Cronholm, Peter F. MD, MSCE; Straton, Joseph B. MD, MSCE; Bowman, Marjorie A. MD, MPA
The health care needs of the underserved and, indeed, the entire nation, suffer from an inadequate supply of primary care clinical researchers who can bring the unique perspectives of practicing clinicians to clinical epidemiology and outcomes research. The National Institutes of Health (NIH), the Institute of Medicine, and the Veteran's Administration have identified a great need for well-trained clinical researchers in the areas of clinical epidemiology, health services research, and biostatistics. Further, the NIH Roadmap describes the need for researchers skilled in translational science and leaders in multidisciplinary interventions and evaluations. The NIH asked the National Research Council's Committee on National Needs for Biomedical and Behavioral Scientists to comment on the need for new research personnel in the biomedical, behavioral, and clinical sciences. The committee's report recommended that
the NIH, the Agency for Healthcare Research and Quality (AHRQ), and the Health Resources and Services Administration (HRSA) should intensify their efforts to train and retain physicians in clinical research until the decline in their numbers has been reversed and the clinical research workforce includes substantially more MDs and other health care doctorates than is now the case.1
Since this issuance, the primary response of the NIH regarding a clarified role for primary care researchers has been in developing linkages for clinical and translational science awards.
Primary care research comprises contributions from internal medicine, general pediatrics, obstetrics-gynecology, and family medicine. Family medicine clinical researchers are strategically positioned to study and manage chronic illnesses, both physical and mental, that are especially harmful to the elderly, ethnic minorities, and the underserved. They can also effect the lifestyle changes that ameliorate the costs and suffering of these disabilities. Yet, as the youngest of all of the primary care specialties, family medicine struggles to gain entry into academics and NIH funding. The shortage of family physician academic faculty has been well described.1,2 Family medicine research awards and dollars are scarce; in 2006, departments of family medicine received only 154 awards, representing $44.6 million (out of the approximate $28 billion NIH budget),3 representing only 0.20% (0.15% for dollars) and 0.33% (0.22% for dollars), respectively, of total NIH awards.4 In contrast, 11.3% of NIH funding in 2005 supported pediatric research.5 On top of that, only 10 of the nation's 132 family medicine departments (7.6%) earn almost 50% of all NIH family medicine awards.4 Recent surveys have shown that the average medical school family medicine faculty publishes fewer than one article per year and spends a half day or less each week on research.6,7 Successful programs have addressed these issues and increased the percentage of faculty who publish, but the numbers are still low.8 Further, a smaller proportion of family medicine residency graduates choose an academic career than in any other primary care specialty.9 More than half of graduating medical school students who had stated a desire to enter family medicine when they started medical school, but who ultimately entered other specialties, listed “lack of research opportunities” as an important factor in their change of course.10 One of five medical students who considered family medicine residencies cited a lack of support for research as the reason they decided against entering the field.10 This was substantially higher than students who considered but then decided against internal medicine or pediatrics.
Research training is critical for developing family medicine clinical researchers who can secure federal funding and conduct research that affects health care on a national scale. However, it is not well known what elements of research training produce successful family medicine clinical researchers. National programs exist to train primary care researchers, but none focus on family medicine as a profession.11 Faculty development programs in family medicine are common, but they often focus on clinical and medical education skill sets. To help bridge this gap, we describe in this article the structure (curricular components and rigorous training methods) and outcomes of a faculty development fellowship program that trains capable and committed family medicine clinical researchers to conduct independent research into the needs of primary care patients and underserved populations. The study protocol used to obtain and present the data for this article was approved by the University of Pennsylvania's institutional review board.
Description of the Fellowship Program
The faculty development fellowship in the University of Pennsylvania's Department of Family Medicine and Community Health trains family physician clinical researchers to succeed in competitive and demanding academic health systems and to substantially contribute to the nation's health. The fellowship, which accepts one to three fellows per year, is a two-year program with an optional third year for those who desire to extend the development of their research program. To increase the fellows' potential for success, the program's infrastructure gives them the skills and support they need to build a track record of academic productivity in the form of published manuscripts and successful grant proposals.
Key personnel in the fellowship include 70% full-time equivalent (FTE) effort from core faculty within (four family medicine faculty members—two with 20% FTE, one with 10% FTE, and a third with 5% FTE) and outside of the department (5% FTEs each from a biostatistician, an epidemiologist, and an expert in community-based participatory research from other departments at the University of Pennsylvania) and 55% FTE for program coordination.
The curriculum focuses on two domains: research training and academic career development, framing them in content-appropriate coursework and clinical practice in underserved communities. The key components of the research training are master's degree programs and intensive mentored research experience. All fellows complete a substantive research program and are exposed to methodology relevant to their selected master's degree. In addition to the mentoring they receive in the master's programs, they also participate in a product-focused curriculum, structured around their research projects, that exposes them to existing research programs and secondary datasets, introduces them to manuscript and grant proposal development, and brings them into regular research seminars to help them bridge the gap between the knowledge acquired in the master's programs and the skills needed to develop a primary care research career. All fellows are expected to publish articles and write grants during their fellowship, and intensive small-group sessions are used to critique and advance these skills. The fellows also participate in a structured career-development curriculum that addresses academic promotion, personnel management, budgets, organizational dynamics, interpersonal skills, writing and publishing skills, and strategic planning. The entire curriculum is conducted within the context of each fellow's clinical practice in an underserved area. This practice maintains and expands the fellows' clinical skills, supplies them a clinical basis for their research protocols, and provides a venue for developing their skills as medical educators. Table 1 lists the key components of the curriculum, the training methodologies, and the skills targeted for development. Figure 1 illustrates the breakdown of time during the fellowship and the contextual framework of the program's structure. The fellows' focus shifts from coursework and acquiring skills to developing research programs as their collaborative networks grow stronger and their grant and manuscript production increase, all under the close oversight of their mentors and core fellowship faculty.
We have had multiple sources of funding to support the fellowship. The primary source for the academic portion of the fellowship was HRSA (Faculty Development in Primary Care: D55-HP-05164), which has provided support for our administrative infrastructure, tuition costs, and stipend fees for the majority of fellows. Some fellows have been funded using National Research Service Awards or T32 grants, and one fellow received partial support from Veterans Affairs. Fellows do some self-funding, occasionally covering some of their own tuition, paying taxes on tuition that was paid for them, and bearing the opportunity cost of receiving a fellow's versus an attending's salary. All fellows have additionally had modest clinical practice income. The department provides space and some gap coverage for faculty effort, particularly that of the department chair.
In this article, we describe outcomes for program completers during the 10 years from 1997 to 2007. We define program completers as “all faculty participants finishing the training program each year regardless of source of stipend support.”
Intensive research training
Master's degree programs.
Completion of advanced degree programs has been shown to support a career in academic medicine.12 Research-intense master's degree programs develop the skills needed to further an academic career, such as developing research protocols, writing grant proposals, and producing manuscripts.13 A major focus of our fellowship was to engage trainees in research-intensive graduate degree programs with an output of at least one published manuscript based on their thesis project. While one program completer received a master's degree in bioethics, a second fellow received a master's degree in business administration, and a third, who already had a clinical research PhD, was exempted from the requirement, the majority (12 out of 15) completed a master's of science in clinical epidemiology (MSCE). The MSCE program, administered by the Center for Clinical Epidemiology and Biostatistics at the University of Pennsylvania, is a well-known entity with a strong track record of training successful clinical researchers. The MSCE program (1) provides an in-depth, working knowledge of the techniques of clinical epidemiologic research (including clinical trials, cohort and case-control studies, surveys and quasi-experimental studies, meta-analysis, analysis of large epidemiological and claims data sets, outcomes research, decision modeling, population interventions, and clinical economics), (2) provides research experience with mentors in clinical epidemiology and health services research, and (3) brings together, through participation in a seminar series in clinical epidemiology/health services research, faculty and fellows.
Fellows and core faculty meet every other week for 90 minutes to advance the fellows' research training through an interactive meeting focused on academic productivity. The seminars focus on four aspects of the fellows' research: (1) protocol development, including methodological issues regarding different research designs, development of research questions, and plans for analysis, (2) poster and podium research presentations, (3) manuscript development, and (4) grant proposal submission.
Our fellowship was structured on the philosophy that researchers in training succeed best when closely mentored. The development of mentoring relationships between faculty and fellows is a cornerstone of our program. Departmental fellowship mentoring complements any mentoring components of the fellows' master's degree programs. For example, fellows entering the MSCE program have thesis committees established with a primary mentor, a secondary clinical research mentor, and a biostatistician. Committee members come from within or outside of our family medicine department and are selected based on thesis-specific expertise. To increase the breadth of mentoring, on entry to the program, fellows identify areas of clinical and research interest. On the basis of individual interests, fellows are matched with faculty known to be outstanding mentors of junior investigators. Within the first six months of beginning their two-year program, fellows submit an individual development plan outlining short-term and long-term goals, a strategy for meeting those goals, and a plan to evaluate their progress. The plan is reviewed by the core faculty and discussed with the fellow individually by the appropriate mentor. Fellows are expected to meet with their mentors at least twice monthly, with increased frequency as the required research project is designed and carried out.
In addition to their thesis topics, fellows may design projects nested within an existing research endeavor, often one headed by their mentor. The core faculty encourage the fellows to pursue primary and/or secondary data analyses using either quantitative or qualitative methodologies. Primary and secondary data analyses are acceptable for participant protocol development and eventual thesis production.
Access to high-quality datasets is a key to jump-starting research careers by increasing the chances of producing high-quality publications and substantially reducing the time and fiscal barriers associated with primary data collection. However, fellows can select either primary data collection or secondary data analyses.
Career development curriculum
Alternating with the research seminars, fellows attend a series of seminars that address leadership skills in academic family medicine. The leadership and management skills sessions enhance the scientific components of the fellows' training and give them the tools they need to become effective academic leaders. The curriculum includes seminars, workshops, case studies, group partnering exercises, and peer coaching. Fellows are exposed to the foundations of leadership, including organizational management and change, health system administrative structure, critical and nontraditional thinking, negotiation, personnel management, time management, and budget development.
Clinical practice and practical experience as a medical educator
A root cause of health disparities in our nation is the provision of care in a culturally restrictive manner.14 To improve the health care quality and outcomes of ethnically diverse patients, we designed the clinician-educator component of our curriculum to give our fellows the skills and knowledge necessary to deliver culturally competent care to multicultural clients. Fellows conduct their clinical practice in a medically underserved area. They also interact with learners (medical students and family medicine residents) through lectures, small-group work, and one-on-one clinical precepting sessions informed by the educational techniques covered in the seminars.
Outcomes of the Fellowship Program
The outcomes we describe here cover two domains critical for the academic success of junior faculty: academic productivity and funding. For the purpose of analysis, we define academic productivity as a composite measure of manuscript publication that includes original research, scientific reviews, book chapters, and editorials published in peer-reviewed and non-peer-reviewed medical journals. We assessed funding using a composite measure of research and career development funding.
Other outcomes we considered include the clinical settings in which our fellows work (medically underserved communities/areas [MUC/MUA] versus non-MUC/MUA) and the recruitment and retention of underrepresented groups. For the latter, we used the HRSA-defined categories of educationally disadvantaged (an individual comes from an environment that has hindered him or her in obtaining the knowledge, skills, and abilities required to enroll in and graduate from a health professions school), ethnically underrepresented (an individual categorizes him- or herself as American Indian, Alaska Native, Asian [other than Chinese], black/African American, Hispanic/Latino, Native Hawaiian, or other Pacific Islander), and economically disadvantaged (an individual comes from a family whose parents' annual income is below levels established by HRSA).
Seventeen fellows have matriculated since 1997. Two fellows dropped out, leaving 15 program completers. Six program completers were women, and nine were men. Fellows matriculated directly from residency training with the exception of one who had been in clinical practice. Tables 2 illustrates the productivity and funding outcomes for each of the fellows.
Completion of master's programs
All program completers graduated from their master's degree programs, except for the one who had previously earned a scientific PhD and was exempted from the requirement.
All but one program completer have published peer-reviewed articles. Only one fellow had published before entering the fellowship (data not included in Table 2). To date, the cumulative manuscript-associated productivity for the 15 program completers is an average of 7.6 (range: 0–30; median: 6) peer-reviewed research manuscripts and 8.1 (range: 0–27, median: 7) non-peer-reviewed publications (book chapters, letters to the editor, editorials, and invited commentaries). Dividing by the median time since completion of the fellowship (five years) results in an average of 1.6 peer-reviewed publications and 3.4 total publications per fellow per year. Table 2 illustrates the publication productivity of the program completers.
Twelve of the 15 program completers have received research grants. Program completers have been the principal investigators on 39 funded studies and coinvestigators on 24 funded studies. Five program completers have secured funding through an NIH career development award (CDA), and nine have received other awards (e.g., National Research Service Awards or foundation funding). Of the six program completers who applied to the NIH loan repayment program, four successfully obtained support. Table 2 illustrates the grant productivity of program completers.
All program completers demonstrated commitment to improving the health of underserved or vulnerable populations through translational research. Our fellows' research has focused on reducing disparities in maternal-child health, decreasing intimate partner violence, improving treatment adherence among older persons with depression, reducing the symptom burden in older persons with cancer, reducing disparities in cancer prevention, and reducing the progression of HIV infection. All program completers provided care in MUC/MUA settings during their training, and 12 continue to practice in MUC/MUA settings. Fourteen program completers hold academic faculty positions.
Recruitment and retention of underrepresented groups
Four of our program completers came from underrepresented or disadvantaged backgrounds. One came from an economically disadvantaged background, one from a background both economically and educationally disadvantaged, and two from ethnically underrepresented groups.
We believe that our fellowship program has met its goal of developing family medicine clinical researchers who can and will conduct independent research that addresses the needs of underserved populations. It succeeded through a rigorous curriculum that includes training in research-oriented master's degree programs. Since its inception in 1997, our program has trained 15 program completers whose commitment to fixing health disparities is demonstrated in their choice of research questions (e.g., mental health, women's health, intimate partner violence, health literacy, aging, end-of-life care, and HIV care), modalities (community-based participatory action and mixed-method models), and site of practice (underserved areas). Twelve program completers still practice in medically underserved communities, with one serving as medical director at a federally qualified health center. The outcomes presented in this article show that we have a strong track record of training leaders in primary care research whose clinical efforts greatly contribute to the mitigation of health disparities nationally and globally.
Some limitations should be considered. The most prominent is generalizability. This article describes a single program developed within a resource- and research-rich institution. According to figures released by the NIH, the University of Pennsylvania's School of Medicine ranks second in the nation in terms of total research awards to academic medical schools. For fiscal year 2005, the university received 1,153 total awards, including research and training grants, worth a total of more than $470 million. The University of Pennsylvania also has an established interdisciplinary tradition and commitment to disease-related programs, as evidenced by its 43 centers and institutes and the more than $97 million in NIH-sponsored program projects. Institutes, centers, and multidisciplinary projects run extensive seminar and educational programs, as well as pilot grant support programs. And so we acknowledge the argument that our outcomes result in part from unique resources. At the very least, our outcomes strongly support the development and expansion of family medicine training opportunities in other resource- and research-rich institutions. We also believe that the fundamental model of faculty development presented in this article can be implemented in a wider variety of settings. Mainous and colleagues7 found that whereas publication productivity did not vary significantly between research-intense and non-research-intense institutions, grant productivity was higher in research-intense institutions. With additional resources directed at supporting grant productivity, faculty training programs such as ours can succeed even without our unique resources.
The training of primary care faculty has often focused on education rather than research. The training of primary care researchers is challenging on a national level.5,15,16 Internal medicine, pediatrics, and obstetrics-gynecology have larger and longer track records in NIH research than does family medicine because of the ages of their specialties and the concordance of NIH funding preferences with the subspecialty training divisions within their professions. To develop an awareness of the family medicine perspective at NIH and academic institutions, we need to draw on our discipline's strengths in training clinical researchers skilled in community-based, multidisciplinary research. Only the federal programs funded under Title VII of the Public Health Service Act are specifically designed to support family medicine training. However, in 2008, Title VII funding was dramatically reduced to about $150 million—half of what it was in fiscal year 2006, and only a small percentage of which is devoted to research training.
Other mechanisms of training primary care researchers include programs supported by foundations, such as the Robert Wood Johnson Foundation's Clinical Scholars Program. For the past 30 years, the Clinical Scholars Program has trained its fellows using mechanisms similar to ours.11 It proportions time in the same way, making 20% of the fellows' time available for clinical work and setting aside 80% for training in program development, research methods, and leadership development. It also develops trainees' skills in health care research via formal coursework, individual mentorship, and guidance in project development with a similarly defined product of a graduate-level research project. Unlike our program, however, it has the advantage of The Robert Wood Johnson Foundation's national network of scholars and mentors and assets of some $10.7 billion.17
A lot has been written about what factors make a successful faculty researcher. Clearly, individual, departmental, and institutional characteristics are all important.18 Individual characteristics that favor success include adequate exposure to and a passion for research. Access to fellowship training provides a clear advantage,19,20 as does having a definable research agenda, research networks, and several simultaneous projects available for collaboration.6 Perhaps the most important factor is simply having the time to conduct research. But unless such time is protected, it is often lost to clinical and teaching responsibilities or stymied by family medicine departments' inadequate funding. Family medicine departments with a clear research priority and strong research leadership can produce successful researchers, as measured by peer-reviewed papers and grants.19 And institutions can help by providing networks and mentors outside of the department as well as by encouraging research via promotion standards.6,18
One of the many interdependent components making up the structure of our model of faculty development—and a major one—is the quality of the relationships established between our fellows and their mentors. To build a strong core of research mentors, we aggressively recruit experienced family medicine researchers and other faculty from outside of our department (or jointly appointed with our department) for their expertise in biostatistics, epidemiology, health services, community-based participatory methodology, and medical anthropology. Our fellows' collaborative relationships with their mentors lead to academic success in their graduate programs and professional success in publications and grant applications. The fellows also learn the value of successfully competing for one's own funding and are encouraged to apply for both private and federal career development awards to support their continued development as independent researchers. Of course, this foundation for the future depends on ongoing availability of quality mentors. Our program completers have now begun filling mentorship roles for the next generation of fellows.
This model of rigorous research training and multifaceted career mentoring successfully trains family medicine clinical researchers, even in the face of limited research funds and the demanding requirements of the rank and tenure systems common at research-oriented universities. We believe it can, and should, be successfully replicated at other institutions.
This publication was supported in part by a Health Resources and Services Administration grant (D55-HP-05164) for Faculty Development in Primary Care.