Strom, Brian L. MD, MPH; Kelly, Thomas O. MSEd, PhD; Landis, J. Richard PhD; Feldman, Harold I. MD, MSCE
The practice of population medicine has a long history at the University of Pennsylvania (Penn) School of Medicine, now the Perelman School of Medicine at the University of Pennsylvania, highlighted by a century-long struggle to identify that discipline's proper home, its integration with clinical medicine, and how it should be taught.1–3 Multiple programs were established and then dissolved or modified as school leaders attempted to include a population perspective in the curriculum and establish a scientific base for this discipline.
Similar struggles occurred nationally during the latter part of the 19th century and through the middle of the 20th century. During that time, the fields of clinical medicine and public health operated as increasingly independent disciplines. Clinicians provided for individual patients; those trained in public health focused on prevention and the determinants of health and diseases in populations. Research and training programs developed within schools of public health generally were independent and not conducted collaboratively with schools of medicine. Further, universities with both schools of medicine and schools of public health often maintained separate departments of epidemiology and biostatistics in each, commonly existing without school- or institution-wide interdisciplinary and collaborative research missions. This divide was unproductive and sometimes exacerbated by distrust.3
In the late 1970s, a senior faculty member newly recruited into what was then the Department of Community Medicine sought to develop a program in clinical epidemiology in Penn's school of medicine. His goal was to integrate epidemiology into the clinical specialties, in recognition that the questions addressed by epidemiology were central to pressing medical and societal concerns. This represented a new approach at Penn and elsewhere. The Clinical Epidemiology Unit (CEU) was formally recognized in 1979 as an intradepartmental unit housed within the Department of Medicine's Section of General Internal Medicine.
In the 1980s, the CEU established collaborative research programs with faculty members who were based in clinical departments and divisions throughout the school of medicine and developed a training program for clinicians in clinical epidemiology.4 Further, CEU faculty members, including former trainees who were given joint school appointments in clinical departments and the CEU, demonstrated their ability to obtain extramural funding and develop large and successful research programs.
Despite these accomplishments, the organizational structure of the CEU created two major problems that inhibited its growth and development. First, although the CEU was based in the Section (later Division) of General Internal Medicine, its faculty also included individuals with appointments external to the Section of General Internal Medicine. Yet, because the CEU was identified with one clinical program, formal ties to other clinical programs were difficult to form. The CEU's increasing relationship to multiple and varied clinical programs needed to be addressed if the CEU was to grow, thereby expanding population-based research within the school of medicine.
Second, again because of its base in a clinical division, the CEU was unable to award tenure-track positions to nonphysicians. As a result, the CEU was unable to recruit biostatisticians onto the standing faculty, an impediment that prevented growth in the biostatistics program sufficient to meet the needs of researchers throughout the school of medicine. These limitations also made difficult the task of ensuring that Penn's medical students graduated with adequate knowledge of biostatistics, a discipline critical to their lifelong continuing education.
The Creation of a New Organizational Structure and Mission
To address the organizational and structural barriers that confronted the CEU, in the early 1990s the dean of the school of medicine approved a new model: an interdepartmental center, designed to foster interdisciplinary clinical research, and a smaller basic science department designed to support the mission of the center. The Center for Clinical Epidemiology and Biostatistics (CCEB) was created in 1993; the Department of Biostatistics and Epidemiology (DBE) was formed in 1994. The objectives in forming the center were to provide
* a structure that would encourage interdisciplinary research in clinical epidemiology and support primary research in biostatistics and epidemiology;
* a collaborative environment in which clinical epidemiologists and biostatisticians would contribute to and lead multidisciplinary research programs;
* a school of medicine consulting service; and
* training in clinical epidemiology.
The objectives of the department, linked to the new center, were to provide
* an academic home for biostatisticians and nonclinician epidemiologists;
* an academic structure to support primary research in biostatistics and epidemiology;
* a structure to better facilitate recruitment of the nonclinician faculty members needed for collaborative research involving clinicians, biostatisticians, and nonclinician epidemiologists;
* consultation on issues of study design and statistical analysis; and
* training in biostatistics.
The appointment of a clinician (B.L.S.) as the founding director of the CCEB was particularly notable, given the goal to expand recruitment of nonclinical faculty members in biostatistics and epidemiology. This was considered to be necessary to ensure that the CCEB would maintain a clinical focus. The same individual was also appointed as chair of the DBE. This dual appointment of a physician-clinical epidemiologist as leader of both entities cemented the relationship between them.
Interested and qualified faculty members were invited to apply for appointments as core or affiliate members of the CCEB. Many already held appointments in the CEU, including clinical epidemiologists holding academic appointments in different school of medicine clinical departments and divisions. Clinical epidemiologists given core CCEB faculty appointments were expected to play very active roles as epidemiologists in the CCEB while maintaining their primary academic appointments in clinical departments. These clinical faculty members were not forced to align exclusively with their clinician or epidemiologist identities; they were encouraged to embrace both. To reinforce these dual affiliations, the school agreed to give financial credit both to the CCEB and each faculty member's home department for funded research and programs led by their department's faculty. Further, the school instituted a funding model for the CCEB based on growth in extramural support. This arrangement enabled the CCEB to expand on the basis of its productivity without creating adversarial relationships with the school's clinical departments. With the faculty appointment structure and financial crediting system described above, it was envisioned that the CCEB would be able to forge relationships easily with multiple clinical departments and divisions within the school.
The creation of the DBE expanded on the alignment of the academic disciplines of biostatistics and epidemiology by colocating them within one department, emphasizing the connection between the two and, thereby, enhancing collaboration. Nonclinician epidemiologists and biostatisticians were given primary appointments in the DBE; clinician epidemiologists were given secondary appointments with an equal voice in departmental governance conferred through voting rights.
Core faculty members were expected to maintain appointments in both the CCEB and DBE. Faculty members would collaborate closely as coinvestigators on research projects requiring selected biostatistics and epidemiology expertise and act as principal investigators on primary research projects in both biostatistics and epidemiology. Two entities, which were both units and divisions, were established within both the CCEB and DBE: the CEU/ Epidemiology Division and the Biostatistics Unit/Division.
Interested individuals with the appropriate level of expertise were given affiliate faculty appointments in the CCEB and secondary academic appointments without voting rights in the DBE, if desired. Affiliate faculty members committed to providing service to the CCEB, much of which involved participation in the CCEB's educational programs. Generally, financial and other support for affiliate faculty members was not provided by the CCEB.
The CCEB Today
The CCEB has matured as envisioned, providing the primary home for epidemiology and biostatistics at Penn, and is meeting the objectives identified on its inception. The CCEB is multidisciplinary; its faculty members have advanced degrees in various disciplines, including applied and pure mathematics, biostatistics and statistics, clinical pharmacy, decision science, economics, epidemiology, genetic counseling, genetics, nutrition, exercise physiology, health services and administration, health policy, history and sociology of science, information science, pharmacology and toxicology, public health, sociology, and social work. The CCEB's efforts are positioned between those of the Institute for Translational Medicine and Therapeutics—created in 2005, which supports research at the interface of laboratory and clinical research—and the Leonard Davis Institute of Health Economics, established in 1967, which analyzes how health care is organized, financed, managed, and delivered. Many CCEB faculty members have appointments in one or both of these other entities.
As of November 2011, the core CCEB faculty consists of 36 clinician epidemiologists, 11 doctoral epidemiologists (nonclinicians), 30 biostatisticians, and one clinician biostatistician (these totals do not count the 88 affiliate faculty members). Annually, CCEB faculty members serve as authors and coauthors for approximately 1,000 peer-reviewed publications. All epidemiology and biostatistics faculty members were fully funded extramurally for all effort not devoted to administrative, teaching, or clinical roles.
Training in clinical epidemiology has been provided by the CCEB/CEU since 1979. The master of science in clinical epidemiology (MSCE) degree, one of the first of its kind in the United States, has been awarded since 1985.4 The MSCE program, designed to be completed in two to three years of full-time training, or longer for part-time trainees, is designed for clinicians seeking careers as independent academic investigators. Approximately 35 first-year trainees enroll annually; approximately 75 individuals, including first-, second-, and third-year trainees, are enrolled in total. Nearly all graduates seek and obtain full-time positions in academic medicine. Doctoral (PhD) training also is provided. Approximately 8 to 10 trainees are enrolled at any one time. This program, also designed for those seeking careers as independent academic investigators, includes clinicians interested in training in research methods beyond that which is provided by the MSCE, as well as nonclinicians with strong backgrounds in a biomedical science or another health-related area.
The first students were accepted into the MS and PhD programs in biostatistics in 2000. These graduate training programs are designed for those interested in basic theory and applications of statistical methods, especially as applied to problems in biomedical sciences (MS), and for those who wish to be independent researchers in biostatistical applications and methodology (PhD). Approximately 5 students are enrolled in the MS program, which is designed to be completed in two years, and approximately 25 students are enrolled in the PhD program, which usually takes four or five years to complete. All graduates from these programs have found positions as biostatisticians in academia, private industry, or government.
The CCEB also conducts the clinical evaluative sciences curriculum required for all medical students, consisting of two required courses during the preclinical years: (1) Introduction to Clinical Epidemiology and Biostatistics, and (2) Clinical Decision Making. The first course meets during the fall of the first year. Its objective is to teach medical students to apply the basic principles of epidemiology and biostatistics to the evaluation of clinical research studies and to begin to consider how to translate results into the practice of medicine. The second course meets during the spring of the first year and is designed to teach students to find and use scientific information in the context of clinical decisions. Elective opportunities also are available to medical students, including summer research rotations after the first year and one- to three-month independent studies during the third and fourth years.
CCEB faculty also lead a series of clinical research methods courses leading to a clinical research certificate designed primarily for medical residents, fellows, and faculty members without prior clinical research training who desire careers in collaborative research. Those who complete courses covering research methods, biostatistics, practical applications in clinical research, and scientific and ethical research conduct (four or five courses, depending on specific courses taken) and also complete curricula on the responsible conduct of research and protection of human subjects5 are eligible to receive a clinical research certificate. Eight courses are taught annually: introductory research methods (taught twice), introductory biostatistics (taught twice), practical applications, critical appraisal, database management, and clinical trials/translational research.
The CCEB's Impact on the School of Medicine
The CCEB's 166 core and affiliated faculty members represent more than 30 academic units at Penn, and there are 115 former CCEB trainees who now hold faculty appointments in the school of medicine. This distribution of faculty members, as well as the CCEB's matrix organization, facilitates the conduct of clinical research across the school. Biostatistics and epidemiology faculty members collaborate with investigators representing nearly every clinical department in the school and most divisions within each of these departments. Many core CCEB faculty members are now in leadership positions (including four vice or associate deans), and the integration between clinical and population medicine continues—for instance, the head of epidemiology is a nephrologist, appointed in the Renal-Electrolyte and Hypertension Division of the Department of Medicine. His deputies are in the Cardiology Division and the Infectious Diseases Division.
Epidemiologic and biostatistical research have become engrained in the Penn culture; these disciplines are recognized and faculty members are rewarded for their contributions through Penn's promotion process. Faculty members are appointed and promoted within one of two available academic tracks. The “tenure track” requires that the majority of effort be committed to leading empirical or methodologic research; collaborative work is expected at a lower percentage. By contrast, the “clinician educator” track requires that the majority of effort be committed to collaborative research; independent work is expected at a lower percentage. Faculty members in both tracks are members of the standing faculty. The majority of epidemiology faculty members focus on leading their own research and are in the tenure track. The majority of biostatistics faculty members focus primarily on collaborative research and are in the clinician educator track.
CCEB faculty members manage large and diverse research programs of their own, including primary research in biostatistics and epidemiology, and epidemiologic investigations that are both hospital- and community based. Some investigations are more clinically oriented, such as studies of the utility of diagnostic tests and the economic impact of medical interventions used to manage disease. Others are traditional epidemiologic studies of the etiology, frequency, prognosis, and treatment of disease. Many investigations focus on content-specific areas of interest shared by multiple faculty members and include relevant work developing new research methods. Long-standing and mature primary research programs include investigations in cancer, causal inference, dermatology, gastrointestinal diseases, genetics and complex traits, infectious diseases, injury, pharmacoepidemiology, pharmacogenomics, pulmonary diseases, renal medicine, reproductive medicine, statistical genetics and genomics, and therapeutics. In FY2011, core faculty members led five different multiinvestigator extramural efforts, mostly National Institutes of Health (NIH) program and center grants, and data coordinating centers (DCCs). Most faculty members also collaborate on projects led by others, providing critical assistance to faculty members external to the CCEB. These include hundreds of project grants, many program projects, and other multicomponent grants. In FY2011, CCEB faculty members participated in 21 ongoing funded multicomponent grants, with four first-submission proposals for additional multicomponent grants under review.
The CCEB also houses service centers and cores specifically designed to support the conduct of research projects; many are led by faculty members external to the CCEB, and others are led by CCEB faculty members. The Clinical Research Computing Unit, including 33 staff, provides project and data management, applied informatics, and research computing services, and also leads DCCs for nationally focused, multicenter, clinical research networks, including large-scale, multicenter, clinical trials and observational studies. A second service center, the Biostatistics Analysis Center, which has 32 staff members, provides biostatistical services for collaborative investigations of health outcomes, clinical trials, genomics, large prescription and diagnostic databases, and observational epidemiologic studies. Its cores include the Biostatistics and Epidemiology Consultation Center, which provides short-term research support for research proposals, selection of statistical methods for applications to research data, statistical analysis of data, and statistical review of manuscripts. Its Outcomes Measurement Methods Core assists investigators in the selection, modification, and/or creation of measurement tools needed for the development, conduct, and analysis of translational and clinical research projects. Its Cartographic Modeling Laboratory specializes in spatial analysis using geographic information systems technology and conducts projects featuring mapping and other visual displays of data, Web-based information systems, and innovative analysis using administrative data.
Education and training
Some of the education and training programs that are managed by the CCEB, specifically the MSCE program, the clinical research certificate program, and the clinical evaluative sciences curriculum for medical students, serve as important resources for school of medicine medical students, residents, fellows, and junior faculty.
The MSCE is a major resource for clinical departments throughout the school. More than 100 former MSCE trainees currently hold school of medicine faculty or instructor appointments. Trainees have been supported by 13 NIH institutional training grants (T32) led by CCEB faculty members, specifically designed to support MSCE trainees. Of the 34 MSCE trainees who matriculated in 2010, 23 are clinical fellows at a school-of-medicine-affiliated hospital, 2 hold school of medicine faculty appointments, 1 is a senior research investigator in a Penn-affiliated clinical department, and 1 is a postdoctoral fellow in a Penn-affiliated clinical department. In particular, Penn's clinical fellowship directors perceive the MSCE program to be a valuable recruiting tool for their clinical training programs.
The clinical research certificate program, developed by the CCEB in 2003–2004 at the request of school of medicine clinical division chiefs and fellowship directors, also serves as an important resource for clinical departments. Nearly all students who enroll in these courses have school affiliations. Most are clinical fellows, but many junior faculty members also take these courses. Enrollment in 2010–2011 averaged approximately 25 students per course. To date, 112 certificates have been awarded.
The CCEB's influence in the curriculum for medical students has been important as well. At the time of the creation of the CEU, epidemiology content was included as a minor component of two basic science courses taught in the first year of medical school. A separate, first-year course in epidemiology became part of the curriculum in 1984. This was expanded to two courses in the late 1990s when the Penn curriculum was overhauled. The existing courses, required for all medical students, consist of approximately 50 hours of contact time in years one and two. Elective opportunities attract approximately 5 first-year students annually for a funded summer epidemiology experience and up to 20 third- and fourth-year students annually for one- to three-month mentored research experiences. Finally, with the large number of clinician epidemiology faculty members, there is a considerable probability that a student will have one or more clinical epidemiologists as a ward or clinic attending.
The resources provided by other CCEB training programs, specifically the graduate programs in biostatistics and the PhD program in epidemiology, serve a narrower school of medicine constituency. Many of these trainees serve as research assistants, and others are consultants on ongoing research projects. Several graduates of the MS in biostatistics program work at Penn, including some who provide staff support for the CCEB. Five core CCEB faculty members are graduates of the PhD in epidemiology program, and another has an affiliated appointment.
Two additional programs, FOCUS on Health and Leadership for Women and Bridging the Gaps (BTG), are housed within the CCEB. FOCUS was founded in the CCEB and now has grown into a nationally recognized school of medicine dean-funded program designed to improve the recruitment, retention, advancement, and leadership of women faculty members and to promote women's health research. In addition to numerous innovative faculty development initiatives, FOCUS obtains extramural funds for pilot grants to support junior faculty members' research and medical student fellowships for six-month, mentored research projects in clinical and basic science areas of women's health. BTG is a program that links the training of health and social service professionals with the provision of health-related service and the promotion of public health among vulnerable populations. It is a multiinstitutional collaborative effort with programs in Philadelphia (including all of Philadelphia's academic health centers), Erie, Pittsburgh, and the Lehigh Valley, as well as New Jersey. The program collaborates with over 100 community organizations serving vulnerable populations each year. Of the 3,589 students who have participated in the BTG Community Health Internship Program from 1990 to 2010, 795 were Penn students, and, of those, 321 were Penn medical students. There are two additional Philadelphia BTG program components (seminar series and clinical program), which are smaller and also are available to Penn medical students.
In FY2011, the CCEB's expenditures from extramural funding were $48.6 million. The CCEB's service center expenditures were $9.4 million, virtually all extramural in origin, which supported investigators in most school departments. Fifty-four percent of this funding supported projects based in the CCEB and is included in the CCEB expenditures above. Tuition revenue generated by the CCEB's educational programs in FY2011 was $4.1 million. Twenty-seven percent of this was provided through CCEB awards received and is included in the CCEB expenditures above.
Critical Success Factors
The CCEB's success and ability to thrive are due primarily to three factors: structure, environment, and funding. As described above, the CCEB is structured as an umbrella organization under which all clinician epidemiologists, nonclinician epidemiologists, and biostatisticians at Penn work collaboratively. Core CCEB faculty members look to the CCEB, rather than the DBE or their clinical department, for their primary research support, including space, research computing, secretarial/administrative assistance, and financial management. Although the promotion and tenure process is managed by departments, including the DBE, the CCEB's assessment of a core faculty member's contribution to the CCEB's mission and academic productivity is crucial in this process. Further, the CCEB is a single resource for Penn faculty members seeking epidemiologic and biostatistical research support and expertise. The CCEB provides easy access to this support and expertise and has created service centers to aid the conduct of research for faculty members who do not want to hire and train their own staff.
Penn's environment has facilitated the CCEB's growth and development. The CCEB benefits from Penn's resources as well as from the rich resources of the school of medicine and two nationally ranked hospitals: the Hospital of the University of Pennsylvania and the Children's Hospital of Philadelphia. All 12 Penn schools and many of the affiliated hospitals are located within close proximity on Penn's compact campus, facilitating interaction and collaboration. Further, research and research training are valued enterprises at Penn. Clinical and basic research flourish side by side. The CCEB also has been fortunate to work within an academic community in which the disciplines of epidemiology and biostatistics have been accepted, although not without occasional struggle.
Population-based research, including epidemiology and biostatistics, has grown nationwide in importance and influence in the last 30 years.6 The CCEB was fortunate in the timing of its development; it emerged as research funding opportunities expanded. Calls to increase clinical research and clinical research training by renowned researchers, Institute of Medicine committees, and the NIH were timely.7–12 With this came recognition of the central role of epidemiology and biostatistics in conducting clinical research. The NIH budget doubled between 1998 and 2003. Also, T32 funding has become more available for clinical research training programs.
Conducting clinical research well and properly is expensive. This has been exacerbated in recent years with regulatory demands that have increased costs for conducting clinical research. Although space and equipment costs for clinical research are very low compared with those of basic science research, personnel costs are much higher. The conduct of clinical research requires large research and administrative staffs to manage projects, collect and manage data, oversee budgets, allocate costs, and track faculty members' effort, given their participation in multiple grants and their constantly changing efforts. Further, although the grants that CCEB faculty members lead and/or participate in provide substantial indirect funds to the institution, most administrative staff cannot be supported directly on federally funded grants. CCEB leaders sometimes have struggled with school administrators to justify infrastructure costs that require support from school sources. They also sometimes have encountered school faculty members who desire to use CCEB service centers but who are reluctant to acknowledge the full costs to run these service centers, which are reflected in the rates charged for services. Finally, when appropriate, we need to support national efforts led by clinical research advocates to review federal policies that act as barriers to human subjects research.
We also have concerns regarding the ways in which faculty tracks are perceived, both external and internal to the school of medicine. As above, the tenure and clinician educator tracks are targeted for equally outstanding faculty members whose foci are different. However, it can be difficult to explain the clinician educator track to potential faculty recruits, especially biostatistics recruits, whose role would primarily focus on collaborative research. Many recruits are concerned that a clinician educator track appointment is similar to an appointment not in the standing faculty, such as a position in Penn's research track, which lacks the stability of a standing faculty position. Further, we continue to encounter faculty members outside of the CCEB who consider clinical research to be inherently inferior to laboratory research and, therefore, not deserving of tenure-track positions. Such opinions make more difficult the recruitment and retention of the best clinical research scientists. Responding to these viewpoints continues to be an ongoing effort.
The Current National Environment and the CCEB as an Organization Model
With the formation of the CEU and, later, the CCEB, Penn's school of medicine became one of the first medical schools to integrate population medicine and clinical medicine, although many analogous programs at schools of medicine have since been formed. The CCEB and similar organizations are well positioned to respond to the increased national effort focusing on clinical research, as demonstrated by the 59 K30 programs awarded by the NIH between 1999 and 2004 and the 60 CTSAs granted by the NIH since 2006. Further demonstration of this heightened national interest is reflected in the priorities for the NIH expressed by Francis S. Collins, MD, PhD, the NIH's director. Three of the five priorities he has identified are directly relevant to population-based research: the translation of research into medicine; putting science to work for health care reform through comparative effectiveness research, behavioral science, health information technology, health research economics, and research on health disparities; and global health.
Although not all population-based research units within U.S. schools of medicine have access to some of the environmental advantages we describe above, notably Penn's compact campus and access to excellent multidisciplinary collaborators, Penn's structural model has wide applicability. Population medicine and clinical medicine can be integrated, even in a traditional, departmentally oriented institution such as Penn. The key components are (1) an interdepartmental center to link clinical epidemiologists, doctoral-trained epidemiologists, and biostatisticians, sufficiently empowered and resourced to support these academic disciplines, and (2) a basic science department to provide an academic home for nonclinical faculty members.
The CCEB's success in bridging the gulf that existed between clinical medicine and public health is due, initially, to the vision of the Dana, Rockefeller, and Mellon foundations and their willingness to provide financial support sufficient to create the initial organizational structure for clinical epidemiology at Penn. Building on this foundation funding, the dean's support in the early 1990s of the concept to develop the CCEB and DBE, and his willingness to fund a major expansion in both biostatistics and clinical epidemiology, solidified this vision. The effort of the CCEB faculty and their dedication to clinical research brought this vision to reality.
The authors wish to acknowledge the contributions of Paul D. Stolley, MD, MPH, founding director of the Clinical Epidemiology Unit of the University of Pennsylvania School of Medicine, which later evolved into the Center for Clinical Epidemiology and Biostatistics. The authors also would like to thank Dean William N. Kelley, whose initial trust and investment in the CCEB and its leadership made this possible, as well as Dean Arthur H. Rubenstein for his continued support after Dean Kelley stepped down. Finally, the authors would like to thank Kerr L. White, MD, whose vision to bring population medicine back to clinical medicine was central to the philosophy that guided the formation of the CCEB and DBE.