Abstract

The physician–scientist workforce is vital to the advancement of biomedical research and the practice of medicine. Physician–scientists are a unique link between basic science research and clinical practice that facilitates new discoveries in the understanding and treatment of human disease. However, the number of physicians reporting science as a primary career activity in the United States has remained stagnant despite an increasing number of total practicing physicians. In 2014, just 1.5% of all practicing physicians reported research as their primary activity. In addition, there are concerns that women and minorities are disproportionately underrepresented in the physician–scientist workforce relative to the overall physician workforce. The present review synthesizes the trends and concerns of many prominent voices in the physician–scientist community. Furthermore, it provides recommendations for interventions aimed at repairing the “leaky” pipeline that recruits and trains medical students for careers in research.

Introduction

Concerns over the state of the physician–scientist workforce were first brought to the attention of the research community in 1979 by James Wyngaarden, then professor at Duke University and later director of the National Institutes of Health (NIH), in a symposium addressed to the Committee on Medical Education of the New York Academy of Medicine. In his symposium, Dr Wyngaarden famously warned that the clinical investigator had become an “endangered species.”[1] At the time, Dr Wyngaarden was distressed by waning numbers of young MD and MD/PhD graduates involved in research careers. This early call to action would be echoed in the following decades by other prominent research.[2][5]

To date, the state of the physician–scientist workforce continues to be a source of great concern for many in the research community.[6][9] In their 2014 report, the NIH Physician–Scientist Workforce Working Group found that of the nearly 1 million physicians in the United States, only 14 000 (1.5%) reported research as their primary focus.[6],[10] In a review of the literature, Kosik et al identified more than 1200 peer–reviewed publications as of the year 2014 devoted to the topic of rescuing the physician–scientist workforce.[7] In 2008, the Association of Professors of Medicine published a special report proposing recommended actions for repairing the “leaking physician–scientist pipeline.”[11] Furthermore, an increasingly sparse funding environment at the NIH has only further exacerbated the situation.[9],[12]

Physician–scientists play an essential role in biomedical research and the advancement of medicine. With a unique blend of experiences in research and clinical practice, they are poised to ask and answer scientific questions aimed at improving clinical outcomes. Physician–scientists often play leading roles in patient–oriented research, including clinical trials, behavioral studies, and epidemiology. Without their contributions to these endeavors, the transition of basic science research to clinical practice would be significantly compromised.

Given these concerns, the present paper reviews the current preoccupations and recommendations of many prominent figures in the research community regarding the uncertain future of the United States’ physician–scientist workforce. In doing so, it aims to highlight the importance of proper training, mentorship opportunities, and incentives for medical trainees who seriously aspire to become physician–scientists.

Methods

PubMed and Google Scholar were searched electronically for peer–reviewed publications relevant to the topic of the physician–scientist workforce. Key phrases included physician–scientist workforce, clinical research, biomedical research, medical scientist training program, NIH, and pipeline model. Articles were included and reviewed in full based on their relevance to the topic.

Review

Past and Present Issues

The trend that drew Dr Wyngaarden’s concern in 1979 were the reduced number of total NIH applications and awards belonging to young MDs and MD/PhDs. Between 1968 and 1978, NIH postdoctoral traineeships awarded to physicians dropped from 3200 per year to just 1400 per year. During the same period, physicians qualifying for NIH research career development awards dropped from 43% to 24%. Furthermore, new principal investigators on NIH research projects who were physicians declined from more than 41% to less than 20%.[1]

Similar trends would continue to emerge in the following decades. In 2000, Zemlo et al noted a 6% decline in the percentage of physicians reporting research as their primary career over the previous 17 years, from 15 377 in 1980 to 14 340 in 2002.[13] During the same period, physicians reporting patient care as their principal career activity nearly doubled, from 376 512 in 1980 to more than 700 000 in 2002.[13],[14] This increase was proportionate to the total number of physicians in the United States between 1980 and 1997. By 2014, physician–scientists made up only 1.5% of all physicians, down from 5% in 1987 (Table 1).[6],[13] In other words, a growing number of graduating physicians disproportionately elect to pursue nonresearch activities.

TABLE 1. Total Number of Physician-Scientists, 2003-2012
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Total* 864355 872458 884276 897293 915641 927258 943972 954115 964251 980377
MD 4297 4325 4413 4287 4226 4162 4132 4149 4204 4192
Asian 439 482 522 541 568 602 631 664 679 693
Black 75 85 84 85 79 78 74 85 89 95
Hispanic 165 166 170 171 184 182 188 194 195 189
White 3483 3463 3505 3362 3261 3156 3079 3024 3019 2954
Other 135 129 132 128 134 144 160 182 222 261
MD/PhD 3859 4002 4061 4075 4102 4041 3955 3991 4069 4086
Asian 700 797 875 907 962 1004 1007 1056 1107 1110
Black 50 58 55 53 53 56 57 53 53 51
Hispanic 146 153 168 160 169 171 174 177 184 180
White 2904 2929 2889 2874 2830 2703 2600 2574 2574 2577
Other 59 65 74 81 88 107 117 131 151 168
*Total number of physicians practicing in the United States
Source: NIH Physician Workforce Report, 2014

These trends were reflected in the makeup of faculty members at academic institutions. Between 1982 and 1996, there was a decline in the number of MD faculty members in basic science departments, from 2381 to 2195.[14] The number of MD faculty members actively involved in research as principal investigators receiving NIH funding had also declined, from 481 in 1982 to 348 in 1997.[14]

Meanwhile, a “large and persistent” gender gap continues to exist among physician–scientists, starting with the negative effects of medical school experiences among women interested in research careers.[15] The NIH Physician–Scientist Workforce report noted that among physician–scientists applying for NIH grants, there were 3 times as many male as female applicants, despite the relative gender parity in medical and graduate school applications.[6],[10] One cohort study following women through medical school showed that they were less likely than men to maintain any intentions of pursuing a career as a physician–scientist by the time of graduation.[15] Other studies have shown that women are also less likely than men to participate in clinical scientist training programs, such as the Medical Scientist Training Program (MSTP) and the Howard Hughes Medical Institute/NIH–sponsored Cloister Program.[5],[15],[16]

Explanations

Various explanations for the declining numbers of physician–scientists have been explored. In the pipeline model, missed opportunities occur as early as “recruitment into the pipeline,” or during premedical and early medical school.[9],[13] First, although many medical school admissions committees do place an emphasis on research experience, the priority for most institutions is the selection of future physicians, not necessarily future physician–scientists.[9],[17] Second, available opportunities for students interested in research careers are limited. The NIH–funded MSTP accounts for just 2% of all matriculating first–year medical students, a number that has remained stagnant for decades.[16],[18] For the other 10% of medical students who have strong intentions to pursue a career in research,[17] structured training programs are available but limited. Some notable examples include programs established by the NIH, the Howard Hughes Medical Institute (HHMI), and the Doris Duke Charitable Foundation.[19][21] A select group of medical schools currently also offers 12 months of dedicated research time during the 4 years of medical school, including the Cleveland Clinic Lerner College of Medicine, Duke University School of Medicine, and Baylor College of Medicine.[9]

Only approximately 50% of all graduates of physician–scientist training programs choose careers in academic medicine.[7] The most effective of these are the NIH–funded MSTP, from which 83% of graduating students pursue careers in academic medicine.7 Non–MSTP, such as the yearlong fellowships funded by the HHMI and Doris Duke Charitable Foundation, are less effective, with only approximately 20% of graduates choosing careers in academic medicine.[7] Finally, postgraduate training programs, including research training during and after residency, are similar to the MSTP in their success at graduating physicians who ultimately go into academia.[7]

For those who have completed graduate medical education with strong intentions to pursue a career in research, “attrition from the pipeline” also occurs at the transition point from traineeship to independent physician–scientist.[9],[22] Specifically, Daye et al point to the time period from when young researchers obtain their first NIH career development awards (K awards) to when they obtain a research grant (R award). Reasons for this attrition include a lack of support at the transition point and the inflexibility of K awards in terms of minimum percent effort and maximum total duration.[9] With heavy clinical duties, high minimum percent effort requirements and time period restrictions make balancing research and clinical practice a difficult proposition for many.

In addition to the “leaky pipeline” metaphor, authors have attempted to explain the declining numbers of physician–scientists as the result of a lack of interest in research among medical trainees. On the other hand, some sources cite increasingly heavy debt accumulation through many years of training and relatively low wages that follow in residency and fellowship programs.[13] In recent years, the NIH has established Loan Repayment Programs (LRPs) to address this issue, offering $35 000 annually in loan repayment to trainees with heavy debt who are engaged in NIH mission–relevant research.[23] Still other authors cite length of training as a deterrent, which for students in the MD–PhD track has increased from an average of 6.6 years in the 1980s to a current average of 8 years.[9],[16],[17] The average age to first R01 or equivalent grant has increased from less than 38 in 1980 to more than 45 as of 2013.[24] Burnout may also be a source of midcareer attrition, with physician–scientists specifically citing work overload, time limitations, and a lack of incentives.[25],[26]

The Contribution of Physician–Scientists

Physician–scientists are essential to the advancement of the field of medicine. Physician–scientists, traditionally defined as MDs who perform medical research as a primary professional activity,[13] are a crucial link between bench and bedside.[27] Their clinical experience provides an ideal setting for asking research questions,[2] making physician–scientists an important force in the transition from clinical observations to testable and applicable research hypotheses.[14] In other words, physician–scientists are the bridge between basic science and its application to human disease.[28] Examples of their contributions include new or improved surgical techniques; an increased understanding of adverse drug reactions; new insights into hereditary diseases, including novel diagnostics and therapeutics; and the identification and treatment of emerging infectious disease.[14] Indeed, patient–oriented research—namely, clinical trials, behavioral studies, epidemiology, and disease pathophysiology studies—would suffer without a robust physician–scientist workforce.[29]

Outside of their research endeavors, physician–scientists also play a critical role in the education of medical students, imparting the foundations of clinical reasoning in the form of the scientific method and its use in evidence–based medicine.[14] For trainees interested in careers in research, physician–scientists serve as important role models and mentors. Thus, with their involvement in research, education, and patient care, physician–scientists are among the few individuals who independently embody the tripartite mission of most major US academic medical centers.

Looking to the future, physician–scientists are also poised to lead innovations in medicine. One example is the current trend toward personalized medicine. Physician–scientists with expertise in genetic medicine, pharmacogenetics, and bioinformatics would be uniquely suited to lead patient–oriented discoveries in the evolution of medicine toward personalized care.[14]

Recommendations

If we consider the pipeline model of physician–scientist development, then it is important to engage students early—at the level of premedical or medical training—to foster an interest in research careers.[30] One way to increase recruitment to the pipeline may be to involve more research faculty in the medical school admission process.[17] Doing so would secondarily shift the focus of medical school admissions to include applicants’ intentions to pursue research careers.

Additionally, interested medical schools could commit a set portion of their incoming class to a non–MSTP research track. One example is the Scientific Discovery Path of Excellence at the University of Michigan, which spans all 4 years of medical school. As part of this research track, medical students are paired with a faculty mentor and complete a research project before graduation.[31] The path gives students opportunities to engage in high–quality research experiences with dedicated research time during medical school.[31] This and similar research tracks at other medical schools give students an additional entry point into the physician–scientist pipeline.

Medical school is a second important point of intervention for reversing the troubling trends in the physician–scientist community. Early exposure of non–MSTP medical students to mentored research projects could open students to the possibilities of integrating clinical research into their careers. A study from the University of Tennessee Health Science Center and Vanderbilt University that followed 1000 students over 25 years enrolled in NIH–sponsored Medical Student Research Fellowship programs found that interest in academic careers significantly increased.[32] On follow–up, approximately one–third of respondents reported that research was a significant part of their postresidency careers.[32]

Another approach might be to increase the available positions in structured training and mentorship programs for young scientists, as exemplified by the HHMI, NIH Medical Research Scholars Program, HHMI Medical Fellows Program, and Doris Duke Charitable Foundation programs. Although limited, outcome data from these programs have been shown to increase interest in research and academic careers.[32] In one study, 21% of HHMI NIH Medical Research Scholars Program graduates and 24% of HHMI Medical Fellows Program graduates had received NIH postdoctoral awards.[19] These percentages were significantly higher than for nonawardees, of whom between 10% and 13% had received NIH postdoctoral awards.[19]

Furthermore, changing medical school curricula to include more problem–based learning and dedicated research time has been shown to increase medical students’ interest in research.[33] In 2003, 70% of medical school deans reported using problem–based learning in their curriculum.[34] However, only 6% used problem–based learning for more than half of formal teaching, and 45% used it for less than 10% of formal teaching.[34] Additionally, only a handful of US medical schools include dedicated research time in their curriculum. Notable examples include Duke University School of Medicine, Cleveland Clinic Lerner College of Medicine, Stanford University School of Medicine, and Baylor College of Medicine. With dedicated research time, medical student journals provide students an additional opportunity to engage in the research process and have been shown to increase students’ interest in research.[35]

From the perspective of pipeline retention, changes including adjustments to K awards may lessen the attrition of physician–scientists.[9] Currently, Mentored K awards (eg, K08, K23) require a minimum of 75% effort, making balancing clinical duties and research a difficult proposition.[36] Furthermore, structured programs blending residency and research training may help retain aspiring physician–scientists. In the standard pathway, medical trainees can go 6 or more years between significant research experiences in medical school (eg, MSTP research years, 1–year medical research fellowships) and their first opportunities to resume research after residency and/or fellowship.[22] Additionally, expansion of programs aimed at addressing the burdens of heavy debt, such as the NIH LRP, may also increase interest in research careers. Finally, making more reasonable clinical and research workloads for physician–scientists may help minimize midcareer attrition and increase the effectiveness of student mentorship.

Conclusion

Various alarming trends, which started to emerge as early as the 1960s and continue to the present day, indicate that the growth of the US physician–scientist workforce is lagging behind that of the rest of the medical profession. Fewer medical students are expressing an interest in research careers, fewer medical graduates are applying for NIH project funding, and fewer MDs have positions in basic science departments. Reasons for these downward trends may include inadequate recruitment, limited structured training programs, and the burdens of debt and time constraints in the postgraduate years. Physician–scientists are essential to the advancement of patient–oriented medical science and clinical practice. They bridge bench research and clinical application and embody the tripartite medical mission of patient care, research, and education. There are several promising widespread changes to medical school curricula and LRPs are steps in the right direction. Further changes may include expansion of better recruitment to the physician–scientist pipeline, more structured training programs, and changes to postgraduate funding mechanisms. Additionally, for trainees who decide to leave the physician–scientist pipeline, alternative opportunities should exist to maximize their interest in academic–based clinical research or teaching with funding for placements in medical education, public policy, epidemiology, behavioral science, and other fields.[37] Ultimately, a multipronged approach (Table 2) that both fosters interest in research careers and addresses weak points in the trainee pipeline is the most promising solution to our declining physician–scientist workforce.

TABLE 2. Recommendations
1 Actively include physician-scientist faculty in the medical school admission process.
2 Reserve a set number of positions in medical school admissions for students with demonstrated interested in research.
3 Include more problem-based learning in medical school curricula to encourage scientific reasoning and questioning.
4 Develop a research track within the medical school curriculum to give interested students opportunities to engage in research.
5 Increase the number of available positions in structured physician-scientist training and mentorship programs.
6 Change the requirements for early-career NIH research awards to allow more clinical time.
7 Expand loan repayment programs to encourage interest in research careers.

Acknowledgments

The author would like to thank Thomas Gardner, MD, MS, for his support, guidance, and careful revision of the manuscript. He would also like to thank Jason Miller, MD, PhD, for his thoughtful revision of the manuscript and contribution to the Recommendations section.

References

    1. Wyngaarden JB. The clinical investigator as an endangered species. N Engl J Med. 1979;301(23):1254–1259. doi:10.1056/NEJM197912063012303.return to textreturn to text

    2. Rosenberg L. Physician–scientists—endangered and essential. Science. 1999;283(5400):331–332. http://www.ncbi.nlm.nih.gov/pubmed/9925491.return to textreturn to text

    3. Goldstein MJ, Kohrt HE. What happened to the concept of the physician–scientist? Acad Med. 2012;87:132–133. doi:10.1097/ACM.0b013e31823f0eeb.

    4. Gordon R. The vanishing physician scientist: a critical review and analysis. Account Res. 2012;19(2):89–113. doi:10.1080/08989621.2012.660076.

    5. Andrews NC. The other physician–scientist problem: where have all the young girls gone? Nat Med. 2002;8(5):439–441. doi:10.1038/nm0502–439.return to textreturn to text

    6. Milewicz DM, Lorenz RG, Dermody TS, Brass LF. Rescuing the physician–scientist workforce: the time for action is now. J Clin Invest. 2015;125(10):3742–3747. doi:10.1172/JCI84170.return to textreturn to textreturn to textreturn to text

    7. Kosik RO, Tran DT, Fan AP–C, et al. Physician scientist training in the United States: a survey of the current literature. Eval Health Prof. March 31, 2014. doi:10.1177/0163278714527290.return to textreturn to textreturn to textreturn to text

    8. Mirmira RG. Editorial: the vulnerable physician–scientist. Mol Endocrinol. 2014;28(5):603–606. doi:10.1210/me.2014–1085.

    9. Daye D, Patel CB, Ahn J, Nguyen FT. Challenges and opportunities for reinvigorating the physician–scientist pipeline. J Clin Invest. 2015;125(3):883–887. doi:10.1172/JCI80933.return to textreturn to textreturn to textreturn to textreturn to textreturn to textreturn to textreturn to textreturn to text

    10. National Institutes of Health. Physician–Scientist Workforce (PSW) Working Group Report. June 2014. http://acd.od.nih.gov/reports/psw_report_acd_06042014.pdf. Accessed December 15, 2015.return to textreturn to text

    11. Association of Professors of Medicine. Association of Professors of Medicine Physician–Scientist Initiative: Recommendations for Revitalizing the Nation’s Physician–Scientist Workforce. Washington, DC; 2008.return to text

    12. Garrison HH, Deschamps AM. NIH research funding and early career physician scientists: continuing challenges in the 21st century. FASEB J. 2014;28(3):1049–1058. doi:10.1096/fj.13–241687.return to text

    13. Ley TJ, Rosenberg LE. The physician–scientist career pipeline in 2005: build it, and they will come. JAMA. 2005;294(11):1343–1351. doi:10.1016/S0084–3954(08)70175–1.return to textreturn to textreturn to textreturn to textreturn to textreturn to text

    14. Zemlo TR, Garrison HH, Partridge NC, Ley TJ. The physician–scientist: career issues and challenges at the year 2000. FASEB J. 2000;14(2):221–230. http://www.ncbi.nlm.nih.gov/pubmed/10657979.return to textreturn to textreturn to textreturn to textreturn to textreturn to textreturn to text

    15. Guelich JM, Singer BH, Castro MC, Rosenberg LE. A gender gap in the next generation of physician–scientists: medical student interest and participation in research. J Investig Med. 2002;50(6):412–418. http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=12425427.return to textreturn to textreturn to text

    16. Brass LF, Akabas MH, Burnley LD, Engman DM, Wiley CA, Andersen OS. Are MD–PhD programs meeting their goals? An analysis of career choices made by graduates of 24 MD–PhD programs. Acad Med. March 2010:1. doi:10.1097/ACM.0b013e3181d3ca17.return to textreturn to textreturn to text

    17. Neilson EG. The role of medical school admissions committees in the decline of physician–scientists. J Clin Invest. 2003;111(6):765–767. doi:10.1172/JCI200318116.return to textreturn to textreturn to textreturn to text

    18. Bickel JW, Sherman CR, Ferguson J, Baker L, Morgan TE. The role of M.D.–Ph.D. training in increasing the supply of physician–scientists. N Engl J Med. 1981;304(21):1265–1268. doi:10.1056/NEJM198105213042104.return to text

    19. Fang D, Meyer RE. Effect of two Howard Hughes Medical Institute research training programs for medical students on the likelihood of pursuing research careers. Acad Med. 2003;78(12):1271–1280. http://www.ncbi.nlm.nih.gov/pubmed/14660432.return to textreturn to textreturn to text

    20. Nathan DG, Wilson JD. Clinical research and the NIH—a report card. N Engl J Med. 2003;349(19):1860–1865. doi:10.1056/NEJMsb035066.

    21. Gallin EK, Le Blancq SM. Launching a new fellowship for medical students: the first years of the Doris Duke Clinical Research Fellowship Program. J Investig Med. 2005;53(2):73–81. doi:10.2310/6650.2005.00202.return to text

    22. Kolars JC. Threats to our training programs for biomedical scientists. 2015. YouTube Web site. https://www.youtube.com/watch?v=nzStWuPR8Ic. Accessed December 15, 2015.return to textreturn to text

    23. US Department of Health and Human Services. National Institutes of Health: Division of Loan Repayment. 2015. http://www.lrp.nih.gov. Accessed October 23, 2015.return to text

    24. Daniels RJ. A generation at risk: young investigators and the future of the biomedical workforce: Fig. 1. Proc Natl Acad Sci U S A. 2015;112(2):313–318. doi:10.1073/pnas.1418761112.return to text

    25. Mostafa SR, Khashab SK, Fouaad AS, Abdel Baky MA, Waly AM. Engaging undergraduate medical students in health research: students’ perceptions and attitudes, and evaluation of a training workshop on research methodology. J Egypt Public Health Assoc. 2006;81(1–2):99–118. http://www.ncbi.nlm.nih.gov/pubmed/17382086.return to text

    26. Buonomano DV. The physician scientist: balancing clinical and research duties. Nature. 2007;15(12):1092–1094. doi:10.1038/ni.3010.return to text

    27. Rosenberg LE. The physician–scientist: an essential—and fragile—link in the medical research chain. J Clin Invest. 1999;103(12):1621–1626. doi:10.1172/JCI7304.return to text

    28. Arias IM. Training basic scientists to bridge the gap between basic science and its application to human disease. N Engl J Med. 1989;321(14):972–974. doi:10.1056/NEJM198910053211412.return to text

    29. Schechter AN. The crisis in clinical research. JAMA. 1998;280(16):1440. doi:10.1001/jama.280.16.1440.return to text

    30. Zier K, Stagnaro–Green A. A multifaceted program to encourage medical students’ research. Acad Med. 2001;76(7):743–747. http://www.ncbi.nlm.nih.gov/pubmed/11448834.return to text

    31. University of Michigan Medical School. Scientific Discovery Path of Excellence. 2015. https://medicine.umich.edu/medschool/education/md-program/curriculum/longitudinal-learning/paths-excellence/scientific-discovery-path-excellence. Accessed January 1, 2015.return to textreturn to text

    32. Solomon SS, Tom SC, Pichert J, Wasserman D, Powers AC. Impact of medical student research in the development of physician–scientists. J Investig Med. 2003;51(3):149–156. http://www.ncbi.nlm.nih.gov/pubmed/12769197.return to textreturn to textreturn to text

    33. Khan H, Taqui AM, Khawaja MR, Fatmi Z. Problem–based versus conventional curricula: influence on knowledge and attitudes of medical students toward health research. PLoS One. 2007;2(7):e632. doi:10.1371/journal.pone.0000632.return to text

    34. Kinkade S. A snapshot of the status of problem–based learning in U.S. medical schools, 2003–09. Acad Med. 2005;80(3):300–301.return to textreturn to text

    35. Wehrens XH, Leiner T. Publications from extracurricular research. Lancet. 2001;358(9284):846. doi:10.1016/S0140–6736(01)05997–9.return to text

    36. National Institutes of Health. Research Career Development Awards. https://researchtraining.nih.gov/programs/career-development. Accessed January 1, 2015.return to text

    37. Miller J. Career intentions of MD/PhD program graduates. JAMA. 2009;301(6):597. doi:10.1001/jama.2009.73return to text