Fixing the leaky physician‐scientist pipeline: Integrated–dedicated research period programs in emergency medicine

Abstract

Emergency physicians (EPs) are well positioned to perform medical research. EPs are exposed to a wide range of disease types, medical specialties, and treatment modalities. Furthermore, emergency medicine (EM) serves as the safety net for the U.S. health care system. The diverse exposure provides a vast opportunity for EP to perform many worthwhile research projects. Yet, EM has historically had the lowest amount of funding and a lower number of National Institutes of Health–funded research projects. Many suggest the etiology is a “leaky” educational pipeline with loss of many potential physician‐scientists over the training and development course. Current research training options for the EM physician‐scientist includes MD‐PhD, 4‐year EM residency program and postresidency fellowships. While each has its advantages and disadvantages, we describe an additional educational alternative of EM physician‐scientists, which we have named the integrated–dedicated research period within an EM residency. We describe the features of these programs and preliminary results from the graduates and current trainees.

NEED FOR INNOVATION

The physician‐scientist is an uncommon breed, one who has dedicated a career toward healing of patients but also seeks to improve the field through their own scientific research. ¹ Emergency physicians (EPs) are uniquely poised to perform such work as they have a wide exposure to disease types, a diverse patient population, and ability to treat/study patients early in their disease process. Furthermore, their clinical practice sits at of unique junction between inpatient and outpatient health care providing insights of the existing health delivery systems. ² , ³ , ⁴ It is self‐evident that improving emergency medicine (EM) practice in evidence‐based fashion requires a body of EM experts to produce that evidence.

However, conducting high‐quality research requires specific training. There are concerns that future EM residents will not pursue this additional research training due to burnout or financial pressures, among other factors causing attrition from academic positions. ⁵ This is supported by relatively low uptake of current research training pathways by EM physicians. ⁶ Furthermore research funding in EM has historically ranked among the lowest in comparison to other medical specialties. ⁷ , ⁸

Multiple studies have suggested improving training and mentoring as a solution to this issue. ¹ , ⁹ , ¹⁰ , ¹¹ , ¹² While there presently remain options for physician‐scientist training, we report an alternative educational model for improving the physician‐scientist training pipeline, where time is provided during residency training to focus solely on research training. We have deemed this educational model within an EM residency as an integrated–dedicated research period (IDRP).

BACKGROUND

To fully understand the structure and utility of an IDRP, it is important to understand the historical context of EM physician‐scientist training. In 1959 MD‐PhD programs were created to address the increasing demand for formal research training by physicians. ¹⁰ Unfortunately, EM consistently ranks among the lowest specialty selected by MD‐PhD graduates. ¹³ , ¹⁴ Although the reasons for this have not been specifically studied, it can be assumed intuitively that the time, expense, and other resources required to obtain a PhD are significant barriers to EM graduates.

The 1994 Macy Report called for a defined research agenda within EM. In response the Society for Academic Emergency Medicine (SAEM) established a clearinghouse of fellowships and reported which of those focused on research. ¹⁵ Two subsequent status reports established that out of 211 total EM fellowships only 12% had a research focus and of those, less than 70% were filled. ⁶ , ¹⁶

As an alternative to fellowship training, some proposed expansion of research education during residency training. This became especially relevant in the discussion of EM residency training length: 3‐year versus 4‐year programs. While articles have shown the additional year provides improved academic output, ultimately, “there is insufficient time [in four‐year residency programs] for formal research training.” ¹⁷ In essence, residents not only need time to perform research but also time for research education.

In 1999, the National Institutes of Health (NIH) developed a training award mechanism (K23) to provide additional time for clinical investigator education. Dr. Jeremy Brown of the NIH provided a review of use of these awards by EM physicians from 2008 to 2013. The review found that the number of applicants from EM was much less than from alternative specialties. ¹⁸ The factors for these lower numbers is likely multifactorial, but a response by SAEM and the American College of Emergency Physicians suggested in part that the lower number of applicants was a result of a self‐selection process due to inadequate training and preparation to apply for K awards. ¹⁵ Furthermore, they suggested rather than focusing on postresident fellowship training that “[we should explore] providing our trainees with dedicated, longitudinal time during their [resident] training for in‐depth research experiences, similar to those provided in surgery residencies and medical subspecialty fellowship programs.”

As seen in the preceding paragraphs, each research training format has provided various benefits but also significant limitations. In the next section we provide a layout and the theoretical framework for the concept of an IDRP.

OBJECTIVE

The objective of the IDRP is to produce successful academic EM researchers. The aforementioned historical review provided the inspiration for the formation of the IDRP as seen in Figure 1. In the early 2010s residency and departmental leaderships recognized the dearth of dedicated scientific research training in EM training programs. This educational gap was especially noted in comparison to the dedicated research training tracks available in other specialties' residencies and subspecialty fellowships. As a result of this discrepancy in education, at the start of their academic careers, EM physician‐scientists were competing with more experienced investigators for the same funding. To support future physician‐scientists, leadership's underlying conceptual framework envisioned three main benefits to an IDRP over the traditional research training methods.

FIGURE 1.

Integrated–dedicated research period (IDRP) format for Duke and Yale Departments of Emergency Medicine.

Shorter duration

While MD‐PhD programs have broadly provided physician‐scientists with a rigorous and thorough training education, these applicants are unlikely to select emergency medicine as their specialty. ¹⁹ The decision to not pursue an MD‐PhD was, in part, related to duration of training. ²⁰ As such, the goal for the IDRP was to provide a training program, shorter in duration than a formal MD‐PhD period.

Dedicated research time

While maintaining an abbreviated timeline, the goal is also to provide a more dedicated training period versus the current residency programs. As such, providing at least 1 year of 25% clinical effort (or roughly 39 weeks) greatly expands upon the average 13 weeks of elective provided by most 4‐year residencies. ²¹

Physician‐scientist retention

Lastly, this format aims to improve retention of prospective physician‐scientists. By moving up their research training to earlier in their career, studies have demonstrated reduced burnout and improved satisfaction with residency training. ²² , ²³ It was our hypothesis that obtaining specific research skills and experiencing research success earlier in their career would lead trainees to be more likely to continue in academic careers and pursue research.

DEVELOPMENT AND IMPLEMENTATION

To our knowledge, our two programs are the only two such programs that currently exist nationally and were developed independently. As two large academic EM programs, the infrastructure available to both programs were somewhat readily available but required adjusting and administration confirmation.

At Yale, administration was approached for approval including the associate dean for student research to open the Master of Health Science (MHS) degree to its residents. Furthermore, as part of this approval, an application package was created to apply to the MHS‐granting program. Others approached include Residency Review Committee (RRC), the EM department chair, residency directory, and graduate medical education (GME) committee. Overall, the ideation and implementation for this program progressed over approximately 1 year.

The Yale Emergency Scholars (YES) program was given its own NRMP match number, although residents are also given the option of transferring into the program after their intern year. Fellows are selected similar to traditional residency applicants. ²⁴ In this program, the fellow begins their dedicated research time (75% effort) after their second year of traditional residency training and spends 2 years conducting research while also completing their residency training. As a fifth‐year fellow, they work clinically as an attending much like a traditional postgraduate fellow and part of their billing goes toward funding their MHS.

At Duke, the existing infrastructure within the department of surgery was leveraged since at the time of development EM was a division in that department. Resident research fellows were allowed to join the department of surgery's existing residency research program to provide the necessary dedicated time for research. Structurally this is a non‐ACGME program that provides the resident fellow time away from typical clinical training program requirements to focus on learning and conducting research. The fellow begins their dedicated research time (75% effort) after their second year of traditional residency training. The range of research interests has been broader than those pursed at Yale. As such, funding for the salary of the fellows during their fellowship year has been mostly provided by the department, but fellows have received funding as part of an existing Department of Defense grant and an NIH R38 grant. Educational opportunities for fellows have been provided by classes hosted at Duke and other professional organizations and independent academic conferences. The total time for creation of this fellowship was approximately 1 year but it has been continually refined since. In terms of fellow selection, residents are made aware of this program during interviews, but residents are selected by department leadership after their intern year based on clinical skills, prior research experience, and aptitude.

Both programs were able to balance clinical shifts with lab requirements if performing basic research by planning projects with mentors and residency administration without difficulties. Furthermore, mentors are selected by the IDRP applicant and program leadership. An individualized plan for each fellow is created, outlining a timeline for research project tasks, research skill development, and clinical commitment at the start of the fellowship year. As part of selecting a mentor, expenses related to the research itself are primarily the responsibility of the mentor. Furthermore, staffing required to perform other research activities relevant to the research project was allocated by the mentor as available.

OUTCOMES

Globally, these programs sought to improve the physician‐scientist education pipeline by (1) increasing the number of researchers and (2) improving the success of physician‐scientist performing research. Unfortunately, objectively defining success in research education is somewhat problematic. Although measures such as number of academic appointments, grant dollar amounts, and publication rates are generally used as quantitative markers of success, a 2020 study demonstrated that there exists a wide range of opinions on what defines a successful physician‐scientist. ²⁵ Ultimately, due to precedence in other specialty programs, we have utilized publication rates and academic appointments as crude markers of “success” in the evaluation of the IDRP. ²⁶ , ²⁷

Duke University

Since its inception the program has matriculated five fellows; three graduated and two are current. Interests have ranged from translational research and data science to engineering and global health. All graduates have retained academic positions. In aggregate, IDRP residents produced 21 peer‐reviewed articles through March 2023. Residents have also subsequently received funding from a variety of sources including NIH, SAEM, and internal grants; degree‐earning scholarships; and the Department of Defense. Over the same time period, only one resident in our traditional residency program has received grant funding during their training. Comparatively three of the five IDRP fellows have received some form of funding. Additionally, during that time frame, on average, residents entering into traditional (nonresearch) fellowships have published 0.37 articles per resident (even when accounting for an additional year). Comparatively our IDRP fellows have published 2.2 papers per resident.

YES program

The YES program started in 2013 and matriculated 10 fellows since its inception; five have graduated and five are still in training. They have had a wide range of interests including toxicology, health services, EMS, and ultrasound. All graduates thus far have accepted positions affiliated with research institutions and universities. Scholars in the program have generated six extramural grants and 54 peer‐reviewed, PubMed‐indexed publications through October 2022, making the publications per resident 5.4. While the publication rate for Yale's standard residency is unknown, another institution formally evaluated over a 10‐year period and demonstrated a 0.4 publications rate per resident. ²⁸

DISCUSSION

Organizations such as Association of American Medical Colleges (AAMC) and NIH have established the importance of work performed by physician‐scientists, yet EM has lagged behind other specialties in number and funding of research projects. ¹⁹ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ Some propose the cause to be inadequate researcher workforce, dubbing the phenomenon a “leaky educational pipeline.” ⁹ , ¹⁰ , ¹¹ , ³⁴ , ³⁵ Current physician‐scientist education formats include MD/PhD programs, 4‐year residency with emphasis on research, and postresidency research fellowships. Each of these methods have its advantages and drawbacks. MD/PhD programs can dissuade potential researchers due to length of education and lack of clinical experience prior to research topic selection. ²⁰ Four‐year residencies tend to have higher publication rates but lack the comparative dedicated research time needed to develop their skills. Postresidency training provides the relevance and dedicated time but may lose potential applicants to burnout and/or unwillingness to further delay salary increase. As previously proposed, we feel the IDRP provides an alternative solution that addresses these concerns.

While limited, our data demonstrate a higher publication rate of IDRP participants in comparison to residents of a standard residency. It is important to note that it is difficult to assess what defines a successful physician‐scientist and thus what defines the success of a physician‐scientist training program. While more objective measures such as funding and publications are mentioned, they are not necessarily a uniformly accepted method of evaluation. ²⁵ Furthermore our data are limited in providing a broader evaluation of impact and comparison in that: (1) there is a lack of data to compare postresidency fellowship outcomes to the IDRP outcomes and (2) the impact the IDRP has on long‐term grant applications and success rates (i.e., R01) in comparison to alternative training methods.

While successful at our institutions we note a few limitations to more generalized implementation. Firstly, institutions need adequate research infrastructure. Successful training necessitates multiple experienced faculty with actively funded research programs to provide opportunities for successful project completion and mentorship. In addition, ancillary staff such as statisticians, research coordinators, regulatory coordinator, etc. (variable based on type of research to be completed), are essential to a well‐constructed research project. Secondly, we recognize funding as a limiting factor. Although finance mechanisms may change, currently, our programs have used a mixture of department funds and NIH research fellowship grants to support IDRP residents. Additionally, dedicated time by program leadership and research mentors may be limiting in an already otherwise busy schedule.

This paper presents two examples of an innovative IDRP training program as an alternative to current research training. Using this dedicated and integrated method we provide residents with an abbreviated, but thorough, research education resulting in a synergistically improved residency experience to produce a new generation of successful physician‐scientists. While further adjustments may be required to sustain such a program, ultimately, we feel our model provides an additional patch of the “leaky” physician‐scientist pipeline.

AUTHOR CONTRIBUTIONS

Study concept and design: Charles J. Gerardo, Alexander J. Limkakeng, Steven L. Bernstein. Acquisition of the data: Mitchell C. Veverka, Caitlin R. Ryus. Drafting of the manuscript: Mitchell C. Veverka. Critical revision of the manuscript for important intellectual content: Charles J. Gerardo, Alexander J. Limkakeng, Steven L. Bernstein, Mitchell C. Veverka, Caitlin R. Ryus.

CONFLICT OF INTEREST STATEMENT

Mitchell C. Veverka is an active participant in an IDRP program at Duke University. Caitlin R. Ryus is a graduated IDRP resident from Yale University's program. The other authors declare no conflicts of interest.

Veverka MC, Ryus CR, Gerardo CJ, Bernstein SL, Limkakeng AJ. Fixing the leaky physician‐scientist pipeline: Integrated–dedicated research period programs in emergency medicine. AEM Educ Train. 2023;7:e10919. doi: 10.1002/aet2.10919