The work of the National Institutes of Health (NIH)—a U.S. research establishment whose agenda ranges from the very basic to the highly applied—has long been recognized as critical to advancing the quality of health care in the nation and the world. The virtual elimination of polio resulted from basic and applied research in virology; the development of blood pressure–lowering drugs came from an understanding of the fundamental regulation of the underlying biological process; the development of cholesterol-lowering drugs is due to studies of the transport and enzymes controlling cholesterol flux in humans; the development of new imaging techniques has led to improved diagnostic procedures; and improved treatments of mental disease are due to a better understanding of brain function. As a result of NIH research, major diseases such as AIDS, stroke, congestive heart failure, and diabetes are treated more successfully each year. The net result has been a dramatic increase in Americans’ longevity. At the turn of the 20th century, life expectancy for women and men was 53 and 50, respectively; in the year 2000 it was 80 and 74.1
Virtually all of these improvements in health care were derived from basic research that led to an understanding of human physiology. In many cases the basic research occurred decades before its application, often with little or no obvious expectation that an application to health care might develop. Who could have imagined, for example, that being able to orient hydrogen nuclei in a magnetic field would lead to today’s magnetic resonance imaging techniques? Or that understanding the basis of enzyme action and regulation in bacteria would lead to specific drugs? One might go so far as to say that a greater understanding of basic physiology is the key to successful medical applications and that this knowledge can only come from research.
The work goes on, in new and constantly evolving ways, to keep improving the methods and outcomes of health care. The sequence of the human genome and the linkage of specific genome sequences with diseases such as Alzheimer’s, cystic fibrosis, and many others are intriguing developments. Similarly, the linkage of genome sequences with mental disorders promises greater understanding—and, ultimately, improved treatment—of such diseases as major depressive disorders, which affect 5 percent of the population (9.9 million Americans) annually.2 Meanwhile, the addition of computer science and bioinformatics to the arsenal of biomedical, social and behavioral, and clinical research holds enormous promise and is stirring considerable excitement among scientists. Further research will undoubtedly lead to better medical therapies.
To continue to derive these benefits, a highly trained workforce is required. This workforce must have a steady infusion of highly trained people with new approaches if it is to be successful. Support of this workforce’s training, therefore, is an investment in the health of the country.
The National Research Council has been evaluating workforce needs in the biomedical, social and behavioral, and clinical sciences on a continuing basis since 1975, as mandated by Congress. This report and its predecessors monitor the current workforce of these areas and attempt to predict their necessary size and composition for the future. It also makes recommendations on how the National Research Service Award program in particular can optimally contribute to overall training efforts for the biomedical, social and behavioral, and clinical sciences.
Research Training and the National Institutes of Health
The origins of research training at NIH date to 1930, when the Ransdell Act changed the name of the Hygienic Laboratory to the National Institute of Health (a single institute at that time) and authorized the establishment of fellowships
for research into basic biological and medical problems. While the harsh economic realities of the Great Depression imposed constraints, this legislation marked a new commitment to public funding of medical research and training. The National Cancer Act of 1937, which established the National Cancer Institute (NCI) within the Public Health Service (PHS), funded the first training programs targeting a specific area. This legislation supported training facilities and the award of fellowships to outstanding individuals for studies related to the causes and treatment of cancer. In 1938, 17 individuals received fellowships in cancer-related research fields such as biochemistry, physiology, and genetics.
NCI became part of NIH with the passage of the Public Health Services Act of 1944—the legislative basis for NIH’s wartime and postwar expansion of research and training programs and more generally for a major federal commitment to support biomedical research. This expansion was supported by legislative actions that converted existing divisions within NIH to institutes and centers and the establishment of new institutes or centers, each with field-specific training and research missions. In particular, the first of these laws—the National Heart Act of 1947—established the National Heart Institute and changed the name of the National Institute of Health to the National Institutes of Health.
Throughout the 1940s, 1950s, and 1960s there was substantial growth in the NIH budget, with annual increases averaging 40 percent from 1957 to 1963 (with dollar increases ranging from $98 million to $930 million). This funding raised the number of grants to academic institutions and enabled greater federal assistance in both the construction of research facilities and the establishment of fellowship and training programs for research personnel; this generous funding even allowed for limited investment to support research in foreign countries. The growth in research and training support slowed in the late 1960s, to about 6 percent annually, with a consequent decline in the number of research grants, both foreign and domestic, and a curtailment of facilities construction.
Support in the 1970s reflected public and congressional interest in specific diseases. Research areas such as cancer and pulmonary and vascular disorders were identified by legislation for increased funding, and the eleventh institute on the NIH campus, the National Institute on Aging (NIA), was established in 1974. The NIA also brought a new perspective to NIH in that it was authorized to support not only biological research but also social and behavioral research. While funding for research in targeted areas was welcomed at NIH, this also meant that research in less visible areas tended to decline. Institutes such as the National Institute for General Medical Sciences (NIGMS) and the National Institute of Allergy and Infectious Diseases saw annual average reductions of about 10 percent.
By the early 1970s, training support was authorized through the different institutes and centers by 11 separate pieces of legislation. However, in its FY 1974 budget recommendations, the administration proposed the phasing out of research training and fellowship programs over a five-year period by making no new awards and honoring only existing commitments. The reasons it cited for this proposal were as follows: the need for such programs and the manpower trained by them had never been adequately justified; people trained in these programs earned incomes later in life that made it reasonable to ask them to bear the cost of their training; large numbers of those trained did not enter biomedical research or continue their training; alternative federal programs of support for this training were available; and the programs were not equitable because support was not available equally to all students.3
The administration’s proposal met with virtually universal opposition by members of the nation’s biomedical research community. As a result, the administration revised its position and proposed a new, but smaller, fellowship program at the postdoctoral level. This proposal also met with objections, and in 1974 Congress enacted the National Research Act (P.L. 93-348), which amended the Public Health Services Act by repealing existing research training and fellowship authorities and consolidating them into the National Research Service Award (NRSA) program. The legislation authorized support for individual and institutional training grants at the predoctoral and postdoctoral levels, with the stipulation that an individual could be supported for no more than 3 years. Moreover, to safeguard against some of the cited abuses of the former programs, it restricted training support on the basis of subject-area shortages and imposed service obligations and payback requirements.
In the years since the National Research Act was signed, the law governing the NRSA program has been modified several times in order to include new areas of research training and establish funding levels for selected disciplines. The first change came in 1976, when Congress extended the program to encompass research training in nursing.4 Then, in 1978, Congress expanded the NRSA program to cover training in health services research.5 In 1985 the program was enlarged once again to include training in primary care research.6
Specific funding targets for training in health services and primary care research were established with the Health Research Extension Act of 1985, when Congress required that 0.5 percent of NRSA funds be allocated to each of the two fields.7 The same law directed that funds for training in health services research be administered by the Agency for Health Care Policy and Research (AHCPR) and its successor, the Agency for Healthcare Research and Quality (AHRQ). Research training in primary care originally came
under the purview of NIH but was delegated to the Health Resources and Services Administration by Congress in 1988 after concerns were raised that NIH was interpreting the meaning of “primary care” too broadly.8,9 Funding levels for training in health services and primary care research were increased to 1 percent of the NRSA budget with the passage of the NIH Revitalization Act of 1993, and these two fields remain the only ones for which specific funding levels have been established by law.10
Minority Programs at NIH
The recruitment of underrepresented minorities into research careers has been a long-standing activity at NIH. In 1972, about the time the NRSA program was established, the Minority Schools Biomedical Support program—under the administration of the NIH Division of Research Resources—began awarding grants to faculty and students at minority institutions. That same year research awards were made to minority faculty under the Minority Access to Research Careers (MARC) Visiting Scientist and Faculty Fellowship program; and in 1974, MARC was officially established within NIGMS as a formal program to stimulate undergraduates’ interest in biomedical research and to assist minority institutions in developing strong undergraduate curricula in the biomedical sciences. In 1977 the MARC Honors Undergraduate Research Training (HURT) program was established, and in 1981 the MARC Predoctoral Fellowship program was created to provide further incentive for graduates of the HURT program to obtain research training in the nation’s best graduate programs.
These programs continue today with some modifications, such as the replacement of the MARC HURT program with the MARC Undergraduate Student Training in Academic Research program, designed to help meet the need for continual improvement in institutional offerings. Other additions have included the Post-Baccalaureate Research Education Program Award, MARC Faculty Predoctoral Fellowships, MARC Faculty Senior Fellowships, MARC Visiting Scientist Fellowships, and MARC Ancillary Training Activities.
Concurrent with the growth of the MARC programs, the Minority Schools Biomedical Support program also has been evolving. When eligibility for the program was expanded in 1973, it was renamed the Minority Biological Support program; its name was changed again in 1982 to the Minority Biological Research Support (MBRS) program in order to reflect its research scope. This MBRS program was transferred to NIGMS from the Division of Research Resources in 1988, and the NIGMS established the Minority Opportunities in Research (MORE) program branch to serve as the focal point for efforts across NIH to increase the number and capabilities of minority individuals engaged in biomedical research and teaching. In 1996 the MORE Faculty Development and Initiative for Minority Student Development awards were established, and in 1998 the Institutional Research and Academic Career Development Award was announced to encourage postdoctoral candidates’ progress toward research and teaching careers in academia. Both the MARC and the MBRS programs have benefited from their coordination in NIGMS and the regular conferences that are held to promote program activities.
In 1989, NIH introduced the Research Supplements for Underrepresented Minorities program. This program allows principal investigators interested in mentoring minorities to add students or researchers to an existing grant by applying for a supplement. Individuals from the high school to junior faculty level are eligible, but nearly 90 percent of the awards are for training at the predoctoral, postdoctoral, and faculty levels. Like others offered by NIH, this program provides an opportunity for promising minority researchers to gain experience that will help them build a research career—but it is not a vehicle for attracting minorities into science, as seen by the low participation rate at the precollege level. The fact that minority candidates initiate the process by expressing their interest to a principal investigator may be the reason for the focus at the higher career levels; about 35 percent of the program’s funding is for faculty support.
Another action that helped increase minority participation was an NIH directive in the mid-1990s to encourage training directors to support more minorities on their grants. As a result, the percentage of minorities on training grants more than quadrupled from the 1990s (2 percent) to 2003 (9 percent).
Career Development Programs
While the education and training of graduate students or postdoctoral researchers prepare individuals for research careers, in the 1980s NIH recognized the need for programs to help individuals establish strong and productive research careers. To that end, it began programs to facilitate the transition from trainee to research status and give established scientists the opportunity to pursue new research directions.
One such mechanism, referred to as the K award, has the goal of providing Ph.D. scientists with the advanced research training and additional experiences needed to become independent investigators and holders of clinical degrees with the research training needed to conduct patient-oriented research. Over 20 different K award types have existed over the years, though the number at any one time has varied—with some being replaced by other awards or combined to make administration more efficient and others being created to address special needs. Currently, there are 15 K awards, which target different parts of the research population with different forms of support and training features. Not all NIH institutes or centers offer K awards. However, for those that
do, they are only for research training in areas that are consistent with the mission of that particular institute or center.
The NIGMS established the Medical Scientist Training Program (MSTP) in 1964 to encourage research training that would lead to the combined M.D./Ph.D. degree. The program was designed to train investigators who could better bridge the gap between basic science and clinical research by providing both graduate training in the biomedical sciences and clinical training offered through medical schools. Students receive a combined M.D./Ph.D. in an average time of about 8 years. The impact of this specialized training program on increasing the productivity of physician-scientists is noteworthy: M.D./Ph.D.s make up about 2.5 percent of medical school graduates each year, yet they hold approximately 33 percent of the NIH grants going to physician-scientists.
What began in 1964 with three programs—at the Albert Einstein College of Medicine, Northwestern University, and New York University—has now grown to 32 MSTP programs, and almost all of the programs have had continuous funding from their beginning. There were 66 trainees in the program in 1964, and by 2000 the number had grown to over 900 and about 300 program graduates. The establishment of these programs has also prompted institutions to develop additional dual-degree programs that are not MSTP funded.
Since the inception of the MSTP, several assessments documenting the success of the programs have been conducted. The most recent study in 1998 included graduates of all of the funded MSTP programs, graduates of non-MSTP programs, and graduates of MSTP programs with either an M.D. or a Ph.D. Over a range of measures, the study showed that MSTP graduates appear to have been successful in establishing research careers, and their recent publication records suggest that members of all cohorts continue to be productive researchers. While the program has clearly been a success, the level of funding for it has not kept pace with the overall NIH budget.
NATIONAL RESEARCH SERVICE AWARD PROGRAM
In its nearly 30-year history, the NRSA program has provided research training in the biomedical, clinical, and behavioral and social sciences to more than 140,000 students and young investigators. It has done so through a combination of individual fellowship awards and institutional training grants to some 465 universities, research institutes, and teaching hospitals. Moreover, as the NIH and the PHS have grown over the last quarter-century, the NRSA program has evolved to encompass new fields in the basic biomedical sciences such as genome research and neuroscience. NRSA has also expanded in breadth to include research training in fields such as communication disorders, health services research, primary care, and nursing.
Since 1980, when the NRSA program was well under way, the number of individual training positions has increased by over 4,500, or about 40 percent—from 5,884 predoctoral and 6,173 postdoctoral positions (for a total of 12,057) to 9,308 predoctoral and 7,457 postdoctoral positions (for a total of 16,765) in 2003, as shown in Figure 1-1. Over the same period, the training budget at NIH increased from about $196 million to $681 million, though in 1980 dollars the 2003 expenditure for training was just a little over $230 million.11 Although the NIH training budget grew slightly over that period, it did not kept pace with the overall NIH budget. Between 1980 and 1998, for example, while the NIH budget went from $3 billion to $5.6 billion (in 1980 dollars), the training budget declined from $196 million to $177 million. By another measure, the NIH training budget declined from a high in 1981 of 7.9 percent of the extramural research grant funding (that is, for work done in institutions outside the NIH campus) to 4.3 percent in 1998. The training budget leveled off at 3.8 percent in 2003.
There was significant change, however, from 1998 to 2003, during which time the training budget increased by about $253 million, or 59 percent, in current dollars. This gain resulted in part from growth in the number of predoctoral and postdoctoral awards (about 1,000 and 500, respectively) as well as from a significant increase in the NRSA stipends. These recent increases in the training budget are more in line with the total NIH budget, which doubled—from $13.6 billion to $27.2 billion—during those 5 years.
Data on the number of awards for 2004 are not complete, but the FY 2004 budget for NIH sets the full-time NRSA training positions at 17,197. This marks an increase of 80 budgeted positions over 2003 and a training budget of $716 million, up 3.9 percent over 2003.12 The president’s budget request for FY 2005 again has a proposed increase for NRSA training. The request for training positions at the individual and institutional levels is up by 225 from 2004, with a budget of $764 million, and the budget proposes to hold the stipends at 2004 levels for both predoctoral and postdoctoral positions. However, no information is currently available on the actual number of positions that will be supported for 2005. In these difficult budgetary times, with many agencies receiving cuts, NIH and health-related research is still in a preferential position, but it is unclear what affect the small increases in the NIH budget will have on research training or even research in general in the future.
Institutional training grants, which funded the education
of about 84 percent of NRSA participants in 2003, are widely regarded as one of the best vehicles for learning the theories and techniques of biomedical and behavioral research.13,14 The NIH, as well as the AHRQ and the Health Resources and Services Administration, make such awards only after competitive review; institutions are assessed and compared on their previous records in research training, the objectives and designs of their programs, the caliber of their preceptors, their ability to recruit high-quality trainees, their institutional environment, and their commitment to research training. Over the past few years the success rate of proposals for T32 training grants has averaged 62 percent, with the rate dropping to 53 percent in 2003 as the result of a significant increase in the number of proposals. The number of proposals went from an average of 570 between 1996 and 2001 to 666 in 2002 and 753 in 2003.
The remaining 16 percent of the awards in 2003 were individual fellowships, which are also awarded on a competitive basis and provide what is often a first step toward professional independence. Fellows at the predoctoral and postdoctoral levels develop their own proposals and, once an award has been made, are generally accorded a great deal of autonomy in pursuing their educational and research goals. Through much of the 1990s, the majority of these awards—about 73 percent—were at the postdoctoral level, but in 2003 the percentage dropped to 64 percent. For example, applications for F3215 postdoctoral fellowships declined from 2,556 in 1996 to 1,552 in 2002; in turn, the number of awards also declined—from 968 to 614. The reasons for these declines are unclear but could possibly have resulted from the low NRSA stipends before the NIH increases that set $45,000 as a five-year goal. A recent turnaround in the number of applications brought the 2003 total to 1,949 and the 2004 level to more than 2,000, thus possibly supporting the stipend-level conjecture. As a result of these increased application levels, the number of F32 awards increased from 614 in 2002 to 713 in 2003. Meanwhile, applications for F31 predoctoral awards increased steadily between 1990 and 2003 (from 275 to over 972, respectively), and the number of grants awarded grew from 141 to 414.
The NRSA program accounts for only 22 percent of NIH’s total funding for graduate education in the biomedical, behavioral, and clinical sciences. Nevertheless, the pro-
National Research Council. 1995.
National Research Council. 1998c.
See Appendix B for a complete explanation of the NRSA awards.
gram occupies a leadership role in research training in these fields. NRSA awards are important as they:
Serve to attract quality people into biomedical research. Perhaps the best example of this phenomenon is the MSTP, which has a well-established track record for launching physicians into outstanding research careers.
Help direct training into specific research areas, such as cell biology or biophysics.
Establish training standards—the requirements imposed on individuals supported by NRSA training grants are also imposed on trainees supported by other means.
Offer the possibility of providing support for training in emerging areas for which other mechanisms may not be available.
Provide graduate students, during the early years of their training, the opportunity to explore different areas of research.
While NRSA grants are awarded with the expressed purpose of providing research training, not all supported graduate students or postdoctorates actually pursue research careers as independent investigators or researchers. This is documented by the 1998 analysis of the career progression of NRSA predoctoral trainees16 and may be even more the case in the current employment market, where few tenure-track faculty positions are available to new doctorates and postdoctorates.
Over the years the NRSA program has been responsible for several improvements and new developments in research training. For example, NIGMS has focused on multidisciplinary research training right from its first NRSA training grant awards in 1975. As a result, almost all universities now structure the education of basic biomedical scientists in a way that cuts across disciplinary and departmental lines. Similarly, after the NRSA program required (in 1990) that institutional training grants provide education in the responsible conduct of research, most universities began to offer such instruction to all students and fellows engaged in health research training, regardless of the source of their financial support. Finally, although M.D./Ph.D. training had been introduced by NIH prior to the passage of the National Research Act in 1974, dual-degree training programs have grown considerably since; they are now supported not only by NRSA funds but by private and institutional sources as well. In addition, dual-degree training has been extended to the oral health sciences. In 1996 the National Institute of Dental Research established a D.D.S./Ph.D. program.
To lessen the sizable personal investment required of those who pursue careers in biomedical, behavioral, or clinical research, the NRSA program provides its fellows and trainees with stipends and tuition subsidies. For predoctoral NRSA recipients, at least, the results have been measurable. Graduate students with NRSA support are more likely to begin their careers without education debt than their fellow Ph.D.s in the life and social sciences. The effects of education debt are particularly striking for those who participate in NIH-sponsored M.D./Ph.D. programs: in 1996 their median debt was $1,000, in contrast to their medical school classmates who graduated with a single degree and whose median debt approached $60,000.
Beyond the financial differences, there are a number of other distinctions. A 1984 evaluation of formal NIH-sponsored research training, which included programs predating the establishment of the NRSA, found that participants in the programs were more likely than their counterparts to complete their doctoral programs and go on to NIH-supported postdoctoral training. Further, those supported by NIH during their predoctoral studies applied for and received NIH research grants with greater success, authored more articles, and were cited more often by their peers. At the postdoctoral level, regardless of whether trainees were appointed to institutional training grants or had received individual fellowship awards, they were more likely to pursue research careers than their colleagues without formal NIH research training. These differences were true for M.D.s with postdoctoral research training as well as Ph.D.s.
A follow-up to the 1984 evaluation of the NRSA predoctoral program was conducted in 1998 to examine the characteristics of doctorates from three groups between 1981 and 1992: NRSA-supported doctorates, Ph.D.s at institutions with NRSA training grants who did not receive this type of support, and doctorates at institutions without NRSA grants.17 Many of the same conclusions were reached as in the 1984 study. In particular, it revealed the following:
As an indicator of the high quality of the NRSA program, 80 percent of the NRSA trainees or fellows received their Ph.D.s from 50 institutions in the top quarter of the biomedical sciences programs, and nearly 60 percent received their degrees from the top 25 institutions.
The completion rate for students supported by the NRSA program was an estimated 76 percent, a figure comparable to those of other merit-based national fellowship programs and students in high-quality doctoral programs.
NRSA trainees and fellows spent less time enrolled in graduate school than their non-NRSA-supported counterparts—three months less than doctorates without NRSA support at their institutions and seven months less than those at institutions without NRSA grants.
Nearly 58 percent of the NRSA trainees and fellows received their doctorates by the age of 30, compared to 38.9 and 32.3 percent, respectively, for the non-NRSA-supported doctorates from NRSA institutions and non-NRSA institutions.
Following graduate school, NRSA recipients were, by some measures, more likely to make progress toward re-
search careers. In fields in which postdoctoral study is very common, 93 percent of the NRSA trainees and fellows reported definite postdoctoral commitments, compared to 84 and 80 percent, respectively, at NRSA institutions and non-NRSA institutions.
NRSA trainees and fellows appeared to be more likely to move into faculty or research positions. About 37 percent of the NRSA recipients held faculty positions seven to 8 years past the doctorate, compared to 24 and 16 percent, respectively, for their non-NRSA-supported counterparts at NRSA institutions and non-NRSA institutions. Similarly, 87 percent of the NRSA trainees and fellows, compared to 77 and 72 percent, respectively, for the NRSA and non-NRSA institutions, were in research-related positions in academia, industry, or other research settings.
NRSA trainees and fellows were more likely to have grants. For example, among the doctorates who received their degrees between 1981 and 1988 and applied to NIH by 1994 for research grant support, the success rate for NRSA recipients was 67 percent, compared to 55 and 47 percent, respectively, for the NRSA and non-NRSA institution graduates.
NRSA trainees and fellows had more publications—another indicator of productivity.
NRSA predoctoral trainees and fellows in the 1981–1982 cohort had a median number of publications (8.5) that was more than twice that of doctorates from institutions without NRSA grants (4) and 70 percent more than that of non-NRSA-supported Ph.D.s at NRSA institutions (5).
Such studies do not explain whether the success of former NRSA trainees and fellows reflects the training they received, the selection process, or other factors. Nonetheless, these findings do suggest some of the strengths and achievements of the NRSA program.
STUDY ORIGINS AND RELATED ACTIVITIES
In the legislation that established the NRSA program, Congress decreed that awards be made only in areas for which “there is a need for personnel,” and it directed that the National Academy of Sciences provide periodic guidance on the fields in which researchers were likely to be needed and the numbers that should be trained (see Box 1-1). In response, the National Research Council (NRC)—the operating arm of the National Academy of Sciences, Institute of Medicine, and National Academy of Engineering—has issued regular reports since 1975 on the supply of biomedical and behavioral researchers in the United States and the likely demand for new investigators. The present study is the twelfth one completed by the NRC.
In each of the 11 assessments of national need for research personnel in the biomedical and social and behavioral sciences submitted thus far by the NRC, the congressional committees adhered to the purpose of those assessments and forwarded the information to NIH and Congress for their use in making budgetary decisions. However, the manner in which the assessments should be conducted and the scope of the investigations were left to the discretion of the NRC. As a result, in many of the reports the characteristics, mecha-
Sec. 472. (a) (3) Effective July 1, 1975, National Research Service Award may be made for research or research training in only those subject areas for which, as determined under section 473, there is a need for personnel.
Sec. 473. (a) The Secretary shall, in accordance with subsection (b), arrange for the conduct of a continuing study to—
(c) A Report on the results of the study required under subsection (a) shall be submitted by the Secretary to the Committee on Energy and Commerce of the House of Representatives and the Committee on Labor and Human Resources of the Senate at least once every four years.
nisms, and quality of NRSA training programs were also addressed.
Prior to the first assessment, the NRC conducted a feasibility study in late 1974 to determine if it was possible to collect the data, perform the analyses, and determine the need for biomedical and behavioral research personnel as outlined in the legislation. In early 1975 the committee for this study returned a report18 which recommended that a follow-on study be conducted. While it recognized the difficulties of the task, the committee viewed it as necessary and feasible. Shortly after conclusion of the feasibility study, a committee was formed to conduct the first study. This study was preliminary and was to be completed in 90 days. The committee viewed its task as one of bringing readily available data and professional judgment to bear on the study requirements and looked to the 1976 study and beyond to provide the basis on which the effectiveness of continuing studies can be judged. The 1975 assessment limited its scope to the demand for faculty, basing findings on federal support for university-based research and enrollments in higher education.19 With this information, it recommended that the training levels in FY 1976 be the same as in FY 1975, while emphasizing the importance of vesting quality in a smaller number of training programs and the need for a balance between supply and demand. It also set a precedent for later reports by broadly interpreting the research areas to include the basic biomedical sciences, the behavioral and social sciences, the clinical sciences and health services research and by providing training levels in each field. This committee, as well as later committees, noted the difficulties of making personnel projections on the basis of the available data; it singled out in particular the lack of data on medical doctors doing either basic biomedical or clinical research. It also drew attention to the growing postdoctoral pool in the early 1970s; while this population seemed to fall within reasonable bounds, the committee expressed concerns about its future size.
Initially the studies were conducted on an annual basis by a standing committee. In its first full-length report in 1976 the committee concluded that Ph.D. production in most areas was more than adequate to meet existing demand, though it recommended some changes from the 1975 level—in particular, that the number of predoctoral training positions in the biomedical sciences be reduced by 10 percent in 1976 and remain at that level through 1978. In the behavioral and social sciences, the committee recommended a dramatic shift from predoctoral to postdoctoral institutional training positions. While in 1975 the apportionment was about 90 percent predoctoral and 10 percent postdoctoral, the committee suggested that the proportion in the behavioral and social sciences instead be 30 percent predoctoral and 70 percent postdoctoral. Its reasoning was based on the perceived need for research in special areas of health and the belief that individuals with a Ph.D. or an M.D. are better able to address those needs. The change was not to affect the funding level and was to take place over time, with an annual reduction of 250 to 350 predoctoral positions and an increase of 150 to 200 postdoctoral positions. Over the next few years NIH partially implemented this recommendation by cutting support at the predoctoral level from 1,754 in 1975 to 653 in 1980, but it increased support at the postdoctoral level from 212 in 1975 to only 349 in 1980. In addition, because the committee believed that the clinical area was one where there was a need for more support to increase the flow of M.D.s into clinical research careers, it recommended a 10 percent annual growth rate through 1978 at both the predoctoral and postdoctoral levels.
In the 1977, 1978, and 1979 studies the committee continued to express concerns about the increased number of doctorates, and the growing postdoctoral pool, in the biomedical sciences. In 1977 it recommended another 10 percent decrease in predoctoral support. In 1978 and 1979 it did not recommend additional decreases due to a concern that the loss of or a reduction in the size of training programs would seriously affect their quality. Each of these studies also reaffirmed the shift from predoctoral to postdoctoral support in the social and behavioral sciences. While recommendations on the number of positions were made for each of the four broad fields—the basic biomedical sciences, the social and behavioral sciences, the clinical sciences, and health services research—the committee in 1978 did draw attention to subfields in the biomedical sciences, such as toxicology and biostatistics, and presented reasons why they should have increased support. The 1979 assessment was the last of the annual reports. In 1978, amendments to the original NRSA Act of 1974 changed the report cycle to every 2 years.
The findings from the next study in 1981 were similar to those of the earlier studies. It expressed concern for career prospects in the biomedical sciences, a “soft” job market in the nonclinical social and behavioral sciences, and a continued need for physician-scientists in health-related research. Thus the committee’s recommendations followed those of earlier studies: no change regarding the biomedical sciences, a shift from predoctoral to postdoctoral support in the social and behavioral sciences, and increases in predoctoral clinical support. The committee also focused on two other needs: to increase minority participation and to stabilize federal support of research training. The committee’s report contained a rather complete list of federal programs aimed at increasing minority participation and recommended that assessments be made of these programs in order to determine their effectiveness. The committee’s concern on the stability issue was that sharp fluctuations in training support could have serious consequences; it therefore suggested that a core level of support, independent of short-term demand, be maintained.
The reports in 1983 and 1985 found signs of increased demand for biomedical research personnel, such as the leveling off of the postdoctoral pool and a decline in the number of graduate students in this area. These committees also believed that recent biological discoveries would create a demand for researchers in industry, government, teaching hospitals, and other settings, and they incorporated employment trends in those sectors into their analyses of national need. By 1985 the biotechnology industry indeed began to recruit significant numbers of Ph.D.s, and that committee called for additional research training in the basic biomedical sciences. In the social and behavioral sciences, the 1983 and 1985 reports restated the need for increased training at the postdoctoral level in clinically related subfields and the return of predoctoral support to the 1981 level. With regard to clinical awards, there was an increase in 1981 over the numbers in the late 1970s, and committees in 1983 and 1985 supported this increase and recommended that it be maintained through 1987. But because they saw a need for increased training of physicians at the postdoctoral level, they stipulated that 85 percent of the positions at this level should go to physicians.
The next (1989) committee report, citing a decline in predoctoral biomedical positions from about 4,000 in 1985 to less than 3,700 in 1987, recommended a significant increase to 5,200 positions by 1995. For the social and behavioral sciences, the committee projected a fairly stable job market and recommended no change in current support levels. With regard to clinical fields, it did expect an increased demand for physician-scientists, but given the lack of reliable data it recommended no change in current support levels.
By 1994 demand for personnel in the biotechnology industry appeared to be slowing, and that year’s committee recommended that NRSA training support in the basic biomedical sciences be maintained at the present levels—5,175 predoctoral and 3,835 postdoctoral trainees—through 1999. The committee also called for a gradual increase in research training in the social and behavioral sciences—at the predoctoral level from 673 positions in 1993 to 900 by 1996 and at the postdoctoral level from 323 to 500. This recommendation was not based on an increased demand for faculty but rather on the continuing gains being made by behavioral scientists in areas of national interest and thus on an anticipated demand for mental health–related research. In the clinical sciences the committee recommended an increase in MSTP grantees from 822 in 1993 to 1,020 in 1996 and a decrease in other clinical programs to offset the MSTP increase. The committee’s 1994 report also expanded the scope of its investigation by highlighting several issues of particular concern to the administration of the NRSA program. These included the growth of the Ph.D. population in the biomedical sciences, the decline in the number of physician-researchers, the recognition that the social and behavioral sciences should play a more important role in health care, the decline in the proportion of graduate students funded by training grants, and the lack of promising research career options for young scientists. These and related issues have subsequently been addressed in later NRSA reports and other policy studies (as discussed below).
In response to the major recommendations put forth by the 1994 study committee, NIH focused on increasing the stipends for trainees and fellows and on evaluating the NRSA program. The committee’s suggestions for maintaining training levels in the basic biomedical sciences and for attracting underrepresented minorities also were pursued. However, recommendations for increasing the number of NRSA training grants and fellowships in the behavioral and clinical sciences, oral health, nursing, and health services research were not acted on; this prompted Congress to request a report on NIH’s implementation of the 1994 study.20 In explaining its actions to Congress, NIH indicated that it had focused on the highest-priority recommendations: it was likely to continue to direct additional research training monies to NRSA stipends until their levels were comparable to those of other sources. The level of support for the social and behavioral sciences, however, was a point of controversy, and this issue appeared again in a minority report in the next study.
The most recent committee to assess the need for research personnel began its work in 1997, concentrating its attention on the three broad fields of biomedical, behavioral, and clinical research, with dental, nursing, and health services research included in the latter category. Two major changes from earlier reports were this committee’s movement away from detailed recommendations on the number of individuals who should be trained under the NRSA program and its use of a demographic life-table model, proposed in the 1994 report, to estimate the size of the workforce each year up to 2005. The life-table model was adopted in 1997 because previous models of supply and demand could not be relied on for valid forecasts. This report’s analysis considered such factors as the average age of current investigators in the biomedical and behavioral sciences, the number of Ph.D.s expected to join the workforce in the years ahead, and the likely effect of retirements and deaths. The committee supplemented this analysis by reviewing indicators of short-term demand—trends in faculty and industry hiring, for example, and perceptions of the job market by recent Ph.D.s.
Implemented for the biomedical and social and behavioral sciences, the model showed that the supply of doctorates, even if at a low level, would be much greater than the demand for researchers during the projection period. This result prompted the committee to suggest that degree production be maintained at current levels in all three broad fields. However, it did recommend increases in clinical research related to patient care and in interdisciplinary research. Many of the committee’s recommendations concerned administration of the NRSA program, and NIH
responded by establishing new guidelines for stipends at the predoctoral and postdoctoral levels, encouraging early completion of education and training and establishing limitations on the period of NRSA support at the predoctoral and postdoctoral levels.
A PERSPECTIVE ON RESEARCH TRAINING ACTIVITIES
Since the establishment of the NRSA program, several reports—other than those mandated by the NRSA Act of 1974—have addressed the state of health-related training, either in general or from the perspective of individual fields. Particular attention has been paid to clinical research and the need to train more physician-scientists. For example, a 1976 study21 by the NRC found that postdoctoral training was essential to the research productivity of M.D.s. Similar results were obtained in a 1986 study by NIH.22 The research activities of medical school faculty were addressed in a 1987 report,23 which found that only 29.5 percent of the M.D.s in departments of medicine devoted 40 percent or more of their time to research and almost 30 percent of the M.D.s spent no time in research at all. These and related clinical research issues were also described in detail by Ahrens in 1992.24
In 1994 the Institute of Medicine (IOM) issued a report, Careers in Clinical Research: Obstacles and Opportunities,25 that made the following recommendations for research training:
Further evaluate clinical research training programs.
Redirect funds to the most effective forms of clinical research training.
Emphasize training programs that provide opportunities to earn advanced degrees in the evaluative sciences.
Increase the number of M.D./Ph.D. and D.D.S./Ph.D. programs that train investigators with expertise in patient-oriented research.
Expand initiatives that reduce education debt, either through tuition subsidies (as in the case of dual-degree programs) or loan forgiveness.
This report was followed in the spring of 1995 by the convening of an NIH director’s panel to review the status of clinical research in the United States and to consider, among other topics, the recruitment and training of future clinical researchers. When the panel presented its report in late 1997, a number of the suggestions for clinical research training paralleled those put forth by the IOM in 1994.26 It included the following recommendations:
Initiate clinical research training programs, such as M.D./Ph.D. programs specifically for clinical research, aimed at medical students.
Ensure that postdoctoral training grants include formal training in clinical research.
Provide new support mechanisms for young and mid-term clinical investigators.
Take steps to reduce the education debt of clinical investigators.
Even before the panel issued its recommendations, NIH made a number of changes to support the training of clinical investigators. These included a program to bring medical and dental students to its Maryland campus for a one- to two-year clinical research training experience and new guidelines from the NIGMS to assure that its M.D./Ph.D. training grants would encourage research training in fields such as computer science, social and behavioral sciences, economics, epidemiology, public health, bioengineering, biostatistics, and bioethics. In response to the panel’s recommendation regarding support mechanisms for young and midcareer investigators, NIH established three new career development awards aimed at advancing careers in clinical research.
Despite these efforts, the issues surrounding clinical research training remained unresolved, so in 1999 the Association of American Medical Colleges, in conjunction with several other organizations, convened a clinical research summit. The resulting report outlined nine core problems and made recommendations for each. For example, it cited the lack of a comprehensive clinical research agenda and suggested that the solution might lie in the formation of a National Clinical Research Roundtable. Such a roundtable was in fact formed by the IOM in 2000, and it has met on a regular basis since then to conduct workshops and discuss issues. Recent workshop reports have addressed topics ranging from the needs of the education infrastructure to engaging the public in the clinical research enterprise.
In 1995 another IOM committee published the results of a study in a related area—the training and supply of health services researchers—and the report27 endorsed the recommendation of the 1994 NRSA “needs” study to substantially increase the number of training positions in health services research. The IOM report also encouraged the AHCPR, now the AHRQ, to focus its training funds on areas in which research was needed—such as outcomes measurement, biostatistics, epidemiology, health economics, and health policy—and to set up institutional training grants for innovative
research training programs. In response, AHCPR made “innovation awards” to 10 institutions in 1998 in order to support the design and implementation of new models of health services research training.28
One of the primary concerns in many of the NRSA “needs” studies for the biomedical sciences was the possible oversupply of researchers, as indicated at various times by the increasing number of doctorates being awarded and the growing size of the postdoctoral pool. While many books and articles, such as In Pursuit of the PhD, discussed the quality of doctoral education,29 it was not until 1995 that the NRC published the findings of a major study directed at graduate training’s effects on career progression. The report, Reshaping the Graduate Education of Scientists and Engineers, reviewed graduate education across the biological, physical, and social and behavioral sciences and engineering and called on universities to offer programs that allow for a broader range of career options; it also called on federal agencies to encourage this trend by supporting graduate education through training grants.30 The report urged universities, government, industry, and professional organizations to work together to develop a national human resources policy for scientists and engineers.
In response to a congressional inquiry about how NIH was planning to adapt its policies to conform with the 1995 NRC report’s recommendations, the agency noted that it would take steps to require institutions with training grants to expose their students to a range of career options. The NIGMS in particular announced new guidelines in 1997 for biomedical sciences graduate programs, so interested trainees might be able to take internships in industry as well as gain experience in teaching. In addition, graduate programs were urged to supply trainees with information on the career histories of previous graduates and to offer workshops on employment opportunities and career counseling.
A 1995 examination of supply and demand regarding Ph.D. scientists, published by Massy and Goldman (RAND Corp.), suggested that U.S. universities were producing more Ph.D.s than necessary in engineering, mathematics, and fields such as the biological and geological sciences.31 Their paper maintained that this overproduction would create a group of chronically underemployed doctorates. An expansion in research funding, furthermore, would likely worsen future job prospects even further as more Ph.D.s were produced. Contending that enrollment of doctoral students in fields such as the biosciences was driven more by the need for research and teaching assistants than by the labor market, Massy and Goldman called for academic restructuring to bring the production of Ph.D.s into balance with demand. The authors noted that improvements in the development and dissemination of data on the scientific and engineering labor market could help restrain production rates, but they ultimately concluded that Ph.D. overproduction would end only if departments were required to reduce their dependence on the research and teaching services provided by doctoral students. In a follow-up report, The PhD Factory,32 in 2001 the authors predicted that without such requirements the “employment gap” (the difference between the supply and the demand) in the biosciences would exceed 25 percent into the future.
In 1996 a consensus conference sponsored by the Federation of American Societies for Experimental Biology addressed some of the issues raised by Massy and Goldman. It concluded by opposing national regulation on the size of graduate programs. However, participants called for data on employment trends to be made available to students and for universities to self-regulate the size of their graduate programs. In addition, institutions were urged to refrain from enrolling graduate students simply to meet teaching or research needs.
In 1994 an NRC report, The Funding of Young Investigators in the Biological and Biomedical Sciences, noted an increase in the age at which young researchers were obtaining their first R01 grant; as a result, an NRC committee was formed to examine the career paths of young investigators. The 1998 report, Trends in the Early Careers of Life Scientists, noted that the number of Ph.D.s awarded annually may be too high and called for restraining the rate of growth in the number of graduate students in the life sciences.33 The report also suggested several policy options for the federal government to consider, such as restricting the number of graduate students supported by research grants and emphasizing research training via training grants and fellowships. However, in the end the committee noted that the rate of Ph.D. production was the cumulative result of individual decisions—by faculty, departments, universities, and students—and it maintained that these groups should bear the primary responsibility for implementing the committee’s recommendations.
As the annual number of new doctorates in the biomedical sciences continued to grow during the 1980s and the 1990s, so did the postdoctoral pool. Aside from concerns about size, the working conditions of postdoctorates and the career guidance they were receiving also became serious issues. The NRC was prompted to release, in 2000, a widely distributed guide that set down a body of principles for advisors, institutions, funding organizations, and disciplinary societies to follow in addressing these issues.34 At that time a number of institutions recognized the problems their postdoctorates were encountering and at a minimum began to count and keep track of postdoctoral appointments. Dedi-
cated offices were established, and local postdoctoral organizations were formed. Eventually, a national association was founded to provide a central voice in helping policy-makers address postdoctoral employment and training issues.
It was not the size alone of the postdoctoral pool that has been of concern but the fact that postdoctorates are not independent researchers and cannot in most cases develop their own research agendas. This is a particular issue at NIH, which conducted a workshop in the fall of 2003 to investigate possible mechanisms for moving researchers it would normally support as postdoctorates into positions of independent research status. NIH also sponsored a meeting at the NRC in June 2004 to address the same subject. The report of that meeting will be published in early 2005.
The significant growth in the number of postdoctorates in the biomedical sciences may have been fueled in part by the growth in the NIH budget. The doubling of the NIH R&D budget—from $13.6 billion in 1998 to $27.2 billion in 2003—resulted in an increased number of new, competing, and noncompeting research grants (33,570 in 1998 to 45,922 in 2003) and an increase in the amount for these grants (from $9.58 billion to $18.04 billion). Although grant amounts have almost doubled, the number of awards has only increased by 36 percent, indicating that more personnel at the graduate and postdoctoral levels will be able to be supported on them. Now that the NIH budget has returned to relatively modest single-digit increases, graduate enrollment and the size of the postdoctoral pool may be reduced—though these effects are yet to be determined. It is also interesting to note that the number of Type I or new awards increased by 45 percent and that the amount of these awards more than doubled during the growth in the NIH budget, but a return to normal growth may have serious consequences for young investigators obtaining independent research status.
A new initiative at NIH that may have an impact on research training is the NIH Roadmap, developed during 2002 and 2003 at the urging of NIH Director Elias A. Zerhouni. This program, resulting from a process of internal discussions and consultation with leaders of the biomedical community, has three major themes:
New Pathways to Discovery—a set of proposals that will create and develop new tools and technologies for bench scientists.
Research Teams of the Future—a set of initiatives to promote public-private partnerships and to train investigators to work in interdisciplinary and multidisciplinary teams.
Reengineering the Clinical Research Enterprise—a complete reorganization, expansion, and streamlining of clinical research in the United States.
NIH has been criticized in the past for its conservative approach—for funding research that may incrementally add to results already in hand. The NIH Roadmap is designed to break with this image and allow for greater innovation. Each of these themes encourages programs that cut across the NIH institutes and centers, with the aim of fostering new interdisciplinary areas of research.
Of the Requests for Applications (RFAs) issued so far, several are for research training. One of these, the Curriculum Development Award in Interdisciplinary Research, supports the creation of courses or curricula at the undergraduate, graduate, or postdoctoral levels that integrate the principles and conceptual approaches of diverse disciplines in emerging areas of biomedical research. Another RFA, Short Programs for Interdisciplinary Research Training, will support grants to develop 8- to 10-week training programs that combine lectures with lab experience. These courses are targeted at not only trainees but also established scientists seeking to develop interdisciplinary research projects. The program is open to a broad spectrum of the scientific community, including foreign as well as domestic institutions.
Using the NRSA mechanism, the National Institute of Mental Health is sponsoring an initiative that seeks to support as many as 40 postdoctoral scientists under an institutional training grant to integrate behavior, environment, and biology into a single discipline. Possibly, the most novel approach is a program called Training for a New Interdisciplinary Research Workforce, which uses a new funding mechanism (the T90) and seeks to capitalize on existing multidisciplinary and interdisciplinary research programs. This new training program, which combines aspects of research grants and NRSA training into a single mechanism, is open to all organizations, foreign and domestic. Trainees can be at any level, from undergraduate to postdoctoral, and do not need to be U.S. citizens or permanent residents.
THE PRESENT STUDY
The purpose of the current study is to advise NIH and Congress on the appropriate level of training for the NRSA program. The committee for this study, in much the same spirit as the preceding study committee, has interpreted its charge more broadly and examined the contributions of the NRSA and career development programs in the general context of health-related research training.
The committee began its work in September 2002 and over the next 28 months held five meetings, drawing on members’ own expertise—as well as information provided by government agencies and disciplinary professional organizations that support health-related research—to determine this study’s findings and recommendations. The committee’s objectives were to:
Estimate future needs for biomedical, behavioral, and clinical research personnel.
Make recommendations about the size and composition of the NRSA program and other NIH research training
and career development programs, using its estimates of future needs as well as information about the rate of Ph.D. production.
Assess the quality of existing NIH research training and career development programs.
Recommend modifications in the existing programs so that they may better prepare the research workforce to meet anticipated needs.
The committee’s initial attention was focused on the biomedical, behavioral, and clinical fields, but it expanded its analysis to include oral health, nursing, and health services research. These three new fields were not addressed in the last report, nor were they initially identified for consideration in the current study. However, the corresponding NIH institutes and centers and the AHRQ asked the committee to examine training in these fields, given an apparent shortage of research personnel in oral health and nursing and a need for more training in health services research.
For this report a life-table model was developed to simulate the changing characteristics of the research workforce in the biomedical, behavioral, and clinical sciences and to assess the future need for personnel in each of these fields. Data from various sources were used to estimate the input of research personnel into the model and to estimate the outputs with mortality and retirement data. In assessing the results of this analysis, several qualifications are in order. First, the model is necessarily fairly simple. Second, available data for the model are incomplete, since it is difficult to determine what proportion of the workforce is actually engaged in research, the number of foreign training scientists that are researchers, and whether available data represent current trends. Third, the models base projections only on the current situation—or, to be more precise, the situation for which the data were valid (typically 2–3 years earlier). Fourth, extrapolation to the future depends in large part on unknown variables, such as the strength of the economy and the amounts of federal spending for research. This analysis was not conducted for oral health, nursing, or health services research for the same reasons as above, and in addition, the results would be even more questionable, since the size of the workforce is small.
It should be noted that the life-table model and much of the analysis in this report address the supply side for research personnel. Metrics for the demand side of the analysis are either unknown or are difficult to quantify. The demand for research personnel could depend on the state of the U.S. economy, the speed at which discovery takes place, levels of research support from government and foundations, the outsourcing of research to foreign countries or the growth of research in foreign countries, and the importance of a field of research in the general scientific and engineering enterprise. The unprecedented growth in the biomedical research workforce in the 1980s and 1990s was fueled by the important discoveries during and before this period and the financial resources that were available to support the research. Predicting this growth in the 1970s on the basis of the current workforce and degree production would have been difficult.
When developing recommendations for the future, it should be remembered that workforce models are only part of the necessary considerations. As shown in the data presented earlier in this chapter, NRSAs support only a relatively small fraction of the total number of people being trained in the biomedical, social and behavioral, and clinical sciences. This is a significant addition to the pool of trained personnel, but the primary role of the NRSA program is not to just add numbers to the pool. It serves several other more important roles. The NRSA program serves as a beacon to attract quality people into biomedical, behavioral, and clinical research. Perhaps the best example of this is the MSTP, which has a well-established track record for launching physicians into outstanding research careers. The program also serves to facilitate training in specific research areas, such as molecular and cell biology and biophysics. These awards also establish training standards that affect not only NRSA awardees but all trainees. Generally, the requirements imposed on individuals supported by training grants, for example, are also imposed on trainees supported by other means. Finally, they offer the possibility of providing support for training in emerging areas for which other mechanisms may not be available. This has been particularly important in promoting multi- and interdisciplinary training.
Because the NRSA program’s leadership role in research training is so vital to the health of the biomedical, behavioral, and clinical research establishments, a regular assessment is important. Moreover, in formulating this report the committee tried to weigh all of the above factors not only to assess the situation but also to arrive at a set of recommendations for optimizing the NRSA program in the future. In that way the committee anticipates that the NRSA program will continue to be the bellwether for improved health care through research.
Following this introductory and background discussion, Chapters 2, 3, and 4 focus on research workforce considerations in the biomedical sciences, behavioral and social sciences, and clinical sciences, respectively. The next three chapters examine in succession the training issues in oral health, nursing, and health services research. In Chapter 8 the committee addresses aspects of training for emerging fields and interdisciplinary research. Chapter 9 discusses various stages in the career progression of biomedical, behavioral, and clinical researchers, as well as the impact—or lack of it—of foreign-trained scientists and underrepresented minorities. Chapter 10 offers some final overarching comments and recommendations.