Ph.D. scientists, who offer complementary expertise. And for each project, you and your collaborators should try to agree in advance what each scientist’s role will be and who will be the “driver.” Insisting on being the sole principal investigator on every project, or even most projects, will prove to be counterproductive. Can you be a team player and still be a star? Yes. Remember that the greatest sports stars have been the ones who were able to elevate their teams to win championships.

And that, above all, is the major change that has occurred over the last 20 years. Solo investigators running their own little labs are going to have an increasingly difficult time suriving, and that goes for both basic scientists and clinicians. Both are needed to maintain the pace of progress of science-based medicine. If one, the physician-scientist, is endangered, it’s going to be very hard to figure out how to apply all that fancy science, either reductionistic or systems biology, to improve the care of actual patients.