Biomedical Research Wgreport
Biomedical Research Wgreport
Biomedical Research Wgreport
1
United States Government Printing Office, Washington: 1945 - http://www.nsf.gov/od/lpa/nsf50/vbush1945.htm
Final Report 2
Biomedical Workforce Working Group Report
Final Report 3
Biomedical Workforce Working Group Report
ACKNOWLEDGEMENTS
The working group benefited greatly from the input and effort of the modeling subcommittee whose
members generously contributed their effort and wide-ranging expertise on subjects as diverse as NIH
policies, modeling, the science and engineering workforce, and the scientific enterprise. We are grateful
for their tireless input on both the overall landscape and specific aspects of it as well policy responses
that ensured that the subcommittee reports reflected a wide range of perspectives and skills.
The working group also would like to thank:
• Dorit Zuk and Wally Schaffer from NIH for their herculean efforts in managing this entire
process, Dorit for her masterful organizational and writing skills as the principal author of this
report and Wally for his remarkable historical perspective and insights on how we got here.
• Cary Scheiderer, Michelle Dunn, Tiffani Lash, and Erica Rosemond from NIH, who spent
countless hours taking notes, gathering data from multiple sources, building the data
infrastructure and the website, and reconciling the numbers.
• The large group of NIH staff who offered their time and expertise to provide input to the
working group as part of the NIH/TAC NTW committee, and particularly Rod Ulane and Richard
Baird for chairing the committee and providing input to the modeling subcommittee and the
working group.
• Michael Gottesman from the NIH Office of Intramural Research for sharing his expertise with
the working group.
• All those who provided information, data and their time during the data-gathering and analysis
phase of this study. Particularly, Cathee Johnson Phillips (National Postdoctorate Association);
Herschel Alexander (American Association of Medical Colleges, AAMC); Howard Garrison
(Federation of American Societies for Experimental Biology); Mike Finn (Oak Ridge Institute for
Science and Education); Jennifer Sutton (Office of Extramural Research, NIH); Jodi Lubetsky
(AAMC) and members of the AAMC GREAT steering committee; Sheri Sharareh Craig, and
members of the NIGMS SWAM initiative, especially Irene Eckstrand and Shiva Singh.
• Laurel Haak from Discovery Logic/Thomson Reuters, Pawel Sulima from Discovery
Logic/Thomson Reuters, and Patricia Oslund from the Center for Science, Technology &
Economic Policy, University of Kansas for gathering and analyzing data.
• Lars Lefgren (Brigham Young University), Sarah Turner (University of Virginia), William Johnson
(University of Virginia), William Kerr (Harvard Business School) who provided input to the
modeling subcommittee on aspects of the labor market.
Final Report 4
Biomedical Workforce Working Group Report
Table of Contents
WORKING GROUP MEMBERS ............................................................................................................................. 3
ACKNOWLEDGEMENTS ...................................................................................................................................... 4
EXECUTIVE SUMMARY ....................................................................................................................................... 7
INTRODUCTION .............................................................................................................................................. 13
PATH TO A CAREER IN BIOMEDICAL SCIENCES .................................................................................................... 17
Considerations about the Data ............................................................................................................... 17
Graduate Students .................................................................................................................................. 17
Postdoctoral Researchers ....................................................................................................................... 19
Career Outcomes ................................................................................................................................... 23
Compensation of Biomedical PhDs ...................................................................................................... 28
Trends in Age Distribution of Biomedical Researchers ........................................................................... 29
WORKFORCE FRAMEWORKS ............................................................................................................................ 31
RECOMMENDATIONS ...................................................................................................................................... 35
Graduate Students .................................................................................................................................. 35
Postdoctoral Researchers ....................................................................................................................... 36
Staff Scientists ......................................................................................................................................... 38
Salary Support ......................................................................................................................................... 39
Physician Scientists ................................................................................................................................. 40
Information Collection, Analysis and Dissemination .............................................................................. 42
Diversity .................................................................................................................................................. 44
ACTIVITIES OF WORKING GROUP AND SUB-COMMITTEES ................................................................................... 45
Modeling Subcommittee ........................................................................................................................ 45
Stakeholder Input ................................................................................................................................... 47
Soliciting Public Comment ...................................................................................................................... 48
NIH/TAC Workforce (NTW) Committee .................................................................................................. 52
GLOSSARY ..................................................................................................................................................... 56
APPENDICES .................................................................................................................................................. 57
A: Working Group Charge and Modeling Subcommittee Roster ........................................................... 57
B: Additional Workforce Data ................................................................................................................. 59
C: Recommendations to Address Specific Data Gaps ............................................................................. 69
D: Incentives Memo ................................................................................................................................ 72
E: Memo from Modeling Subcommittee ................................................................................................ 81
Final Report 5
Biomedical Workforce Working Group Report
Final Report 6
Biomedical Workforce Working Group Report
EXECUTIVE SUMMARY
A working group of the NIH Advisory Committee to the Director (ACD –charter and roster in Appendix A)
was tasked with developing a model for a sustainable and diverse U.S. biomedical research workforce
that can inform decisions about training the optimal number of people for the appropriate types of
positions that will advance science and promote health. Based on this analysis and recognizing that
there are limits to NIH’s ability to control aspects of the training pipeline, the working group was asked
to make recommendations for actions that NIH should take to support a future sustainable biomedical
research infrastructure.
The working group met 11 times in 2011 and 2012, including 4 in-person meetings and 7
teleconferences, with a goal to provide recommendations to the ACD in June 2012. In addition, a
subcommittee consisting of social scientists (primarily economists) with expertise in the scientific
enterprise and NIH-funded investigators with expertise in mathematical models was formed to gather
and analyze data on the biomedical research workforce and develop a model (see roster in Appendix A).
This report summarizes the workforce data collected and the working group’s recommendations. The
working group did not have either the time or the expertise to propose details on how the
recommendations should be implemented. This will require thoughtful consideration by NIH. The
working group recommends that changes to existing programs be phased in gradually and pilot
programs be conducted to test new ideas. The outcomes of all changes should be evaluated rigorously.
Additional workforce data can be found at http://report.nih.gov/investigators_and_trainees/ACD_BWF.
The overall purpose of the recommendations is to ensure future US competitiveness and innovation in
biomedical research by creating pathways through undergraduate, graduate and postdoctoral training
that provide excellent preparation in a timely fashion to:
• Attract and retain the best and most diverse scientists, engineers and physicians from around the
world to conduct biomedical research as well as increase the number of domestic students from
diverse backgrounds who excel in science and become a part of the Science Technology
Engineering and Mathematics (STEM) workforce
• Prepare biomedical PhD students and postdoctoral researchers to participate in a broad-based
and evolving economy
The working group appreciates that K-12 and undergraduate education are major factors that influence
the success of building of the biomedical research workforce but has confined its recommendations to
graduate training and beyond as NIH funding and training focuses on those stages.
Graduate Students
The working group recognizes that the overall number of PhD students in biomedical research is in large
part determined by the budget of the NIH. The vast majority of graduate students in the US are
supported on a combination of NIH training grants, fellowships and research project grants. The
number of fellowships and traineeships has remained relatively constant over time, but the number of
students supported on research grants has grown substantially without any mechanism in place to
review the quality of training that students are receiving. Although the vast majority of people holding
biomedical PhDs are employed (i.e. unemployment is very low), the proportion of PhDs that move into
tenured or tenure-track faculty positions has declined from ~34 percent in 1993 to ~26 percent today.
In contrast the proportion of non-tenured faculty has stayed relatively constant during the same period,
while increasing in absolute numbers. The percentages of biomedical Ph.D.s in industry and
government have remained relatively constant. The categories that have seen growth are science-
Final Report 7
Biomedical Workforce Working Group Report
related occupations that do not involve the conduct of research and occupations that do not require
graduate training in science.
Despite these changes, graduate training continues to be aimed almost exclusively at preparing people
for academic research positions. Therefore, the working group believes that graduate programs must
accommodate a greater range of anticipated careers for students. Graduate programs should reflect
that range, and offer opportunities for students to explore a variety of options while in graduate school
without adding to the length of training. Graduate programs also should openly communicate the
career outcomes of their graduates to potential students.
Finally, the working group recognizes that there are aspects of the biomedical workforce that make it
less attractive to potential graduate students. The overall length of training in the biomedical sciences
(PhD plus postdoctoral research) is longer than in comparable scientific disciplines such as chemistry,
physics and mathematics. For PhDs graduating in 2001, the median age for biomedical scientists was 32
and the median age for starting a tenure track position was 37; comparable ages for chemistry
doctorates were 30 and 33. Furthermore, academic salaries at public research institutions for assistant
professors in biomedical fields are low compared to other fields. According to the Oklahoma State
University survey of public research institutions; average starting salaries in fiscal year 2011 for
biomedical assistant professors were approximately $68,000 compared to $69,000 for chemistry,
$79,000 for clinical and health fields and over $100,000 for economists. The long training period,
together with disparities in earnings, may make a career in biomedical research less attractive than one
in other scientific disciplines and professional careers.
Recommendations:
• NIH should create a program to supplement training grants through competitive review to allow
institutions to provide additional training and career development experiences to equip students
for various career options, and test ways to shorten the PhD training period. The best practices
resulting from this program will help shape graduate programs across the country. The working
group felt that including diverse types of training (e.g. project management and business
entrepreneurship skills needed in the pharmaceutical and biotechnology industries, or teaching
experiences needed for a successful faculty position in liberal arts colleges) would be particularly
valuable for those who go on to conduct NIH-funded research as well as benefit those students
who do not follow the academic research career track. For example:
o Approximately 30% of biomedical PhDs work in the biotech and pharmaceutical industries in
research and non-research positions. Their transition would be more effective if their
training was better aligned with the required skill-sets for these careers. NIH and the
institutions should explore ways to involve relevant employers in the public and private
sector in designing training paths for those students who seek employment in that sector. It
is possible that the pharmaceutical and biotechnology sectors would be willing to partner in
supporting such programs. Another option would be for institutions to develop pilot
programs in partnership with private foundations and industry to prepare Ph.D. graduates
for careers that involve translational research and development. Finally, NIH should
encourage the SBIR/STTR awardees to provide internships for graduate students and
postdoctoral researchers to enable increased hands-on training at small businesses.
o Institutions also could be encouraged to develop other degree programs, e.g. master’s
degrees designed for specific science-oriented career outcomes, such as industry or public
policy. These could be developed as stand-alone programs or provide sound exit pathways
for PhD students who do not wish to continue on the research career track. However, this
Final Report 8
Biomedical Workforce Working Group Report
would require a change in the definition of “success” in the evaluation of NIH training
grants.
• To encourage timely completion of graduate degrees, NIH should cap the number of years a
graduate student can be supported by NIH funds (any combination of training grants, fellowships,
and research project grants), with an institutional average of 5 years and no one individual
allowed to receive support for more than 6 years. Note that a different cap may be needed for
physician scientists (MD, DDS, MD-PhD etc.). NIH should continue to assess the pre-doctoral
stipend level annually.
• To ensure that all graduate students supported by the NIH receive excellent training, NIH should
increase the proportion of graduate students supported by training grants and fellowships
compared to those supported by research project grants, without increasing the overall number of
graduate student positions.
• NIH should revise the peer review criteria for training grants to include consideration of outcomes
of all students in the relevant PhD programs at those institutions, not only those supported by the
training grant. Study sections reviewing graduate training programs should be educated to value a
range of career outcomes. This recommendation could be phased in relatively quickly.
• The very different requirements and characteristics of training programs at each NIH Institutes
and Center (IC) constitute a substantial burden on the institutions. All NIH ICs should offer
comparable training programs and fellowships and their requirements should be harmonized.
Postdoctoral Researchers
As the number of graduate students doubled over the past twenty years, it is not surprising that there
was a comparable increase in US-trained postdoctoral fellows, along with a significant influx of foreign-
trained fellows. There are very little reliable data on the number of postdoctoral researchers in the US
and the length of their training (see below for specific recommendations to address the lack of data).
This is due to a dearth of information about the numbers of foreign-trained postdoctoral researchers, as
well as changes in the titles of postdoctoral researchers as they proceed through their postdoctoral
positions. The lack of reliable estimates of the population size and rates at which people enter and
leave the postdoctoral pool complicated the analysis.
Nonetheless, after analyzing the available data, the working group believes that the postdoctoral
experience be considered an extension of the training period primarily intended for those Ph.D.
graduates who intend to pursue research-intensive careers. Fellows should be given structured career
development opportunities and there should be incentives provided by NIH to move postdoctoral
fellows to more permanent positions as soon as possible. The working group also recognizes that
postdoctoral fellows have spent years in graduate training, and should be compensated accordingly.
Recommendations:
• To ensure that all postdoctoral fellows supported by the NIH receive excellent training and
mentoring, NIH should increase the proportion of postdoctoral researchers supported by training
grants and fellowships and reduce the number supported by research project grants, without
increasing the overall number of postdoctoral researchers.
• NIH should create a pilot program for institutional postdoctoral offices to compete for funding to
experiment in enriching and diversifying postdoctoral training, including partnerships with other
entities (industry, private foundations, government, etc.).
Final Report 9
Biomedical Workforce Working Group Report
• The current stipends for NIH-supported postdoctoral fellows need to be adjusted to levels that
better reflect their years of training. The working group recommends that the NIH should adjust
the starting stipend levels of the Ruth L. Kirschstein National Research Service Awards (NRSA) to
$42,000 and index the starting stipend according to the Consumer Price Index (CPI-U) thereafter.
Stipend levels should increase with each year of experience in any postdoctoral position
irrespective of their titles by 4% for the second and third years and 6% for years 4 through 7. The
large jump between years 3 and 4 is meant to emphasize a transition from postdoctoral training
to research production, and to incentivize PIs to move fellows to more permanent positions. This
salary scale will apply to postdoctoral researchers supported by research project grants as well,
and institutions should be encouraged to adopt this scale for all postdoctoral researchers,
irrespective of the source of their support.
NIH should evaluate this policy in the decade after implementation to determine whether the
postdoctoral period has shortened. If it is not reduced, then perhaps NIH should experiment with
a cap on the length of funding for postdoctoral researchers.
• NIH should require and adjust its own policies so that all NIH-supported postdoctoral researchers
on any form of support (training grants, fellowships or research project grants) receive benefits
that are comparable to other employees at the institution. Such benefits include paid time off,
health insurance, retirement plans, maternity leave etc.
• To encourage larger numbers of PhD graduates to move rapidly into permanent research
positions, NIH should double the number of Pathway to Independence (K99/R00) awards, and
shorten the eligibility period for applying to this program from the 5 years to 3 years of
postdoctoral experience.
• NIH also should double the number of the NIH Director’s Early Independence awards to facilitate
the “skip-the-postdoc” career path for those who are ready immediately after graduate school.
• NIH should require individual development plans (IDPs) for all NIH-supported postdoctoral
researchers, whether on training grants, fellowships, or research project grants. Assessment of
implementation of this requirement should be included in the review criteria of training grants.
Staff Scientists
The typical academic laboratory consists of a PI and one or a small number of permanent technical staff,
with the majority of the research carried out by trainees. This creates a system in which a large number
of future scientists are being produced each year, well in excess of the number of research-oriented jobs
in academia, government and industry. The working group believes that even a modest change in the
ratio of permanent staff to trainees could have a beneficial effect on the system without reducing the
productivity of the research enterprise. Staff scientists - individuals with masters or PhD degrees - could
play a more important role in biomedical research (one that may become increasingly necessary if the
market for biomedical researchers strengthens outside of the United States in coming years).
Today, these scientists bring stability to many labs and provide important functions as part of
institutional core facilities, but have a wide variety of titles and employment conditions. As an example,
staff scientists constitute an essential part of the NIH intramural research program. In the extramural
program, these scientists do not apply for their own grants, but are supported by research project,
Center and Program Project grants. They should be differentiated from “soft money” scientists, whose
employment depends upon their successful competition for research funds, a category that has been
increasing over the last few years.
Final Report 10
Biomedical Workforce Working Group Report
The working group encourages NIH study sections to be receptive to grant applications that include
staff scientists and urges institutions to create position categories that reflect the value and stature of
these researchers.
Salary Support
Originally the conduct of federally-funded research at universities and other extramural institutions was
based on an understanding that institutions would provide the bulk of facilities and salaries to the
researchers and the NIH would provide the majority of funds for conducting research. Over the past
decades, this distinction has become increasingly blurred, with NIH providing an increasing proportion of
faculty salary support and the institutions covering a larger percentage of the research costs. This is
especially true during the start-up period, which has become significantly longer as young investigators
struggle to receive their first R01 grants. The growth in “soft money” positions in academic medical
schools, in which investigators are required to raise 100% of their salaries and research funds, has
contributed to the negative views of a career in biomedical science, and has had the additional
consequence of encouraging institutions to expand their physical space without making additional long
term commitments to faculty.
The working group believes that institutions should provide some fraction of salary support for their
researchers in order to qualify for NIH funding. That being said, the working group appreciates that any
reduction in NIH salary may have major consequences on institutions.
The working group recommends that NIH consider a long-term approach (over a 20 year period) to
gradually reduce the percentage of funds from all NIH sources that can be used for faculty salary
support.
Physician Scientists
The working group was charged with addressing physician scientist training as well as PhD training. The
economic and educational drivers which affect the training and career paths of the physician scientist
workforce are very different from those underlying PhD research training and career paths and there
was not sufficient time for the working group to examine this important part of the biomedical
workforce in detail. In addition, the changing landscape of health care and the effects these changes
likely will have on academic medical centers need to be projected carefully and considered when
analyzing the future physician scientist workforce.
Therefore, the working group recommends that NIH conduct a follow-on study that focuses on
physician scientists and involves people who train physician scientists, as well as economists who
focus on medical education costs, career choices, and the role of these as incentives.
Information Collection, Analysis and Dissemination
The working group was frustrated and sometimes stymied throughout its study by the lack of
comprehensive data regarding biomedical researchers. The timeframe and resources of the study did
not allow for comprehensive data collection or the implementation of a comprehensive model of the
biomedical workforce. It is evident from the data-gathering and analyses undertaken by the working
group that there are major gaps in the data currently being collected on foreign-trained postdoctoral
researchers and those who work in industry.
The working group also believes that it is imperative to provide as much information as possible to
prospective graduate students and postdoctoral researchers on career outcomes both nationally and at
their specific training programs so they can make more informed decisions about their future.
Final Report 11
Biomedical Workforce Working Group Report
Recommendations:
• Institutions that receive NIH funding should collect information on the career outcomes of both
their graduate students and postdoctoral researchers, and provide this information to prospective
students/ postdoctoral researchers and the NIH. Such information should include completion
rates, time to degree, career outcomes for PhD trainees, as well as time in training and career
outcomes from postdoctoral researchers over a 15-year period. Outcome data should be
displayed prominently on the institution’s web site. This will require institutions to track the
career paths of their students and postdoctoral researchers over the long-term. One way to do
that would be to assign graduate students and incoming postdoctoral researchers an identifier
that can be used to track them throughout their careers.
• NIH, working with other agencies in the Federal Government, should address the identified data
gaps and collect information on the biomedical and scientific workforce on an ongoing basis.
• NIH should create a permanent unit in the Office of the Director that works with the extramural
research community, the National Science Foundation (NSF) and the NIH ICs to coordinate data
collection activities and provide ongoing analysis of the workforce and evaluation of NIH policies
so that they better align with the workforce needs.
Diversity
Increasing diversity of trainees and the workforce is critical to the future of biomedical research in the
US, particularly as the share of the US population comprised of underrepresented groups increases. The
committee recognizes that this is the responsibility of the entire scientific community but feels NIH
should set an example.
Although the working group recommendations are not aimed specifically at increasing diversity, the
group feels that implementation of these recommendations will increase the overall attractiveness of
the biomedical research career and consequently its attractiveness to underrepresented ethnic and
racial minorities and women.
The working group is aware that another working group of the Advisory Committee to the NIH
Director is focused on this issue but would like to highlight the need for much stronger coordination of
the many diversity-related efforts at the NIH and for rigorous evaluation of the outcomes of all
programs.
Conclusion
The working group is aware that similar recommendations have been made in the past by other groups
that studied the biomedical research workforce. Many of those recommendations were not
implemented, in part because of funding constraints and in part because of resistance from the scientific
community. Therefore, the working group urges NIH to provide the funds necessary to implement
these recommendations and encourages institutions to work with NIH on the implementation.
Final Report 12
Biomedical Workforce Working Group Report
INTRODUCTION
Over the years, biomedical research, funded in large part by the National Institutes of Health (NIH), has
contributed enormously to an increase in health and life expectancy in the US. As described in the 2007
NIH biennial report to Congress 2, life expectancy increased by 7.4 years from 1961 to 2004. Infant
mortality has decreased from 26 deaths per 1,000 live births in 1960 to 6.9 in 2005. Biomedical research
has and continues to expand our understanding of the physiology underlying many diseases (often at
the molecular level), contributing, along with other factors such as changes in behavior, to numerous
advances in treatments and improved health care. The change in the prognosis for HIV patients is one
example of these benefits. In the 1990s, the discovery and development of antiretroviral drugs
transformed HIV infection for many individuals from a death sentence into a chronic disease. In
addition, the results of biomedical research have led to important changes in the US economy, launching
the biotechnology industry and changing the way pharmaceutical companies develop new drugs and
treatments.
Successful biomedical research relies on the talent and dedication of the scientific workforce and a
continued supply of highly trained people of the best quality who can bring new insights to our
understanding of biology and advance the translation of these insights into improved health for all. To
this end, NIH supports training of graduate students and postdoctoral researchers both with dedicated
training grants and fellowships and as employees on research project grants.
Training at NIH
The training of biomedical researchers has been an integral part of the NIH mission since its earliest
days. In 1930 the Ransdell Act 3 established the National Institute of Health. By the early 1970s, the
NIH included multiple institutes and the training programs had grown substantially; nearly 15 percent of
NIH extramural funding was dedicated to research training. The National Research Act of 1974
amended the Public Health Service act by repealing existing research training and fellowship authorities
and consolidating these authorities in the National Research Service Awards (NRSA) authority.
In 2002, Congress renamed the National Research Service Award program after Ruth L. Kirschstein in
recognition of her many scientific accomplishments in polio vaccine development, and her tenure as the
first woman director of an NIH Institute. Dr. Kirchstein was a champion of research training and a strong
advocate for the inclusion of underrepresented individuals in the scientific workforce 4.
Today, NIH has authority to award NRSA individual fellowships to support predoctoral and postdoctoral
training of individuals to undertake biomedical, behavioral, or clinical research at domestic and foreign,
public and private institutions. The NRSA legislation authorizes NRSA institutional research training
grants and limits institutional NRSA support to training and research at domestic public and non-profit
private entities. Individuals trained in these programs must be citizens (or noncitizen nationals) of the
United States or have been lawfully admitted for permanent residence by the time of the award.
Individuals receiving postdoctoral support under individual fellowships or institutional research training
grants are required to pay back to the Federal government their initial 12 months of Kirschstein-NRSA
postdoctoral support by engaging in health-related biomedical, behavioral and/or clinical research,
research training, health-related teaching, or any combination of these activities. Arguably the most
important feature of the service payback obligation is the requirement to monitor the payback
2
http://report.nih.gov/biennialreport/
3
P.L. 71-251, 46 Stat. L. 379
4
http://www.nigms.nih.gov/Training/RuthKirschstein
Final Report 13
Biomedical Workforce Working Group Report
obligations, which necessitates careful data collection and tracking of NRSA recipients. This data
collection has allowed for comprehensive evaluation of the programs.
In FY2012, these research training programs comprise 3% of the NIH budget 5. The number of NRSA
training positions awarded has not changed substantially in the past decade 6. For every graduate
student and postdoctoral researcher supported by NRSA NIH research training programs, however,
there are between 2 and 4 individuals who are supported as research assistants and associates working
on NIH research project grants.
Assessment of Biomedical Research Workforce Training
Together with the NRSA act, Congress created a companion act that requires regular assessment of the
needs for research personnel, the fields of training, and the kinds and intent of such training. That
assessment is carried out by the National Research Council (NRC). Initially those studies were required
every year and then every four years.
The last such study was completed in 2011 7. This study, chaired by Roger Chalkley of Vanderbilt School
of Medicine, found that, based on the observation of low unemployment rates of biomedical and
behavioral scientists and models that predicted substantial growth in scientific employment
opportunities over the next decade, the number of NRSA positions is adequate and should remain at the
same level in biomedicine and should be increased in behavioral sciences.
As described later in this report, the data gathered by the ACD working group do not indicate such
growth in employment opportunities. Rather, the numbers of positions available for biomedical PhDs
that take advantage of their long training are less than the number of PhDs produced each year. As a
consequence their career path is marked by uncertainty. Compensation is relatively low compared to
other disciplines such as engineering and the physical sciences, and the NIH funding environment is
highly uncertain for the near future.
The NRC report also recommended increases in the number of Medical Scientist Training Program
(MSTP) students, increases in graduate and postdoctoral stipends, increases in the indirect cost rate on
training grants and career development awards, and increases in efforts to enhance the diversity of the
graduate and postdoctoral training programs. Finally, the report suggested improvements in the way
workforce data are collected and managed, recommended changes in the content of training grant
applications, and made a number of additional discipline and training content focused
recommendations.
Other studies of the NRSA program have been conducted over the years. In 2001, NIH published an
evaluation of The Early Career Progress of NRSA Predoctoral Trainees and Fellows, conducted by
Georgine Pion of NIH and Vanderbilt University 8. The study compared career outcomes of NRSA award
recipients who completed their doctorate between 1981 and 1992 to students who did not receive
NRSA predoctoral support (either in departments that had NRSA predoctoral training grants or in those
that did not have such grants). The outcomes measured included educational attainment, postdoctoral
training, research-related employment, success in applying for NIH and NSF research support, and
research productivity as defined by publication and citation rates
5
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=5&catId=1
6
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=52&catId=17
7
http://grants.nih.gov/training/Research_Training_Biomedical.pdf
8
http://grants.nih.gov/training/career_progress/index.htm
Final Report 14
Biomedical Workforce Working Group Report
The study found that predoctoral NRSA recipients completed their degrees in less time and were more
likely to engage in postdoctoral research training, assume faculty positions, apply for and receive NIH
and NSF grants, and publish highly cited papers than individuals who graduated at the same time in the
same field without the benefit of NRSA support.
In 2006, NIH conducted a study of The Career Achievements of NRSA Postdoctoral Trainees and Fellows:
1975–2004 9. The study evaluated career outcomes of postdoctoral researchers who received support
from fiscal year 1975 through fiscal year 1992, comparing NRSA recipients to postdoctoral fellows
supported by other means. Postdoctoral researchers on training grants were considered separately
from those with fellowships. The outcomes measured were:
• success in obtaining NIH research grant support
• success in publishing in peer-reviewed journals
• success in obtaining and remaining in research-oriented employment
The study found that postdoctoral NRSA fellows performed better in all outcomes measured than
comparison postdocs including those that were supported by NRSA training grants.
In addition to the studies that evaluated the NRSA programs specifically, analyses of the broader
biomedical research workforce and training needs have been conducted over the years. One example is
a study published by the National Research Council in 1998, Trends in the Early Careers of Life Scientists,
chaired by Shirley Tilghman 10. The committee examined the graduate and postgraduate training of life
scientists and the nature of their employment on completion of training.
The study concluded that the level of PhD production in 1998 exceeded the availability of jobs in
academe, government and industry where they can use their training as independent scientists. As a
result, increasing numbers of PhDs occupy postdoctoral and other academic appointments outside the
tenure and tenure track. The structure of the life sciences was built on the premise that the enterprise
would continuously expand and absorb and employ the large number of graduate students and
postdoctoral researchers. In the absence of such expansion there is a growing imbalance between the
rate of training and the growth in research career opportunities. The 1998 committee suggested that
the absence of suitable employment has led to a crisis of expectations that could discourage the best
students from entering the field.
The 1998 committee recommended restraint in future growth in the number of graduate students,
disseminating accurate information about career prospects, improvement in the educational and
training experience of graduate students, funding mechanisms that shorten the postdoctoral period,
and, focusing on preparing students for independent research positions rather than for “alternative”
careers. It is notable that this report was released just before the doubling of the NIH budget, which
may have affected the perception of the urgency of its recommendations.
Recognizing that the behavioral and biomedical research enterprise has grown in both size and
complexity in the past decade - particularly with the doubling of the NIH budget between 1999 and
2003, and that the NIH budget is likely to remain flat or even decline in the near future, the NIH Director
tasked the ACD in December 2010 with forming a workgroup that would develop a better understanding
of current and future needs of the behavioral and biomedical research workforce in various sectors.
These sectors include academia, industry, and government, including those who do research and those
9
http://grants.nih.gov/training/NRSA_report_5_16_06-2.doc
10
Trends in the Early Careers of Life Scientists, Commission on Life Sciences, National Research Council, National
Academies of Science, National Academy Press, Washington DC 1998.
http://www.nap.edu/catalog.php?record_id=6244
Final Report 15
Biomedical Workforce Working Group Report
who use their training in other ways. The working group would collect data on the complete biomedical
research workforce to support a more comprehensive assessment of the workforce are needed to fill
biomedicine-related positions now and in the future (see charter and roster in Appendix A).
The working group appreciates that K-12 and undergraduate education are major factors that influence
the success of building of the biomedical research workforce but has confined its recommendations to
graduate training and beyond as NIH funding and training focuses on those stages.
The working group met a total of eleven times in 2011 and 2012, including four in-person meetings and
seven teleconferences, with a goal of providing recommendations to the ACD in June 2012. In addition,
a subcommittee consisting of social scientists (primarily economists) with expertise in the scientific
enterprise as well as NIH-funded investigators with expertise in mathematical models was formed to
gather and analyze data on the biomedical research workforce and develop a model (see roster in
Appendix A). The subcommittee met three times in 2011 and 2012, including two in-person meetings
and one teleconference.
This report summarizes the workforce data collected and the working group’s recommendations.
Additional data can be found at http://report.nih.gov/investigators_and_trainees/ACD_BWF.
Final Report 16
Biomedical Workforce Working Group Report
Graduate Students
The number of PhDs trained in biomedical science in the US has risen steadily over the past decade as
evident from the data below. In contrast, the number of PhDs trained in Behavioral and Social Sciences
and Chemistry has been stable over the same period (see Figure 1). The steep increase in the number of
biomedical PhDs awarded began in 2004, just after the end of the doubling of the NIH budget (1999-
2003). Given a 5-7 year training period, this illustrates a close relationship between the size of the NIH
budget and the number of biomedical PhD slots.
11
http://www.nsf.gov/statistics/srvydoctoratework/
Final Report 17
Biomedical Workforce Working Group Report
12
Figure 1: US Graduate Degrees Awarded, by Field
NIH supports the vast majority of biomedical graduate students in the US on a combination of training
grants, fellowships and research project grants. Many more students are supported on research grants
than by training grants and fellowships. The number of students on research grants has grown
considerably over the past decade, in parallel with the doubling of the NIH budget (Figure 2). On the
other hand, the number of students supported by traineeships and fellowships has increased only
modestly over the same period. Training grants uniquely provide the NIH with a mechanism for peer
review of training, and permit the NIH to require attention to issues such as outcomes, diversity and
professional ethics training. On the other hand, there are no training-related requirements for students
on research grants. The existence of an NIH training program at an institution, however, can motivate
graduate programs to provide all students at that institution with training that conforms to NIH
guidelines and expectations.
13
Figure 2: Doctorate Students by Type of Support
Along with the increase in the number of biomedical PhD students trained in the US, there is a
perception in the community that the length of PhD training also has increased. As shown in Figure 3,
12
SED
13
GSS
Final Report 18
Biomedical Workforce Working Group Report
the time to degree and age at degree of biomedical PhDs actually have remained stable or decreased in
the past 15 years.
14
Figure 3: Time to Degree and Age at Degree
Nevertheless, the overall length of training in the biomedical sciences (PhD plus postdoctoral research)
is longer than in comparable scientific disciplines such as chemistry, physics and mathematics,
particularly for those scientists who go on to tenure-track research positions. For PhDs graduating in
2001, the median age for biomedical scientists was 32 and the median age for starting a tenure track
position was 37; comparable ages for chemistry doctorates were 30 and 33. This difference can be seen
in the SED data presented below in Figure 7. In addition, the increasing age at which medical school
faculty obtain their first tenure-track position has increased, as shown in Figure 17.
Postdoctoral Researchers
As the number of graduate students doubled over the past twenty years, it is not surprising that there
was a comparable increase in postdoctoral fellows. This increase was augmented by a significant influx
of foreign-trained fellows. As the working group began its work, it quickly became clear that there are
very little reliable data on the number of postdoctoral researchers in the US and how this number has
changed over the years (see below for specific recommendations to address the lack of data). This is
due to a dearth of information about the numbers of foreign-trained postdoctoral researchers, as well
as changes in the titles of postdoctoral researchers as they proceed through their training. The National
Postdoctoral Association defines a postdoctoral scholar as “an individual holding a doctoral degree who
is engaged in a temporary period of mentored research and/or scholarly training for the purpose of
acquiring the professional skills needed to pursue a career path of his or her choosing. 15” We have used
the term postdoctoral researcher throughout this report. The lack of reliable estimates of the
population size, and the rates at which people enter and leave the postdoctoral pool, greatly
complicated the analysis conducted by the working group.
Nevertheless, the available data do provide some insights into the state of US-trained postdoctoral
researchers in the biomedical sciences. Like graduate students, postdoctoral researchers primarily are
14
SED
15
http://www.nationalpostdoc.org/policy/what-is-a-postdoc
Final Report 19
Biomedical Workforce Working Group Report
supported by the NIH with a combination of training grants, fellowships, and research grants. The vast
majority of US-trained postdoctoral researchers are supported on research grants and that number has
increased steadily for a long time (Figure 4). Note that data from other sources including NIH suggest
that the number of postdoctoral researchers in the figure below may be under-estimated by as much as
a factor of 2, due in part to the nomenclature problem and to the fact that the GSS (from which the data
are derived) only includes postdoctoral researchers who are at degree-granting institutions.
16
Figure 4: Biomedical Postdoctoral Researchers by Type of Support
The other source of support that has been growing over the last five years is “nonfederal support”,
defined as support from state and local government; the institution, such as stipends; foreign sources,
such as foreign governments, foreign firms, and agencies of the United Nations; and other US sources,
such nonprofit institutions, and private industry.
A large number of postdoctoral researchers are foreign-trained. The available data suggest that their
number has grown immensely over the past two decades (Figure 5). Non-US citizens (the majority of
foreign-trained postdoctoral researchers) are not eligible for most NIH training grants or fellowships.
The majority of these researchers are supported by research project grants.
16
NSF Graduate Students and Postdoctorates Survey
Final Report 20
Biomedical Workforce Working Group Report
17
Figure 5: Biomedical Postdoctoral Researchers by Citizenship
There is a perception in the biomedical community that the postdoctoral training period has lengthened
over time. With the caveats in accessing accurate data listed above, data from the SDR suggest that
most US-trained biomedical PhDs spend fewer than 5 years in postdoctoral positions, although that
number has been steadily growing with time (Figure 6). Furthermore, there are a significant number
who remain in postdoctoral training between 5-8 years. There is some indication that the researchers
remaining in the postdoctoral position the longest are the ones who go on to tenure-track academic
research careers. For example, in Figure 7, it is evident that the age at first non-postdoctoral job (many
of which are in industry) has been consistently a year or two lower than the age of first tenure-track job.
Note that the latest data in this graph (2002-2003) may be underreported due to a lag-time bias.
18
Figure 6: Years Biomedical PhDs Spend in Postdoctorate Training Positions
17
GSS
18
SDR
Final Report 21
Biomedical Workforce Working Group Report
19
Figure 7: Age at First PhD, First Non Postdoctoral Job, First Tenure Track Job, for US trained Doctorates
In addition, the scientific fields most likely to have postdoctoral researchers coincide with the top fields
funded by NIH 20 (see box).
The current stipend levels of postdoctoral NRSA awardees are listed in Table 1 below. When compared
to stipends of postdoctoral fellows funded by other agencies such as NSF ($45,000 21) and the
Department of Energy ($65,000 22), the NIH stipends for biomedical postdoctoral researchers are low.
19
SED
20
SDR
21
http://www.nsf.gov/pubs/2011/nsf11499/nsf11499.pdf page 6
22
http://www1.eere.energy.gov/education/postdoctoral/benefits.html
Final Report 22
Biomedical Workforce Working Group Report
Career Outcomes
As mentioned above, data from the SDR, which is designed specifically to track the PhD labor force
and is thus the major source for data on US-trained biomedical PhDs, extends only through 2008.
Therefore, much of the information presented below about career outcomes does not take into
account the past four years (including the recent recession). The working group gathered more
up-to-date data from other sources and those are included where possible.
Even as the number of US-trained doctoral recipients in the fields analyzed generally increased from
1993 to 2008, the number of PhD recipients that are employed has declined slowly although
unemployment has remained remarkably flat (Figure 8).
24
Figure 8: Employment Status of US Doctorates
As can be seen in Figure 9, across science and engineering PhD fields 60-80% of graduates report that
they are employed in occupations that are closely related to their PhD field. However, the percent in
biomedical sciences decreased between 1997 and 2008 from 70% to just below 59%. Other fields do
not show a decrease of this magnitude. Figure 10 shows that over 70% of biomedical PhDs begin
23
http://grants.nih.gov/grants/guide/notice-files/NOT-OD-12-033.html
24
SDR
Final Report 23
Biomedical Workforce Working Group Report
working in research positions immediately after graduate school. By 11 years after their PhD 60% still
work in a research occupation, but once again, the percentages among relatively recent graduates have
steadily decreased since 1995. Taken together, these data indicate an increasing imbalance in the
supply and demand of individuals in research-related occupations over time. Given the current state of
the US economy, it is reasonable to predict that this imbalance will continue, and possibly grow.
25
Figure 9: Relationship between Science and Engineering PhD Field and Occupation
26
Figure 10: U.S. Trained Biomedical PhDs in Research Occupations, by Years since Degree
Figure 11 shows employment of US-trained biomedical PhDs at various times after obtaining their
degree. Most of the individuals in the academic non-tenure track group 1-5 years since their degree
probably are postdoctoral researchers. The trend data 6-10 years and 11-20 years after the PhD show
that the proportion of academic tenured or tenure-track positions has decreased over the past decade,
while the proportion of non-academic non-research positions have increased over the same period.
25
SDR
26
SDR
Final Report 24
Biomedical Workforce Working Group Report
27
Figure 11: U.S. Trained Biomedical PhD employment, by Years Since Degree
Focusing on PhD employment in the academic sector, it is evident from Figure 12 that the number of
biomedical PhDs employed in this sector increased from 1993 to 2006. However the percentage of
tenure or tenure-track positions decreased steadily during this period, as colleges and universities chose
to increase their staff by increasing the number of non-tenure-track positions. These positions often are
dependent on obtaining outside funding (mainly from NIH) to cover 100% of salary. There seems to be a
decline in all positions between 2006 and 2008 and it remains to be seen whether this will continue
when the 2010 data are released. Academic employment in other life sciences showed a similar trend
but at a much lower level. In contrast, academic employment in chemistry was stable over the same
time period.
28
Figure 12: U.S. Trained PhDs in academic employment, by tenure track status
27
SDR
28
SDR
Final Report 25
Biomedical Workforce Working Group Report
The AAMC faculty roster provides the opportunity to take a closer look at employment in medical
schools as Figure 12 includes all academic institutions. Figure 13 shows faculty appointments between
1980 and 2010. MD tenure-track appointments represent the largest component of medical school
faculty appointments, but as summarized in the conceptual framework (Figure 20), the vast majority of
those are in Clinical Departments where the primary focus is not research. Non-tenure track
appointments increased during the early part of the millennium and stabilized in 2006.
29
Figure 13: Number of Medical School Appointments, by Degree and Tenure Status
The working group collected data on the MD and MD/PhD workforce, mainly from the AMA and AAMC
and incorporated them into a conceptual framework (see Figure 20). However, neither organization
provides much detail on the fraction of the MD workforce that conducts biomedical research, either as
their main occupation or in addition to their clinical duties.
According to the AMA30, 82% of physicians are strictly involved in patient care. This is consistent with
data from the AAMC Medical School Graduate Questionnaire 31, which indicated that 16.7% of
respondents in 2011 expected to be significantly involved in research during their medical career. This
percentage has been relatively stable over the past 5 years.
One measure of the number of physician scientists conducting research is the percentage funded by
NIH. As can be seen in Figure 14, researchers with an MD or MD/PhD comprise around 30% of NIH-
funded Principal Investigators (PIs). Although the combined percentage has remained steady for many
years, the percentage of MDs has declined and the percentage of MD/PhDs has increased slightly in the
past few years. These percentages correspond to approximately 4,700 MDs and approximately 3,000
MD-PhDs in FY 2011.
29
AAMC Faculty Roster
30
Physician Characteristics & Distribution in the US – 2012 Edition
31
https://www.aamc.org/download/263712/data/gq-2011.pdf
Final Report 26
Biomedical Workforce Working Group Report
32
Figure 14: Distribution of NIH RPG Principal Investigators by Degree Type
The working group gathered data from several sources to evaluate the status of biomedical PhD
employment in industry. The SDR only includes data through 2008, while the Bureau of Labor Statistics
Occupational Employment Statistics (OES) program includes data through 2011. The OES gives the
number of people employed in a field by sector of employment but does not clearly indicate the number
of PhDs. The number of PhDs in each occupation was estimated based on the SDR. Unlike the SDR, the
OES sample includes both US- and foreign-trained workers.
To extend employment information beyond 2008 and try and reconcile the various data sources,
Figure 15 compares the trends in employment of various biomedical occupations in the SDR (1993-
2008) and the OES (2002-2011). The numbers in each occupation category are not identical but the
trends are informative and show a general flattening or decrease in job growth in jobs over the
past few years, with the exception of “medical scientists”. These data should be considered in
light of the fact that the number of newly minted US-trained biomedical PhDs has been increasing
over the past decade, particularly after the doubling of the NIH budget and reached 9,000 in 2009.
Although ~70% of these graduates go on to postdoctoral training, this number is an indication of
the magnitude of the number of US-trained PhDs seeking positions. In addition, there are many
non-US citizen PhDs in postdoctoral positions, most of whom are foreign-trained and many of
these also are seeking positions. The SDR shows growth in biomedical employment (especially
medical scientists) through 2008. However, the OES indicates that between 2008- 2011,
biomedical employment has been flat in some fields and declined in others. Overall, based on the
BLS quarterly census of employment and wages, the pharmaceutical and medicine manufacturing
sector lost almost 16,000 jobs between 2008 and 2011 33. This includes all levels (not just PhDs) and all
parts of the sector (not just research and development) but provides an idea about the magnitude of the
job loss.
32
NIH OSAR. RPGs include the following activity codes: R00, R01, R03, R15, R21, R22, R23, R29, R33, R34, R35,
R36, R37, R55, R56, RL1, RL2, RL5, RL9, P01, P42, PN1, UA5, UC1, UC2,UC3, UC4, UC7, UH2, UH3, UM1, U01, U19,
U34, DP1, DP2, DP3, DP4, DP5, RC1, RC2, and RC3
33
http://data.bls.gov/pdq/querytool.jsp?survey=en; searched for Total US, NAICS 3254 Pharmaceutical and
medicine manufacturing
Final Report 27
Biomedical Workforce Working Group Report
34
Figure 15: Estimated Change in Biomedical Employment Levels in Biomedical Occupations
34
SED and OES
35
SDR
Final Report 28
Biomedical Workforce Working Group Report
36
Figure 16: Discounted Earnings Profiles
37
Figure 17: Age Distribution in 1980 (background) and 2010 (foreground)
36
SDR
Final Report 29
Biomedical Workforce Working Group Report
38
Figure 18: Age Distribution in 1980 and 2010
37
NIH and AAMC
38
Source NIH and AAMC
Final Report 30
Biomedical Workforce Working Group Report
WORKFORCE FRAMEWORKS
Based on the data presented above and additional analyses performed by the modeling subcommittee
(see details in Activities sections below), conceptual frameworks were developed to provide static
models of the workforce – one each for the PhD and the MD and MD-PhD workforces. The frameworks
shown below were populated with information on each career stage and transition, summarizing the
current state of the workforce as can be determined from available data. The post-training workforce
boxes are color coded, with light blue denoting those in research positions and academic teaching
positions.
Due to the lack of data on certain sectors of the workforce (see details of these data gaps in Appendix
C), many of the numbers shown in the diagrams below are estimates. The color of the numbers
indicates the level of confidence in the numbers due to the precision of both the underlying data and
the methods by which estimates were derived (with red denoting rough estimates). The data sources
and methods by which the numbers were derived are described in the footnotes and additional
information can be found at http://report.nih.gov/investigators_and_trainees/ACD_BWF.
Final Report 31
Biomedical Workforce Working Group Report
The PhD biomedical research workforce is summarized in Figure 19. It is important to note that the
“Post-Training Workforce” box contains information on US-trained PhDs only as these are captured in
the SDR. There is very little available information on the career outcomes of foreign-trained PhDs (one
of the major data gaps).
The “Science Related Non-Research” box includes individuals employed by industry, government, or
other who do not conduct research. However, this box is colored dark blue to indicate that many of the
careers represented in this box are closely related to the conduct of biomedical research and require
graduate training in biomedical science. Examples of such careers include program and review officers
at NIH and managers in many biotechnology companies. The 18% in this box is comprised of 13% PhDs
employed by industry, 2.5% by government, and 2.5% other. Therefore, all individuals employed by
industry (research plus non-research occupations) comprise ~30% of the workforce and all individuals
employed by government comprise ~9%.
39
Figure 19: Conceptual Framework of the PhD Biomedical Research Workforce
39
Data Sources:
Graduate Education and Training - Green numbers: NSF GSS or SED; Yellow Time to degree range: NIH statistics,
Council of Graduate Schools completion rates, and NSF time to degree. International - Yellow International return:
Oak Ridge Institute for Science and Education reports authored by Mike Finn (http://orise.orau.gov/science-
education/publications/default.aspx); Red International to postdoc: derived from comparing the ratio of temporary
to permanent residents from the graduate student to postdoctoral populations; this is an estimated range because
Final Report 32
Biomedical Workforce Working Group Report
The MD and MD-PhD workforce is summarized in Figure 20. Although there are a large number of MDs
in the US, only a small fraction conducts research. The AAMC and AMA collect extensive data on MDs
but the available data do not identify clearly those who conduct research. More information about this
workforce is included in the Physician Scientists section of the Recommendations chapter.
40
Figure 20: Conceptual Framework of the MD and MD-PhD Workforce
some people who are temporary residents as graduate students become permanent residents before starting their
postdoctorate training. Postdoctorate Training - Minimum: GSS, Maximum: estimated from many sources, see
additional details at website page, Median length: SDR, excluding postdocs > 8 yrs. Workforce: SDR
40
Sources of Data:
Medical and Graduate Education & Training – Association of American Medical Colleges, Data Book 2011, 2012;
International – FASEB/GSS; Postdoctoral – FASEB/GSS; Residency – Total PGY-1 Positions Offered - National
Residency Matching Program, 2011, Total Residents – American Medical Association, Physician Characteristics and
Distribution in the US – 2012 Edition, Internationally-Trained Residents – American Medical Association, Physician
Characteristics and Distribution in the US – 2012 Edition; Government – FedScope
http://www.fedscope.opm.gov/; Academic – Association of American Medical Colleges, Data Book - 2011, 2012;
Note: Of the 86,527 FTE MDs, 97% are employed in Clinical Departments; Of the 10,158 FTE MD/PhDs, 86% are
employed in Clinical Departments; PIs supported by NIH – NIH OSAR; Non-Research Non-Clinical – American
Medical Association, Physician Characteristics and Distribution in the US – 2012 Edition; Non-Research Patient
Final Report 33
Biomedical Workforce Working Group Report
Weighing all the data that were analyzed, the working group concluded that the combination of the
large upsurge in US-trained PhDs, continued increased inflow of foreign-trained PhDs, and aging of the
academic biomedical research workforce (i.e. increase in older researchers over time) make launching a
traditional, independent, academic research career increasingly difficult. In addition, the long training
time and relatively low early-career salaries when compared to other scientific disciplines and
professional careers may make the biomedical research career less attractive to the best and brightest
of our young people. Finally, the current training programs do little to prepare people for anything
besides an academic research career, despite clear evidence that a declining percentage of graduates
will find such positions in the future.
The working group’s recommendations are aimed at modifying the career paths to biomedical research
so as to:
• Attract and retain the best and most diverse scientists, engineers and physicians from around
the world to conduct biomedical research as well as increase the number of domestic students
from diverse backgrounds who excel in science and become a part of the STEM workforce.
• Prepare biomedical PhD students and postdoctoral researchers to participate in a broad-based
and evolving economy.
Care – American Medical Association, Physician Characteristics and Distribution in the US – 2012 Edition;
Unemployed – American Medical Association, Physician Characteristics and Distribution in the US – 2012 Edition
IMGs: International Medical Graduates
PGY-1: Post-graduate year 1
FTE: full-time employed
GSS: NSF Graduate Student Survey
FASEB: Federation of American Societies for Experimental Biology
Final Report 34
Biomedical Workforce Working Group Report
RECOMMENDATIONS
Graduate Students
Graduate training historically has been aimed almost exclusively at preparing people for academic
research positions. However, as can be seen in the PhD framework summary (Figure 19), less than half
of US-trained biomedical PhDs go on to a career in academia. Almost a quarter conduct research in
industry or government settings and almost one fifth are in science-related occupations but do not
conduct research (e.g. program and review officers at NIH or managers in the biotechnology industry).
Given the changing face of the biomedically trained workforce, the working group believes that
graduate programs must accommodate greater diversity in anticipated career outcomes for students.
Graduate programs should offer opportunities for students to explore options relatively early in
graduate school, so that they are able to adjust their training to the kinds of careers they will pursue.
Graduate programs also should openly communicate the career outcomes of their graduates to
potential students. This would allow potential graduate students to choose graduate programs that are
more aligned with their career aspirations. Some institutions already post comparable data on the web
sites. For an example, Duke University posts information about outcomes of its PhD students by
program (http://gradschool.duke.edu/about/stats.php).
The long training period through both PhD and postdoctoral training (Figures 3 and 6), increased age at
which researchers obtain a faculty position (Figure 17), and the disparities in early-career earnings and
life-long earnings potential compared to other scientific disciplines (Table 2 and Figure 16) may make a
career in biomedical research less attractive than one in other scientific disciplines and professional
careers. Therefore, the working group believes it would be desirable to shorten the overall training
period (PhD and postdoctoral training) in biomedical sciences. This is reflected in the recommendations
related to graduate students described below and those related to postdoctoral researchers in the next
section.
Specific Recommendations:
• NIH should create a program to supplement training grants through competitive review to allow
institutions to provide additional training and career development experiences to equip students
for various career options, and test ways to shorten the PhD training period. This would mean
that NIH-trained and other US-trained students would become available earlier and would likely
be more competitive for the next phase of their career.
• The best practices resulting from this program will help shape graduate programs across the
country. The working group felt that including multiple types of training (e.g. project management
and business entrepreneurship skills needed in the pharmaceutical and biotechnology industries,
or teaching experience needed for a successful faculty position in liberal arts colleges) would be
particularly valuable for those who go on to conduct NIH-funded research as well as benefit those
students who do not follow the academic research career track. For example:
o Approximately 30% of biomedical PhDs work in the biotech and pharmaceutical industries in
research and non-research positions (Figure 19). Their transition would be more effective if
their training was better aligned with the required skill-sets for these careers. NIH and the
institutions should explore ways to involve relevant employers in the public and private
sector in designing training paths for those students who seek employment in that sector. It
is possible that the pharmaceutical and biotechnology sectors would be willing to partner in
supporting such programs. Another option would be for institutions to develop pilot
Final Report 35
Biomedical Workforce Working Group Report
programs in partnership with private foundations and industry to prepare Ph.D. graduates
for careers that involve translational research and development. Finally, NIH should
encourage the SBIR/STTR awardees to provide internships for graduate students and
postdoctoral researchers to enable increased hands-on training at small businesses.
o Institutions also could be encouraged to develop other degree programs, such as master’s
degrees designed for specific science-oriented career outcomes, such as industry or public
policy. These could be developed as stand-alone programs or provide sound exit pathways
for PhD students who decide not to continue on the research career track. However, this
would require a change in the definition of “success” in the evaluation of NIH training
grants.
• To encourage timely completion of graduate degrees, NIH should cap the number of years a
graduate student can be supported by NIH funds (any combination of training grants, fellowships,
and research project grants), with an institutional average of 5 years and no one individual
allowed to receive support for more than 6 years. Note that a different cap may be needed for
physician scientists (MD, DDS, MD-PhD etc). NIH should continue to assess the pre-doctoral
stipend level annually.
• Today, the vast majority of PhD students that receive NIH support are funded by research project
grants (Figure 2) and yet the NIH has no influence over the quality of the training of these
individuals. Training grants uniquely provide the NIH with a mechanism for peer review of
training, and permit the NIH to require attention to issues such as outcomes, diversity and
professional ethics training. Therefore, to ensure that all graduate students supported by the NIH
receive excellent training, NIH should increase the proportion of graduate students supported by
training grants and fellowships compared to those supported by research project grants, without
increasing the overall number of graduate student positions.
• Even though the NIH training programs are able to fund only a limited number of students, the
existence of an NIH training program at an institution can motivate graduate programs to provide
all students at that institution with training that conforms to NIH guidelines and expectations. To
reinforce this, NIH should revise the peer review criteria for training grants to include
consideration of outcomes of all students in the relevant PhD programs at those institutions, not
only those supported by the training grant. Study sections reviewing graduate training programs
should be educated to value a range of career outcomes. This recommendation could be phased
in relatively quickly.
• The very different requirements and characteristics of training programs at each NIH IC constitute
a substantial burden on the institutions. All NIH ICs should offer comparable training programs
and fellowships and their requirements should be harmonized.
Postdoctoral Researchers
Despite the paucity of the data on the current state of postdoctoral researchers, it is evident that the
postdoctoral period has become a holding pattern for many young researchers. Although a postdoctoral
fellow is considered a trainee, in many laboratories fellows receive little additional preparation for their
future careers, even for those in academic research. For example few postdoctoral fellows receive
instruction in grant writing, laboratory and personnel management, and teaching, all skills that are
necessary for a successful academic career.
Final Report 36
Biomedical Workforce Working Group Report
The majority of postdoctoral fellows are funded by research project grants (Figure 4), which are able to
support the growing number of non-US citizens (Figure 5). In addition, although the average
postdoctoral period has increased only slightly over the years (Figure 6), there is some evidence that
those postdoctoral researchers that go on to an academic tenure-track research career are staying in
postdoctoral training for a longer period (Figure 7) while those going on to other careers such as those in
industry move on after a shorter period of time to higher paid positions. The working group concluded
that the decline in growth of academic positions has led to longer postdoctoral periods, in which fellows
hope to generate more papers in order to be competitive for positions. This system leaves trainees in
subordinate positions at a time when they are expected to be highly productive as independent
investigators.
There is little information about the amounts and types of benefits received by postdoctoral researchers
although anecdotal evidence suggests that there is a wide variation among institutions. In December
2011, the NPA conducted a test survey of institutional policies regarding postdoc compensation,
benefits, and professional development opportunities 41. Almost all the 74 institutions that responded
provide health insurance benefits and about two thirds offer some amount of paid time off. Fewer than
one third of the responding institutions provided retirement benefits. There was a significant difference
in regard to postdoctoral researchers who are classified as employees as compared with those who are
not classified as employees. For example, 64.5% of responding institutions provided paid sick leave to
postdoctoral researchers classified as employees, while only 43% provided paid sick leave to those
postdoctoral researchers not classified as employees. Note that almost all the 74 responding
institutions were NPA members and thus may provide a higher than average level of benefits to their
postdoctoral researchers.
After analyzing these and other data and receiving input from stakeholders, the working group
concluded that the postdoctoral experience should include structured career development, and
incentives should be provided by NIH to move postdoctoral fellows to more permanent positions as
soon as possible. The working group recognizes that after a reasonable period of training – ideally three
years – there is diminished value for the trainee in staying in a subordinate position. Also, the group
feels that those postdoctoral researchers who do not go on to conduct research in an academic setting
should receive training in the skills needed and information about other career options. Finally, the
working group also recognizes that postdoctoral fellows have spent years in graduate training, and
should be compensated accordingly including receiving a reasonable level of benefits.
Specific Recommendations:
• As mentioned above, the vast majority of PhD students that receive NIH support are funded by
research project grants (Figure 4). Training grants uniquely provide the NIH with a mechanism for
peer review of training, and permit the NIH to require attention to issues such as outcomes,
diversity and professional ethics training. This is even more important for postdoctoral
researchers than for PhD students as the students are all in a graduate program, regardless of
their source of support, while many postdoctoral researchers supported by research project
grants have no structured training at all. Therefore, to ensure that all postdoctoral fellows
supported by the NIH receive excellent training and mentoring, NIH should increase the
proportion of postdoctoral researchers supported by training grants and fellowships and reduce
the number supported by research project grants, without increasing the overall number of
postdoctoral researchers.
41
http://www.nationalpostdoc.org/policy/institutional-policies
Final Report 37
Biomedical Workforce Working Group Report
• NIH should create a pilot program for institutional postdoctoral offices to compete for funding to
experiment in enriching and diversifying postdoctoral training, including partnerships with other
entities (industry, private foundations, government, etc.).
• The current stipends for NIH-supported postdoctoral fellows need to be adjusted to levels that
better reflect their years of training. The working group recommends that the NIH should adjust
the starting stipend levels of the Ruth L. Kirschstein National Research Service Awards (NRSA – see
current levels in Table 1) to $42,000 and index the starting stipend according to the Consumer
Price Index (CPI-U) thereafter. Stipend levels should increase with each year of experience in any
postdoctoral position irrespective of their titles by 4% for the second and third years and 6% for
years 4 through 7. The large jump between years 3 and 4 is meant to emphasize a transition from
postdoctoral training to research production, and to incentivize PIs to move fellows to more
permanent positions. This salary scale will apply to postdoctoral researchers supported by
research project grants as well (thus also affecting non-US citizens), and institutions should be
encouraged to adopt this scale for all postdoctoral researchers, irrespective of the source of their
support.
NIH should evaluate this policy in the decade after implementation to determine whether the
postdoctoral period has shortened. If it is not reduced, then perhaps NIH should experiment with
a cap on the length of funding for postdoctoral researchers.
• NIH should require and adjust its own policies so that all NIH-supported postdoctoral researchers
on any form of support (training grants, fellowships or research project grants) receive benefits
that are comparable to other employees at the institution. Such benefits include paid time off,
health insurance, retirement plans, maternity leave etc.
• NIH should double the number of Pathway to Independence (K99/R00) awards, which provide a
proven mechanism for postdoctoral researchers to achieve an independent research position, to
encourage larger numbers of PhD graduates to move rapidly into permanent research positions.
In order to hasten the transition, NIH should shorten the eligibility period for applying to this
program from 5 years to 3 years of postdoctoral experience.
• The working group was supportive of the NIH Director’s Early Independence award program that
facilitates the “skip-the-postdoc” career path for those who are ready immediately after graduate
school. Although this program is its early stages and it is too soon to evaluate career outcomes of
those who have received such an award, the working group recommends that NIH also should
double the number of these awards.
• To provide some structured training experience for all postdoctoral researchers, NIH should
require individual development plans (IDPs) for all NIH-supported postdoctoral researchers,
whether on training grants, fellowships, or research project grants. Assessment of implementation
of this requirement should be included in the review criteria of training grants.
Staff Scientists
The typical academic laboratory consists of a PI and one or a small number of permanent technical staff,
with the majority of the research carried out by trainees. This creates a system in which a large number
of future scientists are being produced each year, well in excess of the number of research-oriented jobs
in academia, government and industry. The working group believes that even a modest change in the
ratio of permanent staff to trainees could have a beneficial effect on the system without reducing the
productivity of the research enterprise. Staff scientists - individuals with MSc or PhD degrees - could
Final Report 38
Biomedical Workforce Working Group Report
play a more important role in biomedical research (one that may become increasingly necessary if the
market for biomedical researchers strengthens outside of the United States in coming years).
Today, these scientists bring stability to many labs and provide important functions as part of
institutional core facilities, but have a wide variety of titles and employment conditions. As an example,
staff scientists constitute an essential part of the NIH intramural research program which employed
1367 of these scientists in 2011 42. In the extramural program, these scientists typically do not apply for
their own grants, but are supported by Research Project, Center and Program Project grants. They
should be differentiated from “soft money” scientists, whose employment depends upon their
successful competition for research funds, a category that has been increasing over the last few years.
As a result of the wide variety of titles held by these researchers, it is difficult to determine the exact
number of staff scientists in the extramural biomedical research workforce. However, an estimate can
be obtained from SDR data. Figure 21 presents a rough estimate of these positions, counting people: 1)
whose primary/secondary work activity is basic or applied research, 2) are working in Academia, 3) are
non-tenure track, 4) are non-postdoctoral researchers, and 5) whose primary/secondary work activity is
not teaching.
43
Figure 21: Staff Scientist Positions, by Degree Field and Organization Type
The working group encourages NIH study sections to be receptive to grant applications that include
staff scientists and urges institutions to create position categories that reflect the value and stature of
these researchers.
Salary Support
Originally the conduct of federally-funded research at universities and other extramural institutions was
based on an understanding that institutions would provide the bulk of facilities and salaries to the
researchers and the NIH would provide the majority of funds for conducting research.
Over the past decades, it seems that this distinction has become increasingly blurred, with NIH providing
an increasing proportion of faculty salary support and the institutions covering a larger percentage of
42
NIH Office of Intramural Research
43
SDR
Final Report 39
Biomedical Workforce Working Group Report
the research costs. This is especially true during the start-up period, which has become significantly
longer as young investigators struggle to receive their first R01 grants (Figure 17). The growth in “soft
money” positions in academic medical schools, in which investigators are required to raise 100% of their
salaries and research funds, has contributed to the negative views of a career in biomedical science, and
has had the additional consequence of encouraging institutions to expand their physical space without
making additional long term commitments to faculty.
That said, however, there is little or no reliable information on the percentage of salary covered by
federal grant dollars. Data from several sources suggests an overall range of 30% - 50% of salaries for
faculty derived from federal funds. For example an Association of Chairs of Departments of
Physiologists Survey found that the average proportion of faculty salary derived from research grants
was 37% in 2009 44. The AAMC Research Metrics Survey 45 found that medical school faculty with
external research support received an average of 36% of total salary support from grants in FY 2009.
The proportion of salary derived from grants ranges from 14% to 67% at different medical schools. The
average was 29% for MD and 49% for PhD faculty. However, both these surveys included small numbers
of institutions only. In addition, there is almost no information about how the distribution of NIH salary
support has changed over the years.
The working group has identified this area as one of the major data gaps and made recommendations
about how to collect this information in the future (see Appendix C). Collecting reliable information
about salaries most likely would require the use of administrative data from institutions (such as the
data collected by STAR METRICS 46 ) and NIH should require institutions that receive NIH funding to
participate in programs that collect such data.
That being said, the working group believes that institutions should provide some fraction of salary
support for their researchers in order to qualify for NIH funding. The group appreciates that any
reduction in NIH salary may have major consequences on institutions.
The working group recommends that NIH consider a long-term approach (over a 20 year period) to
gradually reduce the percentage of funds from all NIH sources that can be used for faculty salary
support.
Physician Scientists
The working group was charged with addressing physician scientist training as well as PhD training. NIH
has a variety of training and career-development mechanisms for physician scientists, notably:
• The Medical Scientist Training Program (MSTP) supported by the National Institute of General
Medical Sciences 47 which supports combined MD/PhD training.
• The NRSA F30 program that supports individuals for combined MD/PhD and other dual doctoral
degree training (e.g. DO/PhD, DDS/PhD, AuD/PhD). This mechanism is not supported by all NIH
ICs.
• The K08 career development award program that provides support and “protected time” to
individuals with a clinical doctoral degree for an intensive, supervised research career
44
http://www.acdponline.org/Surveys/2010Survey.pdf
45
https://www.aamc.org/download/170836/data/aibvol11_no1.pdf
46
https://www.starmetrics.nih.gov/
47
http://www.nigms.nih.gov/Training/InstPredoc/PredocOverview-MSTP.htm
Final Report 40
Biomedical Workforce Working Group Report
48
http://www.lrp.nih.gov/about_the_programs/index.aspx
Final Report 41
Biomedical Workforce Working Group Report
Final Report 42
Biomedical Workforce Working Group Report
undercount of postdoctoral researchers employed in the US. Anecdotal evidence suggests that another
factor may relate to the institutional contacts that provide the annual report to NSF for the GSS. Some
institutions apparently rely on the graduate dean’s office rather than a postdoctoral contact. There is
some evidence that NSF is addressing these issues. For example, in recent years the GSS survey has
been extended to count recent PhDs in non-postdoc positions that are outside tenure or tenure track
job series. Also in recent years, NSF has attempted to collect more complete demographic information
on postdoctoral researchers (citizenship, sex/gender, race, ethnicity, foreign/domestic doctorate, etc.).
This kind of information should be made available as soon as possible.
The SDR, on the other hand, includes only those individuals who have records in the Survey of Earned
Doctorates (SED), i.e. individuals who have doctoral degrees from domestic institutions. The NSF is in
the process of developing an Early Career Doctorate study where the frame includes individuals who
have earned a doctorate from any domestic or foreign institution within the past 10 years and are in
postdoctoral position, faculty positions, or various non-postdoc, non-faculty positions. The information
for this frame will be provided by GSS institutions along with Federally Funded Research and
Development Centers and the NIH Intramural Program. Unfortunately, data from this survey are not
currently available.
Data on MDs and MD/PhDs are primarily collected by AAMC and the AMA, neither of which identifies
the level of research involvement with enough detail to predict long-term trends in the involvement of
MDs in research. Moreover, the AAMC has preferred to share data in aggregated form rather than the
underlying individual-level data. This makes it very difficult to predict and potentially explain the
observed, slow decrease in the participation of MDs as PIs on NIH research grants or perform rigorous
analysis and modeling.
The best sources of information on industrial employment of biomedical researchers are the NSF
Scientists and Engineers Statistical Data System (SESTAT) and the Bureau of Labor Statistics (BLS) surveys
and data collections that result in the Occupational Employment Statistics (OES) data. However, SESTAT
has a 3-4 year lag and the OES does not include information on education levels, making it difficult to
identify jobs for those with PhDs. Because of changes in survey methodology, the BLS says that OES
data cannot be used reliably for time series or trend analysis. In addition, there is currently no
centralized source of information on the number of job openings and employment opportunities for
biomedical researchers with the PhD.
Finally, the working group believes that it is imperative to provide as much information as possible to
prospective graduate students and postdoctoral researchers on career outcomes both nationally and at
their specific training programs so they can make more informed decisions about their future.
Specific Recommendations:
• Institutions that receive NIH funding should collect information on the career outcomes of both
their graduate students and postdoctoral researchers, and provide this information to prospective
students/ postdoctoral researchers and the NIH. Such information should include completion
rates, time to degree, career outcomes for PhD trainees, as well as time in training and career
outcomes from postdoctoral researchers over a 15-year period. Outcome data should be
displayed prominently on the institution’s web site. This will require institutions to track the
career paths of their students and postdoctoral researchers over the long-term. One way to do
that would be to assign graduate students and incoming postdoctoral researchers an identifier
that can be used to track them throughout their careers. This could be part of a unique
researcher ID system that would allow tracking of all researchers throughout their career. The ID
would need to relate to any NIH ID assigned to the individual.
Final Report 43
Biomedical Workforce Working Group Report
• NIH, working with other agencies in the Federal Government, should address the identified data
gaps and collect information on the biomedical and scientific workforce on an ongoing basis.
• NIH should create a permanent unit in the Office of the Director that works with the extramural
research community, the NSF and the NIH ICs to coordinate data collection activities and provide
ongoing analysis of the workforce and evaluation of NIH policies so that they better align with the
workforce needs.
Diversity
Increasing diversity of trainees and the workforce is critical to the future of biomedical research in the
US, particularly as the share of the US population comprised of underrepresented groups increases. The
committee recognizes that this is the responsibility of the entire scientific community but feels NIH
should set an example.
Although the working group recommendations are not aimed specifically at increasing diversity, the
group feels that implementation of these recommendations will increase the overall attractiveness of
the biomedical research career and consequently its attractiveness to underrepresented ethnic and
racial minorities and women.
The working group is aware that another working group of the Advisory Committee to the NIH
Director is focused on this issue but would like to highlight the need for much stronger coordination of
the many diversity-related efforts at the NIH and for rigorous evaluation of the outcomes of all
programs.
Final Report 44
Biomedical Workforce Working Group Report
Modeling Subcommittee
The modeling subcommittee, composed of social scientists (primarily economists) with expertise in the
scientific enterprise as well as NIH-funded investigators with expertise in mathematical models, was
charged with providing input to the working group, particularly on the data and modeling aspects of the
workforce charge (see roster in Appendix A). The subcommittee met three times in 2011 and 2012,
including two in-person meetings and one teleconference. Additionally, the subcommittee maintained
extensive informal communication with each other and outside experts in various areas who provided
input to the subcommittee.
The subcommittee reviewed the options for constructing a model of the current biomedical research
workforce, taking into account the availability of data and the time period of the study. The
subcommittee decided to take a two-tiered approach that includes descriptive analyses and a
conceptual framework (model). This entailed a number of analyses of key aspects of the workforce and
the development of conceptual frameworks to organize the analyses – one each for the PhD and the MD
and MD-PhD workforces (see section on Current Workforce above). The conceptual frameworks were
populated with information on each career stage and transition. The data were organized according to
the conceptual framework to build a comprehensive resource upon which recommendations were
based. These conceptual frameworks are designed to be developed into a full dynamic model at a
subsequent stage.
In addition, the subcommittee discussed the unintended incentives and disincentives underlying the US
biomedical research enterprise and drafted a memo to the working group outlining the incentives
related to NIH policies and practices that might be affecting the behavior of biomedical research
institutions, researchers, and students, particularly those that operate in directions opposite to other
important NIH goals. This memo is attached as Appendix D.
Summary of Subcommittee Findings and Recommendations
The subcommittee reviewed all the workforce data gathered and analyzed (see below for list), and
drafted recommendations for the working group to consider in its deliberations. These
recommendations are attached as Appendix E and summarized below.
While the US biomedical research enterprise is highly productive, data show a weak market for
biomedical PhDs in the period after attaining a PhD and “boom and bust” dynamics during and after the
NIH doubling period. The weak early career outcomes are visible in the form of the long and uncertain
postdoctoral fellowships frequent in biomedicine, especially for researchers focused on academic
research careers; an age at first research program grant close to 42; and a large outflow of women
during the postdoctoral years. Although people are motivated by a wide range of factors, the relatively
weak market can also be seen in early-career earnings that are among the lowest among research
doctoral scientists. The boom and bust dynamics are also visible in trends in the lengths of postdoctoral
fellowships and space investments.
A number of factors underlie the weak early careers. Although the US has benefitted tremendously from
an inflow of highly-trained foreign researchers, the availability of foreign-trained researchers has
Final Report 45
Biomedical Workforce Working Group Report
depressed the market for domestic biomedical researchers. Moreover, our production of biomedical
researchers exceeds new job openings. Consequently a decreasing share of biomedical researchers is
using their research training on their jobs. Compounding these issues, a large body of established senior
researchers makes it harder for young researchers to launch independent careers. The subcommittee
also noted that relatively little information on either the outcomes of people from specific programs or
information on the overall job market is available to help people contemplating careers in biomedical
research make informed decisions.
Based on this broad analysis, the subcommittee sketched 12 recommendations to facilitate the working
group’s discussions (included in Appendix E). These were organized into 3 categories:
(1) The nature and characteristics of biomedical graduate training and postdoctoral fellowships.
These recommendations combined efforts to improve the quality of training (e.g. by shifting
trainees from research grants to training grants and exploring ways of scoring the training
aspects of research proposals); to address the factors that lead trainees, including
underrepresented minorities, to leave the biomedical research enterprise; and to improve the
early-career outcomes of biomedical researchers.
(2) The allocation of resources by NIH. These recommendations focused on policies that could
stabilize the biomedical research system; shift resources in ways that would assist researchers in
launching independent careers; and gradually shift lab staffing from temporary postdoctoral
researchers to staff scientists to reduce reliance on trainees.
(3) Collecting, analyzing, and disseminating data to inform potential trainees about career
prospects and to inform NIH’s own policy making. Some of these recommendations focused on
informing potential trainees about prospects in particular training programs and in general.
Others focused on filling critical gaps in data on the biomedical research enterprise and
establishing an ongoing effort to analyze the biomedical research enterprise to inform NIH
policy.
The subcommittee emphasized that the first two sets of recommendations should be implemented
gradually.
Data gathered and Analyzed
• AAMC and AMA data on career tracks of MDs and MD-PhDs
• Data from OPM on biomedical doctorates employed in the US federal civilian workforce and in
the Public Health Service
• Demographic Characteristics for Graduate Students and Postdoctoral fellows from the Survey of
Earned Doctorates , the Survey of Doctorate Recipients and the Graduate Student Survey in
Basic Biomedical Science, Clinical Sciences, Behavioral & Social Sciences, and Chemistry (used as
a comparison). Data include:
o US Graduate Degrees Awarded
o Time to Degree and Age at Degree
o Age at PhD, First Non Postdoc Job, First Tenure Track Job
o Doctoral Students and Doctorates Awarded by Citizenship/Visa Status and Field
o Gender, Race, and Ethnicity of Doctorate Students and Doctoral Recipients
o Doctorate Students by Type of Support
o US PhDs with Postdoctoral Research Plans
o U.S. Trained Biomedical PhDs in postdoctoral researcher positions, by Time Since Degree
o Biomedical Postdoctorates by Citizenship
o Postdoctorates by Field and Type of Support
Final Report 46
Biomedical Workforce Working Group Report
• Employment Trends from the National Survey of College Graduates, the Survey of Doctorate
Recipients, the National Science Foundation (NSF) Industrial Research & Development Survey,
the Occupational Employment Statistics (BLS), and job openings from Wanted Technologies.
o Employment trends for college graduates and doctorates in a biology field
o Pharmaceutical industry R&D employment (from NSF data to 2008, and BLS data from
2008-2010)
o Up to date data on trends of job openings in biotechnology and pharmaceutical sectors
from wanted.com
o Trends in Employment Status of US Doctorates
o Age Distribution of PhD Workforce, by Field
o U.S. Trained Biomedical and Chemistry PhD employment, by Years Since Degree
o Employment and Un-/Under-employment of US Biomedical Doctorates, including
Employment in Permanent versus Temporary Positions and Unemployed
o Employment Status by Gender and Field
• Academic Employment and Resource Allocations from the Survey of Doctorate Recipients
o Trends in Employment of US-Trained Doctorates in Tenured Academic Positions, by
Degree Field
o U.S. Trained PhDs in academic employment, by tenure track status and by Degree Field
o Proportion of Faculty, by Degree Field and Tenure Track Status
o Tenure Track Faculty by Degree Field and Institution Type
o Tenure-Track Faculty by Degree Field and Gender
o Tenure Track Faculty, By Degree Field and Race/Ethnicity
o University Assignable Research Space, by Field
o Staff Scientist Positions, by Degree Field and Organization Type
o Proportion of Staff Scientists, by Degree Field and Gender
o Proportion of Staff Scientists, by Degree Field and Citizenship
• Relationship between Degree and Employment from the Survey of Doctorate Recipients
o US Doctorate Employment: Relationship to Degree Field, by Field and Years Since
Degree
o Relationship between Life Sciences, science and engineering PhD field and occupation
o U.S. Trained Biomedical PhDs in research occupations, by Years Since Degree
o US Doctorate Employment: Hours of Work by Degree Field
• NIH Funding Trends
• Market analysis of the biotechnology sector and job market (from publicly available reports).
• Healthcare and Pharmaceutical spending (from publicly available reports).
• Stay rates of foreign doctoral students (from publicly available reports by Mike Finn, Oak Ridge
Institute for Science and Education)
• Earnings of Doctoral Recipients
Suitable comparisons were sought to benchmark outcomes where possible. The resulting data are
incorporated into various sections of this report, included in Appendix B, or posted at
http://report.nih.gov/investigators_and_trainees/ACD_BWF.
Stakeholder Input
The working group met on the NIH campus on June 21, 2011 and heard various stakeholder perspectives
on the biomedical research workforce and its future direction. Specifically, the speakers were asked to
address the question: “Given the current fiscal climate and the prospect of little or no growth in the near
Final Report 47
Biomedical Workforce Working Group Report
future, and within your area of expertise, what in your view are the major issues the task force should
consider?”
The working group heard from representatives of the National Postdoctoral Association and several
scientific societies, a Dean of an academic Medical School, a Director of an MSTP program, and people
involved in two recent National Research Council studies - Study on Research Training and the
Biomedical, Behavioral, and Clinical Research Sciences and the Report on Expanding Underrepresented
Minority Participation, Director, National Research Council’s Board on Higher Education and Workforce.
The participants of the meeting are included in Appendix F.
49
Broadened to “Post-doc training characteristics” in the analysis of RFI responses
Final Report 48
Biomedical Workforce Working Group Report
219 entities (individuals and institutions) responded to the RFI, with 75% providing personal input. A full
report of the responses and the resulting analysis is included as Appendix G. Following are the main
points.
The comments were parsed into 498 “quotations” representing unique ideas, with an average of 2.3
quotations per commenter. The analysis identified four primary issues in addition to the original eight
primary issues included in the RFI 50. The distribution of primary issues, as cited by commenters, is shown
in Figure 22.
120 25%
19%
100 17% 20%
80 12%
11% 15%
60 9%
7% 6% 6% 10%
40 5% 4%
20 2% 1% 5%
0 0%
Commenters found that most, if not all, of the primary issues were critical to the development of a
sustainable biomedical workforce. About two thirds of the comments included a secondary issue in
addition to the primary issue. The secondary issues mentioned by the commenters were recorded to
help describe the overlapping and interlocking nature of the issues.
In cases where the primary and secondary issues are similar, the secondary issue covers only certain
aspects of the primary issue. For example, the secondary issue of Career Appeal covers the specific
issue of working conditions; whereas the primary issue of Biomedical Research Career Appeal includes
all issues related to the attractiveness of biomedical research careers (e.g. salary, availability of research
funding, working conditions).
The overlapping issues were:
• Funding. Uncertainty and lack of funding, distribution of funding, restricted paylines, success
rates, and excessive competition
• Multi-disciplinary training. Need for multi/ inter/ trans-disciplinary research training to prepare
individuals for a wide range of academic and non-academic career opportunities
• Salary. Inadequate compensation and benefits
• Length of Training. Amount of training time too long to be feasible for majority
• Non-US Citizens. Foreign students and post-doctoral fellows
• Career appeal. Working conditions (e.g. heavy workload, perception of being perceived as
cheap labor, long work hours)
• Mentoring. Quality of career development and the need for pre-college preparation
50
Identified with an asterisk in the chart
Final Report 49
Biomedical Workforce Working Group Report
Final Report 50
Biomedical Workforce Working Group Report
appealing today as a result of lower starting salaries for graduates and the increasing competition for
limited research funding. Individuals seemed more concerned than institutions with career aspects such
as family-friendly work environments, long hours, high stress, and benefits packages. Institutions
expressed greater concern with regulatory burdens placed on current investigators. Many individuals
viewed long-term institutional commitment unfavorably, citing the erosion of tenure positions and the
increasing burden to fund one’s own salary.
Some commenters pointed out that lack of information about graduate degrees and career
opportunities led many to be dissatisfied with their career in biomedical research. Improved mentoring
and career development, as well as anything that addresses the supply and demand imbalance, were
perceived as the best solutions.
Individual commenters cited the pressure to increase clinic time as the main reason why the clinician-
scientist path is less attractive and attainable. Commenters felt that the time required to conduct
research is often not adequately compensated by the institution (in terms of pay, recognition or career
advancement). Commenters believed that MD-PhD degrees are valuable, but that the increasing cost of
medical school, requirements and length of training, and the limited opportunities in academia make
the MD-PhD career path less attractive. Multi-disciplinary training and funding levels were the two
secondary issues mentioned most often in these responses.
Institutions and individuals agreed that a broader approach to training was needed and
clinical/translational training should begin sooner – in medical and undergraduate school. Both
individuals and institutions suggested that funding support from NIH is needed to provide clinicians with
protected time to maintain research activities.
Individual commenters supported the idea of creating a permanent staff scientists career track. They
saw this as a way for all parties to reap the benefits of training support provided by NIH. Institutional
commenters where divided, some taking a cautious approach to the idea of utilizing staff scientist in the
lab, citing possible adverse effects including potential loss of innovative ideas (currently provided by
graduates) and the reduction in project budgets to cover the salaries for these positions. Funding levels
was the secondary issues cited most often in these responses.
Commenters believed that the biomedical workforce cannot be evaluated without addressing the need
for diversity. They expressed the fear that the current economy and funding climate will have an adverse
affect on the diversity of the workforce. Most commenters agreed that diversity should remain a priority
in any proposed policy changes but there were few specific recommendations. Half these responses
mentioned funding levels and 21% mentioned salary levels.
Effects of NIH Policies were mentioned throughout the responses. Commenters cited NIH policies and
practices that positively and negatively affected the workforce, and offered possible solutions.
Commenters addressing the issue of funding policies expressed a range of reasons for being dissatisfied
with the current funding review system. The predominant reason, for both institutions and individuals,
was that securing funding has become increasingly difficult, particularly for new investigators. 88% of
these responses mentioned funding levels.
Commenters reported that the quality of mentoring varies immensely, and it has a significant effect on
mentee’s perspectives and career paths. Institutional commitment to career development resources
also was reported to vary. Mentoring can improve or compound other issues such as diversity, length of
training, and biomedical research career appeal. Institutions and individuals both described conflicts of
time and interest for PIs when mentoring students and post-doctoral fellows. Of commenters
addressing non-academic career paths, most agreed that training for these paths is inadequate. Multi-
Final Report 51
Biomedical Workforce Working Group Report
disciplinary training and funding levels were the secondary issues mentioned most often in these
responses.
Anything that would improve current funding success rates could address this conflict. Also, many
commenters suggested that all NIH-funded students and post-doctoral fellows should have documented
individual development plans (IDPs).
Few commenters (19) addressed the issue of training to research grant ratio. Those that did comment
felt that there was a need for more training grants because their flexibility allows for better career
development of funded trainees. This belief was held by institutions as well as individuals. Funding
levels and mentoring were the secondary issues mentioned most often in these responses.
Institutions requested an in-depth evaluation be conducted to understand the potential impact of
moving students and post-doctoral fellow support off of research grants and onto training grants.
Some commenters suggested that some of the primary issues may have roots as early as K-12
education. Early education intervention programs prior to graduate school would likely have
downstream effects on issues such as Biomedical Research Career Appeal, Diversity, and Supply and
Demand. Funding levels and mentoring were mentioned most often. Suggestions generally called for
an increase in funds devoted to programs that would affect the K-12, undergraduate, and post-
baccalaureate student populations.
Industry Partnership – Some commenters viewed the relationship between industry and academia as
imbalanced in terms of benefits and burdens. Specifically, industry was thought to share more of the
benefits and academia was thought to share more of the burdens. Although fewer than ten commenters
made specific reference to partnerships between academia and industry, this was defined as a primary
issue since so many individuals felt that industry was a vital part of the branching career pipeline for
biomedical workers. Individuals submitted all but one of the comments on this issue.
fellowship and career development awards) versus training supported by research grants. Peer review
of the formal training programs includes the science proposed, as well as the proposed training and the
training potential. Formal training awards can be targeted to emerging scientific disciplines and to areas
of national need. Of particular note, formal training programs have played a vital role in creating
incentives for increasing the diversity of the workforce pipeline.
The NTW committee did not distinguish among the careers of individuals engaged in biomedical
research in academia, industry, government, or any other sector; these were not considered ‘non-
traditional’ or ‘alternative’ careers. The committee felt it was important not to confuse these careers
with truly ‘alternative’ or non-traditional careers, such as science policy, law, finance, and teaching at K-
12 levels or at institutions where there is little or no research activity.
The NTW committee drew a distinction between the career paths of the majority of biomedical research
scientists who are PhD recipients, and the much smaller population of clinician scientists (individuals
holding MD, DO, DDS, DVM, DN, or equivalent clinical doctorate degrees who are research scientists).
Clinician scientists generally do not receive support from the NIH for their clinical training, except for
those enrolled in formal combined degree programs. The NTW committee noted the differences in
immediate career options for clinicians, who may fall back on clinical practice for their livelihoods, and
PhD researchers, whose options are more limited. On the other hand, clinician scientists face demands
on their time for research activities that are not generally encountered by PhD scientists.
There are more policy options and leverage at the beginning of the career but it is not easy to address
the current population of individuals already at the Early Career stage. In the Early Career stage, the
committee struggled with how to retain highly trained individuals in the research workforce without
affecting the research grant budget. For this reason, the committee chose not to recommend reduction
the large postdoctoral pool directly but to make changes earlier in the training pipeline and allow
workforce supply and demand to self-correct.
The scenarios in each career stage represent moderate to significant policy changes required of the NIH,
depending on the desired outcome. There were varying levels of support for the different scenarios
from the full NTW committee; however, most disagreement was manifested in the scenarios for the
Early Career stage, for the reasons mentioned above. Committee members noted that solutions that
work for one biomedical discipline may not work for another, and that particular NIH Institutes and
Centers might have unique issues. Finally, for all scenarios, there was concern that some of the policy
changes could affect efforts to create a more diverse workforce, but there was agreement that
adjustments in policies were possible to minimize or eliminate these effects.
1. Predoctoral Career Stage
a. Training Graduate Students Exclusively for Research Careers. Focuses on the selection
(at the time of graduate school admission) of those students most likely to pursue
careers as research scientists. Incentivizes graduate schools to be more selective in
their admissions process, making them more financially responsible for supporting
students who do not successfully compete for formal training awards and restricting the
use of NIH research grants for such support.
b. Tracking Graduate Students into Research or Non-Traditional Careers. Acknowledges
the large number of students who enroll in biomedical science graduate programs but
moves them into ‘non-traditional’ career training tracks. This scenario accepts the
current graduate admissions process, but requires institutions receiving NIH support for
research or formal training to track the career intentions of supported students. Formal
NIH training support would be reserved for those students exhibiting clear plans to
Final Report 53
Biomedical Workforce Working Group Report
Final Report 54
Biomedical Workforce Working Group Report
b. Supporting Career Staff Scientists. Proposes a program to support staff scientists as part
of a team of individuals providing core services to laboratories with NIH research grant
support. These individuals would not be expected to become independent research
scientists but would rather bring new technical competencies to the institution and
provide intellectual input to the supported laboratories. The purpose of this program
would be to provide an NIH supported career pathway for highly trained research
doctorates who might otherwise leave the workforce, while at the same time perhaps
creating a more efficient research infrastructure.
4. Clinician Scientists
a. Making Research Careers More Attractive for Clinician Scientists. While there are many
PhD holders seeking limited research jobs, the number of clinician scientists at the
postdoctoral level (post-residency) is much smaller. Additionally, only a small fraction of
medical students express interest in careers as independent research scientists.
MD/PhD programs, however, have been quite successful in placing graduates in
independent research positions, when compared to PhD programs
The NTW committee focused on the challenges to becoming a clinician scientist and
considered the critical need for a comprehensive workforce infrastructure to support
clinical-translational research as beyond the scope of this exercise. The clinician
scenario includes actions that NIH could take to support the careers of clinician
scientists at three career stages: medical school, residency, and early career. For
medical students, NIH could support more programs for short-term research
experiences during medical school, and create more awards for ‘year out’ programs. At
the residency level, NIH could create awards that support ‘fast-tracking’ in residency
programs, thereby allowing clinicians to begin research careers sooner. At the early
career stage, the NIH Loan Repayment Program could be extended to more clinicians by
broadening the eligibility criteria. Finally, the NIH could create incentives for more team
approaches to research that involve collaborations between clinician scientists and
PhDs.
Final Report 55
Biomedical Workforce Working Group Report
GLOSSARY
Final Report 56
Appendix A Biomedical Workforce Working Group Report
APPENDICES
1. Develop a model for a sustainable and diverse U.S. biomedical research workforce that can
inform decisions about training of the optimal number of people for the appropriate types of
positions that will advance science and promote health. Developing the model will include an
analysis of the current composition and size of the workforce to understand the consequences
of current funding policies on the research framework. The model should include an
assessment of present and future needs in the academic research arena, but also current and
future needs in industry, science policy, education, communication, and other pathways. The
model will also require an assessment of current and future availability of trainees from the
domestic and international communities.
2. Based on this analysis and input from the extramural community, using appropriate expertise
from NIH and external sources, and recognizing that there are limits to NIH’s ability to control
many aspects of the training pipeline, the committee will make recommendations for actions
that NIH should take to support a future sustainable biomedical infrastructure.
Final Report 57
Appendix A Biomedical Workforce Working Group Report
Final Report 58
Appendix B Biomedical Workforce Working Group Report
Graduate Students
Final Report 59
Appendix B Biomedical Workforce Working Group Report
Final Report 60
Appendix B Biomedical Workforce Working Group Report
Postdoctoral Researchers
Final Report 61
Appendix B Biomedical Workforce Working Group Report
Career Outcomes
Final Report 62
Appendix B Biomedical Workforce Working Group Report
Final Report 63
Appendix B Biomedical Workforce Working Group Report
Final Report 64
Appendix B Biomedical Workforce Working Group Report
Final Report 65
Appendix B Biomedical Workforce Working Group Report
Final Report 66
Appendix B Biomedical Workforce Working Group Report
Final Report 67
Appendix B Biomedical Workforce Working Group Report
Demographics
Final Report 68
Appendix C Biomedical Workforce Working Group Report
Final Report 69
Appendix C Biomedical Workforce Working Group Report
4. Data on researchers and especially faculty who receive salary support from NIH research
grants.
a. The NIH should work with the NSF to include questions in the SDR on the percentage of
salary covered by federal grants.
b. NIH should further explore the use of administrative data such as that collected by STAR
METRICS to identify the proportion of salary derived from federal sources.
c. NIH should consider ways to identify and track staff scientists and people with other
titles functioning as staff scientists. One way would be to expand and perhaps
automate the collection of all-personnel data on annual progress reports to increase the
Final Report 70
Appendix C Biomedical Workforce Working Group Report
ease of preparing extramural staff enumeration reports through programs such as STAR
METRICS.
Final Report 71
Appendix D Biomedical Workforce Working Group Report
D: Incentives Memo
MEMO TO: Biomedical Workforce Working Group, Advisory Committee to the Director, NIH
FROM: Modeling subcommittee
SUBJECT: Some unintended incentives and disincentives underlying the US biomedical research system:
Date: October 21, 2011
The subcommittee discussed the response of investigators and institutions to NIH policies. This memo
outlines some of the incentives and disincentives related to NIH rules and practices that might be
affecting the behavior of biomedical research institutions, scientists, and students (including potential
students). We agreed that the emphasis should be on unintended rather than intended incentives, and
of course especially those unintended disincentives that might be in opposition to other important NIH
goals.
The memo reflects the subcommittee discussion (but much of the background is highly nuanced, making
it possible that some of the characterization is not entirely correct) and is offered as input to the
working group. We hope it will serve as a catalyst for further discussion, correction, or clarification.
Please note that while many of these incentives and disincentives seem to be substantially under NIH
influence, others may be mandated by law or regulation outside of NIH control.
The underlying concern is that the US biomedical research system has been showing repeated signs of a
lack of stable equilibrium. An incomplete list of such symptoms might include the following:
1. While the US biomedical research system is heavily funded, far more than for any other fields of
basic science, 51 it appears to have a tendency to expand beyond the available funds. When the
14-15% annual budget increases of the doubling period ended in 2003, there was widespread
discussion of a new “funding crisis”, although the funds available were nearly double those of
only 5 years earlier (somewhat less if adjusted for inflation).
2. According to one well-informed analysis, the biomedical research system is structured for
continuous growth, and becomes unstable unless NIH receives at least a 6 to 8% annual budget
increase. 52
3. Careers for junior biomedical scientists appear to be unattractive relative to other fields
requiring lengthy post-baccalaureate education and postdoctoral training. Average
remuneration for biomedical researchers is among the lowest reported for scientists,
notwithstanding that the largest volume of Federal funding goes to the biomedical sciences.
51
NIH’s research budget is more than 4 times that of the National Science Foundation. Among US government
agencies the NIH budget is exceeded only by the R&D budget of the Department of Defense, but the latter is heavily
concentrated upon the “D” or development side of R&D. For an informative overview, see Congressional Research
Service, Federal Research and Development Funding: FY2011, March 10, 2010,
http://assets.opencrs.com/rpts/R41098_20100310.pdf
52
David Korn, et al.,”The NIH Budget in the ‘Postdoubling’ Era,” Science, 296, 24 May 2002, p. 1402.
Final Report 72
Appendix D Biomedical Workforce Working Group Report
4. The system generates far more research proposals than can be funded. The trends suggest
increasing time devoted by biomedical researchers to proposal preparation, as the numbers of
proposals per researcher increase and success rates remain low or decline further.
Much of these symptoms of instability are structural in nature, driven by a number of destabilizing and
presumably unintended feedback loops (see discussion in Teitelbaum, 2008 53). Careful analyses of these
system dynamics, accompanied by appropriate adjustments of current incentives and disincentives, may
enable NIH to incrementally guide the research system that it funds and supports toward a more stable
equilibrium, without doing injury to its outstanding quality and significant contributions to human
welfare. Indeed, in addressing such system issues it is critical to attend to the system equivalent of the
Hippocratic Oath: any prescriptions to mitigate negative symptoms should, above all, “do no harm.”
To this end, improving our understanding of the outcomes of at least the following incentives and
disincentives may provide useful insights and guidance.
#1: Incentives that may increase financial leverage and vulnerability of research institutions and
researchers
Issue 1-A:
NIH practices allowing unlimited percentages of faculty salaries to be covered by NIH grant funds:
NIH imposes no limits on the percentage of faculty salaries that can be paid from NIH research funds
(although there is a cap on the overall amount of salary that can be paid). 54 Payment of large fractions
of a faculty salary under an NIH grant appears to be most common in medical schools rather than in
arts/sciences faculties.
In contrast the NSF generally limits salary compensation for senior project personnel to no more than
two months of their regular salary in any one year. NSF does allow payment of full salaries for other
personnel such as postdocs.
Unintended incentives:
• Faculty
o Faculty whose own salaries and benefits depend heavily upon external grant funding
face strong incentives to commit substantial proportions of their time and effort in
53
Michael S. Teitelbaum, “Structural Disequilibria in Biomedical Research,” Science 321, 1 August 2008, pp. 644-5.
54Payment must be in proportion to the effort devoted to the research project, and there is a cap on the amount of
salary that can be charged to an NIH grant, linked to the Federal Executive Pay scale – until recently $199,700,
reduced by Appropriation action in later 2011 to $179,700 for grants with Initial Issue Date on/After 12/23/2011..
http://grants.nih.gov/grants/policy/fy2012_salary_cap_faqs.htm
Final Report 73
Appendix D Biomedical Workforce Working Group Report
pursuit of new/renewed external research funding, not only in support of their research
interests but also of their own personal compensation.
o Due to the difficulties and uncertainties of obtaining external research funding, faculty
in such circumstances also are subject to higher risks of funding gaps that threaten
research programs, stipend support for their postdoc research assistants, etc.
• Institutions
o Current NIH policy allows/incentivizes institutions to encourage faculty to maximize the
proportion of their salaries and benefits paid by external but ultimately “soft” funding
sources such as NIH research grants.
o Salary paid under NIH research grants also is included in the base for calculating indirect
costs; the resulting additional indirect funds may be especially attractive to university
administrators.
In the words of Bruce Alberts, “NIH actually rewards institutions for paying
faculty salaries with unguaranteed “soft money” from research grants by
providing increased overhead payments. Amazingly, any institution that draws
on its own finances to pay its professors is doubly disadvantaged: It must not
only use its own funds but also loses the overhead on the salaries that it would
otherwise accrue.” 55
o Since money is fungible, substitution of “soft” for “hard” funds may incentivize the
institution to expand its faculty and attempt to capture additional external direct and
indirect funding beyond levels that otherwise would be possible.
o This seems to be more of an issue in medical schools, less so in the Arts and Sciences
units, where tuition and teaching constitute a larger part the revenue stream and faculty
effort of science departments. 56
o This strategy may work well when NIH funding is growing robustly, but can become a
significant and even threatening burden when real NIH budgets are flat or declining.
o Faculty positions (including even those protected by tenure) and labs become
increasingly unstable as success rates drop.
o It is important to note that NIH limitations on payment of investigator salaries would
make participation in NIH supported research financially more difficult for many
institutions at a time when there are recession-related funding pressures on clinical
revenues, endowment income, charitable donations, and state support.
Issue 1-B:
Indirect cost regulations incentivize debt financing of biomedical research facilities:
OMB Circular A-21, which controls NIH practices regarding institutional indirect costs, generally allows
inclusion of the costs of debt service in indirect cost calculations, but prohibits inclusion of the imputed
value of similar laboratory facilities that are not debt financed. 57
55 Bruce Alberts, “Overbuilding Research Capacity,” Science, Vol. 329 (10 September 2010), p. 1257.
56 S. A. Bunton, W.T. Mallon, “The Continued Evolution of Faculty Appointment and Tenure Policies at U.S.
Medical Schools,” Academic Medicine, 82, 3 (March 2007), 282.
57
Indirect cost claims are based on the assignment of space, i.e. unused research space cannot be included.
Final Report 74
Appendix D Biomedical Workforce Working Group Report
Unintended incentives:
• Presents universities with incentives to borrow funds for building, renovation, or re-financing of
biomedical research facilities. (The extreme case might be the “condo labs” format that
reportedly has appeared at some institutions).
• Resulting increases in financial leverage implies heightened financial pressures on institutions
unless sufficient grant funds can continually be obtained to cover debt-servicing costs. This is
especially problematic for institutions during times of real declines in NIH budgets.
• However, A-21 also has several countervailing provisions, requiring for example that institutions
provide at least 25% of the equity in cases in which debt financing exceeds $1 million. In
addition, NIH staff knowledgeable about OMB Circular A-21 have expressed the view that while
the effects of such unintended incentives are possible in theory, they may not be widespread.
• To address such issues, options might include the gradual phasing in of: (1) a limit on NIH salary
reimbursement to, for example, 50%; (2) An overhead cost penalty proportionate to an
institution’s fraction of soft-money positions; (3) Eliminating incentives that currently favor
debt-financing of laboratory facilities by limiting indirect cost reimbursement for debt service, or
by allowing such reimbursement for the imputed costs of wholly-owned facilities.
#2: Heavy reliance upon NIH research funding for support of graduate students and postdocs
Issue: Approximately 2/3 of NIH-supported graduate students and postdocs are supported under R01
and related research grants. This may be compared with only 22% supported under NRSA and related
training-type funds. 58 As recently as 1980 these proportions were the opposite, with about 2/3 of NIH
support graduate students and postdocs supported under training-type funding.
Unintended incentives:
• Under the current structure, if NIH research funding increases the number of “slots” for
graduate students/postdoc research assistants should be expected to expand accordingly. This
expansion would occur even in the absence of any plausible growth in career demand for those
who are thus supported. Data from the 1998-2003 doubling suggest that the number of
graduate students indeed does increase in response to increased research funding. Meanwhile
the postdoc population seems to grow at about 4 percent per year relatively independent of NIH
research funding (something of a puzzle that warrants some examination).
• The instability/unpredictability of annual NIH budgets further complicates the situation, given
that there are multi-year lags between entry and completion of both the PhD and the postdoc.
• A gradual, incremental shift from the current 2/3 of NIH graduate students supported under
research grants to about 50% was recommended by the 2000 NRC committee report
commissioned by NIH. 59 This would enable NIH to better relate the number of PhD students
and postdocs it supports to plausible demand and career prospects for PhD researchers, vs. the
58
Committee to Study the National Needs for Biomedical, Behavioral, and Clinical Research Personnel, National
Research Council, Research Training in the Biomedical, Behavioral, and Clinical Research Sciences (Washington,
DC: National Academies Press, 2011), p. 13. http://www.nap.edu/catalog.php?record_id=12983 .
59
Committee on National Needs for Biomedical and Behavioral Scientists, “Addressing the Nation's Changing
Needs for Biomedical and Behavioral Scientists,” National Research Council, 2000.
http://www.nap.edu/catalog/9827.html
Final Report 75
Appendix D Biomedical Workforce Working Group Report
current system in which the numbers of graduate student/postdoc “slots” are driven by the
volume of research grant funding available. 60
#3: Insufficient information provided on career outcomes of recent PhDs and postdocs.
Issue: Most biomedical departments provide prospective new graduate students and postdocs with
little or no credible information as to the career outcomes experienced by those most recently filling
such positions.
The situation is notably different for e.g. law or business schools at the same university. (However, see
recent lawsuits filed by former law students against an increasing number of US law schools, alleging
fraud in the career data provided by these schools.) 61
Unintended incentives:
• Prospective new graduate students and postdocs who are offered financial inducements to
enter a lengthy PhD/postdoc sequence may a lack realistic understanding of their prospects in
the biomedical research labor market beyond the PhD/postdoc, particularly their prospects for
obtaining a tenure-track research faculty position. Indeed availability of Federal financial
inducements may lead students to overestimate funding prospects after degree completion, e.g.
by suggesting that NIH is anticipating a shortage of individuals to fill anticipated post-training
positions.
• Non-alignment of demand in the labor market with supply of completing PhD/postdocs likely
contributes to long time-to-PhD, lengthening postdoc and cognate temporary positions. This
can result in frustration (even anger) among those who feel they have been misled.
• This may provide impetus to toward unionization of graduate students and postdocs (e.g.
University of California), and also create a negative feedback loop of communication from
discouraged graduate students/postdocs to undergraduate majors in the department, who may
be thereby deterred from continuing to pursue a science career.
• This may also have unintended effects upon the type of person attracted to careers in
biomedical research. Those individuals with the greatest range of opportunities or those that
have plans and aspirations for a certain standard of living may opt for other careers.
• To address such concerns, NIH could consider: (1) establishing a small expert analytic unit
(presumably in-house) charged with monitoring trends and prospects for education and careers
that are central to the current and future biomedical research workforce; 62 (2) supporting a
60
During the budget doubling, the total number of training slots remained flat although the training budget
increased due to increased stipends.
61
Katherine Mangan, “Law Schools on the Defensive Over Job-Placement Data,” Chronicle of Higher Education, October 16,
2011. http://chronicle.com/article/Crisis-of-Confidence-in-
Law/129425/?key=SzgmclZmNCsRNH00YG0SZDpXPXRqZk4kZ3NJPn5wblpRFw%3D%3D See also:
http://www.law.com/jsp/nlj/PubArticleNLJ.jsp?id=1202517930210&Another__law_schools_targeted_over_jobs_data&slreturn=
1
62
The mandates for such a group might include:
Final Report 76
Appendix D Biomedical Workforce Working Group Report
website with information on outcomes of PhD education and postdocs, including links to other
sources such as scientific society websites; (3) urging or requiring grantee institutions to collect
and provide credible data on career outcomes of their former PhDs and postdoc; (4) supporting
a web-based career clearinghouse for biomedical researchers.
Unintended incentives:
• Use of temporary postdocs as the primary lab workforce brings new ideas, abilities, and energy
into labs, as often noted. However it also brings lack of experience as compared with longer-
term research staff. It also places people in a subordinate role during what is often a highly
creative portion of the career.
• May tend to extend years in postdoc status, increasing opportunity costs (see below) for
trainees.
• Unintentionally limit independent creative contributions by researchers in their late 20s and
early 30s, such as those made in the past by now-prominent senior biomedical researchers.
• May make postdoctoral positions less attractive, especially for women as they extend into the
30s. There is some evidence of declining interest in postdocs among recent US-citizen PhDs, but
this does not affect recruitment of postdocs since institutions have ready access to large
international pools of recent PhDs (see related item, below). The large, available pool of
postdocs reduces the creation of stable, long-term employment of lab technicians and staff
scientists.
- provide ongoing insights on education and workforce trends to the NIH Director, Institutes, and the
public;
- develop new analyses from both existing data collected by the NSF and other Federal agencies, as well as
from internal NIH administrative data;
- develop credible data on postdocs in US institutions, most of whom appear to be in biomedical fields;
- develop new data series such as measures of the changing “inventory” of postdocs relative to prospective
career openings;
- track ongoing trends and composition of enrollments and completions of biomedical PhDs and postdocs
- assess knowledge and views of entering PhD students and postdocs about career paths, both in terms of
general trends and what they think their own personal outcomes will be.
63 National Science Board. 2012. Science and Engineering Indicators 2012. Arlington VA:
National Science Foundation (NSB 12-01), pp. 3-36 to 3-40.
Final Report 77
Appendix D Biomedical Workforce Working Group Report
Issue: NIH funding policies, US immigration policies, and the intersection between them all are lacking in
coherence and coordination. Moreover there has been no serious assessment of the costs and benefits
as well as no reliable way to measure the number of foreign-trained postdocs.
Under NIH rules, international students and postdocs can be financed without limitations under NIH
research grants, but the same students/postdocs are excluded from support from NIH training funds.
Under current immigration law, universities have essentially unlimited access to rapidly growing global
pools of prospective graduate students and postdocs.
For postdocs, the international category may now represent a majority in biomedical research; China
may be largest source country. Unlimited numbers of international postdocs are admissible on
temporary H-1B visas -- the numbers are capped for corporate employers, but there are no numerical
limits for universities and nonprofit research institutions. The large numbers of H-1B visas are
uncoordinated with very limited numbers of permanent visas for those with high levels of education and
skills. Combined with per-country limits this lack of coordination has led to large backlogs of former H-
1B visa holders awaiting permanent visas and competing for employment opportunities in order to
secure these visas.
Unintended incentives:
• Ready access to global labor markets makes large numbers of science workers available but
depresses the domestic market for biomedical researchers. If postdoctoral positions are allowed
to become unattractive to citizen PhDs, US universities can and will readily recruit international
postdocs without making adjustments to postdoc conditions.
• Many international postdocs are self-recruiting, and some bring partial funding from their home
governments. Most have few postdoctoral opportunities available in their own countries.
• Living stipends for graduate students and postdocs are still low by US standards, but may be
high by the standards of a low-income country.
• The primary economic cost to would-be biomedical researchers is the income and benefits they
forego (what economists call “opportunity costs”) in the course of lengthy PhDs and extended
postdocs. As compared with US-citizen PhDs with comparable talents and knowledge, PhDs
from low-income countries have lower opportunity costs -- their earnings at home would be
lower, and unlike US-citizen PhDs they are not eligible for regular employment in the US.
• International postdocs also benefit from additional incentives that do not accrue to citizen
postdocs – the possibility of a permanent visa allowing US employment, or the prestige of
having held an international research position if they decide to return to their home country.
Issue: The annual Congressional appropriations process means that NIH funding is very difficult to
anticipate beyond a year or two. Due to the same Congressional budget process and substantial
lobbying by health-related associations and universities, NIH research funding is subject to booms and
comparative “busts” (e.g. the 1998-2003 doubling, followed by flat/declining budgets until 2009,
followed by 2 years of ARRA stimulus funding).
Final Report 78
Appendix D Biomedical Workforce Working Group Report
Unintended incentives:
• PhDs and postdocs
o If financed by temporary booms in research funding, expanded numbers of completing
PhDs and postdocs emerge following a multiyear lag, by which time the expansive
funding period may well have ended.
o Results may include: extension of temporary employment status; discouragement;
departure from research careers, etc.
• Newly-hired faculty
o Additional faculty hired during boom phase experience increasing competition for
research grant funding after boom wanes.
o Challenging NIH pay lines, especially for more junior investigators, may disrupt
promising research careers.
o If newly-hired faculty are dependent on soft money for substantial portions of their
salaries as well as for research itself, serious personal financial challenges may result.
• Established research faculty:
o When NIH pay lines decline to low levels, even highly productive researchers with
established records of success with NIH grants may experience lengthy gaps in renewal
support, thereby disrupting successful research careers and sometimes even leading to
closure of successful research labs.
o Lower success rates lead researchers to devote more of their time to writing proposals,
which in turn leads to more time/effort expended in the peer review process.
o If the terms of successful grant proposals are shortened or cut to stretch limited
funding, this may also imply more time writing proposals, leading also to more
time/effort expended in peer review.
o Shorter-term grants also are less appropriate for supporting graduate
students/postdocs.
• Institutions:
o Those that either expanded their faculties or used debt-financing to expand research
facilities -- in anticipation of continued increases in research funding and indirect
support -- likely will face serious problems during budget stringencies in covering
unfunded faculty salaries and in servicing their debt.
o Problems are magnified if institutional revenues from endowments or state
contributions are also reduced, as has been happening in recent years.
o Budget instability also may contribute to reliance on postdocs (inherently temporary
positions) and discourage institutions from hiring more-permanent staff scientists.
Issue: High levels of competition for NIH research funding in a peer review process may incentivize
increasingly narrow, focused, and hence less risky research proposals. Even if the grant proposals
themselves remain innovative, the uncertainty of obtaining renewed funding after the first grant period
may lead researchers to pursue potentially less risky (and perhaps less innovative) avenues of research
to achieve their aims.
Final Report 79
Appendix D Biomedical Workforce Working Group Report
Unintended incentives:
• Many biomedical researchers believe that the greatest potential for scientific advances lie in
broader, interdisciplinary research efforts.
• Narrower and more specialized research in turn leads to a narrower focus for PhD and postdoc
training financed by such grants.
• The uncertain career prospects faced by PhD students and postdocs suggests they would be
better prepared to pursue successful research careers if their training provided a broader set of
scientific skills and the flexibility to apply their knowledge to a wider range of scientific
challenges.
Final Report 80
Appendix E Biomedical Workforce Working Group Report
This memo provides the assessment of the Biomedical Modeling Subcommittee (BMW-MSC – see
Appendix A for roster) regarding the current market conditions of the biomedical research workforce
and provides recommendations for the BMW working group to consider.
The BMW-MSC met in person twice (on August 5, 2011 and March 6, 2012) and held a number of phone
meetings and email exchanges to discuss training and career outcomes for biomedical PhDs. These
discussions led to the approach taken to provide the BMW working group with data on the biomedical
research workforce. This approach included:
• Performing a number of descriptive analyses of key aspects of the workforce
• Beginning work on a model to serve as a conceptual framework to organize the analyses, which
could be developed into a full dynamic model
• Populating the conceptual framework with information on each career stage and transition
• Linking to data on each career stage and transition to build a resource upon which
recommendations can be based.
In addition, the BMW-MSC drafted a description of some unintended incentives and disincentives
underlying the US biomedical research system. This document was discussed at the BMW working
group meeting on October 25, 2011.
In general, the market for biomedical PhDs is soft in the period after attaining the PhD or, more
typically, undergoing postdoctoral training. The subsequent career path is marked by uncertainty. It is
also appears to be uneven, with trainees having some skill sets (e.g. quantitative reasoning) doing well
and others much less well. A long time is spent in training; the age of becoming an independent, NIH-
funded researcher exceeds 40. In 2008 only 21% of basic biomedical PhDs within 6-10 years of their
degree had tenured or tenure track academic jobs compared to 28% in 1993. Pay is relatively low
compared to other disciplines such as engineering and the physical sciences and the NIH funding
environment is highly uncertain.
The weakness of the biomedical labor market has been discussed for decades. The current situation is
not merely the result of current economic conditions, but the recent recession surely compounds the
current mismatch. The long-standing weakness is due to an imbalance between the supply of freshly
trained talent and the limited number of permanent post-training positions. The excessive supply stems
from an enterprise that is both labor-intensive and located in academia with its nearly endless supply of
students and postdocs (from domestic and foreign sources). Students and postdocs are highly
motivated and productive and because of the implied training aspects they are willing to work for
reduced wages for extended periods of time. Individuals can wait 10 to 15 years for a more permanent
job with higher rates of remuneration. The problem is exaggerated by a system that is skewed toward
established researchers who themselves have strong incentives to maximize their productivity at the
lowest cost. The subcommittee offers evidence to support this description of the system and offers
Final Report 81
Appendix E Biomedical Workforce Working Group Report
recommendations designed to lessen the dependence on trainees as a source of labor, improve their
training, enhance retention, and thus make careers in biomedical research more appealing.
Following are specific points supporting these conclusions. Also note that the data are incomplete but
we have tried to list the caveats in each case.
Education and Postdoctoral training
Relative to other disciplines, there has been a significant increase in the number of people receiving
basic biomedical PhDs over the past two decades. This increase is correlated with the doubling of the
NIH budget and is driven by an expansion in women and temporary residents obtaining degrees. The
median age at which these basic biomedical PhDs are awarded peaked in the 1990s and fell somewhat
in the most recent decade. In addition, the age of achieving a first tenure track job peaked at 38 years
and trended lower in the most recent decade.
The average duration of postdoctoral training in biomedical research is reported as approximately 4
years. That being said, however, the interval between the age at which the PhD is awarded and that of
the first tenure track job suggests that those postdocs taking academic jobs are spending longer -
approximately 6 years - in postdoctoral training. Doing postdoctoral training does increase the
probability of having a research career in academia, government or industry. However, those who start
a postdoctoral position have approximately 12% lower earnings if they work outside of tenure track
academia compared to those that start in those sectors.
In conclusion, the length of time people spend in training and temporary positions in biomedical
research is generally longer than in other disciplines, the age of starting an independent career is late,
and postdoctoral training is associated with significantly lower lifetime earnings.
Demographics
The average age of receiving the first NIH R01 award for PhDs has been close to 42 since 2000. The age
is higher for researchers with an MD or an MD/PhD. The percent of PIs over 66 years old that receive
NIH funding has increased significantly in the past decade. The share of workers ages 35-50 in basic
biomedical workforce was less than 50% in 2008.
Approximately 50% of PhDs in life sciences are awarded to women. However, women are significantly
less likely to take tenure track jobs in the life sciences. There is a substantial loss of underrepresented
minorities (URMs 64). About 17% of life science bachelor’s degrees are awarded to URMs (2004), but
URMs received only 10% of life science PhDs in 2009.
In conclusion, lengthy postdocs and a lack of academic jobs make biomedicine an unattractive career for
women and URMs. Older researchers perform well in the NIH funding process, raising the hurdle for
younger investigators. NIH has policies in place aimed at mitigating this.
Industry Careers
Almost 70% of biomedical PhDs initiate postdoctoral training after they graduate. By 6-10 years after
graduating, nearly 33% are employed in industry. Industry employment increased almost 5 percentage
points between 1993 and 2008.
The Pharmaceutical industry increased R&D employment significantly through 2008, however it remains
to be seen whether that trend has continued.
64
The following racial and ethnic groups have been shown to be underrepresented in biomedical research: African
Americans, Hispanic Americas, Native Americans, Alaskan Natives, Hawaiian Natives, and natives of the US
Pacific Islands.
Final Report 82
Appendix E Biomedical Workforce Working Group Report
Academic Careers
Since 1973 the share of tenured and tenure-track academic positions in academia has declined while the
share of non-track academic jobs has increased. In 2008 only 23% of basic biomedical positions were
tenured or on the tenure track with the majority (15%) employed at PhD granting institutions. However
a roughly equal numbers (22% ) of biomedical positions were non-tenure track positions.
Staff scientist jobs, which are defined as non-tenure track jobs where individuals primarily do research
and do not teach, have grown significantly in medical schools and universities. In general, these jobs pay
poorly and have few salary guarantees.
In conclusion, there is a shortage of tenure-track research positions relative to the number of people
who are pursuing postdoctoral training after receiving a PhD in basic biomedical research. Staff scientist
positions are temporary and poorly paid.
Career Characteristics
In 1997 70% of basic biomedical PhDs worked in an occupation related to their degree. By 2008 this
number had dropped to 60%. In addition, people with biomedical PhDs at all stages of careers are less
likely to do research in 2008 than in 1993.
Salaries of people with biomedical PhDs are low relative to those with PhDs in other areas or those with
professional degrees. Over 70% of people with basic biomedical PhDs report working more than 40
hours per week.
The returns to experience, the rate at which earnings increase with experience, are much higher for
biomedical PhDs, partially reflecting the low pay associated with the postdoc.
In conclusion, biomedical research jobs are relatively low-paying, especially at the outset of the career
and early-career positions offer few opportunities for independent research.
Overall summary
Careers in biomedical research are unattractive relative to other research doctorate sectors and high-
skilled occupations, despite the fact that the early training stages provide a substantial level of support
to many people. Some contributing factors are:
• It is difficult to launch an independent career.
• There are relatively few opportunities outside of biomedical occupations and industries
• The pay is low when compared to other fields such as engineering and the physical sciences.
The funding of the major supporter of the enterprise, i.e. NIH, is uncertain and favors
established, white investigators
Note: The BMW-MSC has found that there is a substantial amount of missing and/or poor quality data 65.
Therefore, any recommendation will be based on incomplete information. In addition, all
recommendations will involve significant tradeoffs and possible unintended consequences.
65
Estimates of postdoctoral fellows are educated guesses; in particular data on people trained outside of the US
are limited. There is incomplete information on academic positions, in particular data on the sources of support of
academic biomedical researchers are lacking. The data on non-PhD MDs doing research are limited. Finally, there
is limited information on demand.
Final Report 83
Appendix E Biomedical Workforce Working Group Report
Following are policy recommendations drafted by the BMW-MSC, based on the subcommittee’s
discussions as well as ideas drawn from the BMW working group discussions. The recommendation list
is followed by a detailed description of each one, including the rationale behind it, the projected impact,
and options for implementation.
Detailed Description
Final Report 84
Appendix E Biomedical Workforce Working Group Report
Increasing the availability and quality of such opportunities to capable postdoctoral researchers
will increase the attractiveness of biomedical careers for the best undergraduate and graduate
students. This change should be accompanied by an initiative (below) to provide systematic
information to graduate students and postdoctoral researchers about outcomes of people
attending graduate programs and completing postdoctoral training at specific institutions.
Implementation: Gradually reduce support for graduate students and postdoctoral researchers
on research grants and increase support through fellowships and career awards, perhaps aiming
for a 50/50 allocation. Options would be to expand the number of NRSA individual postdoctoral
fellowships or expand the use of the K99-R00 (kangaroo) award.
b. Encourage NIH to investigate ways to use the application and review process to improve the
training experience of those trainees who continue to be supported by RPGs
Background: Many biomedical researchers experience multiple, long postdoctoral training
periods. Even if resources are shifted to fellowships from training grants, many will continue to
be supported by RPGs. The application and review process for RPGs do not place weight
explicitly on the training aspect of these postdoctoral researchers. Thus there is little incentive
to provide a successful trainee experience with high quality outcomes.
Impact: Requiring RPG applications to list the outcomes of recent trainees and place weight on
these outcomes in the review process would provide strong incentives for researchers to
improve placements for their trainees. In addition, requiring information on formal structures
such as independent development plans (IDPs) and requirements for training in responsible
conduct of research in annual progress reports may also provide incentives for mentors.
Implementation: RPG applicants would be required to list all graduate students, postdoctoral
researchers, and other trainees in the past 10 years along with their positions and achievements
(publications, patents, etc.) to date. NIH is encouraged to investigate ways to incorporate
consideration of this information into the review process in a way that does not penalize
placement in non-academic positions that utilize research training. It is possible that NIH could
collect this information and provide it to the reviewers instead of requiring each applicant to
compile the data. In addition, RPG annual progress reports would include information on IDPs
and the training of the students and postdoctoral researchers supported on the grant.
c. Address Career Amenity Issues
Background: There are concerns about the fact that many postdoctoral researchers spend long
periods in temporary positions with relatively low pay and often few benefits such as health
care and access to retirement plans. In addition, attrition from the biomedical research
enterprise is much higher for women than for men. This attrition is due in no small part to
concerns about work-life balance at the early stages of many biomedical research jobs, an issue
that affects both women and men. The combination of poor career amenities and the lack of
work/life balance may also make careers in biomedical research less attractive and thus may be
a factor in the fact that a very low proportion of undergraduate URMs choose this career path.
Impact: Addressing these career amenity issues may address many issues with the current
biomedical research framework. First, it can ensure a more diverse biomedical research
workforce which may improve innovative outcomes and ensure that the issues facing women
are addressed by the biomedical research enterprise. Second, addressing these issues may make
biomedical research careers more attractive to URMs. In fact the two previous benefits will
become more important as the United States population becomes increasingly diverse,
Final Report 85
Appendix E Biomedical Workforce Working Group Report
from institutions that do not provide hard money support to those that do provide it has the
potential to increase total resources devoted to biomedical research. If such a measure results
in a net reduction in translational research and diminished focus on real biomedical problems, it
would obviously be undesirable. However, the net effect could actually be an increase in the
capacity for research, including translational research, since it would lead to a reduction in time
devoted to generating proposals.
Implementation: Given the wide range of practices within and between institutions, any hard
money requirements have the potential to be disruptive, especially for established researchers.
We therefore recommend that changes be implemented very gradually even relative to our
other recommendations and that multiple approaches be considered.
One approach would be to require hard money appointments. To provide flexibility within
institutions, it is worth separating individual and institutional levels of hard money support. A
phased approach to introduce a hard money requirement might 66: start in 2015, with a
minimum average institutional hard money support level imposed at 2.5%, to be increased by
2.5% per year until 2031 when it reaches 40%. Additionally, starting in 2015, a minimum
individual hard money support level might be imposed at 2% to be increased by 2% per year
until 2030 when it reaches 30%. No acceleration of such a progression should be implemented
without three years advanced notice.
An alternative approach would be to impose a hard money requirement only for newly hired
researchers. Startup packages might be counted toward a hard money requirement for the first
5 years of a contract, if necessary.
Reductions in the maximum level of salary that NIH can support provide a third mechanism to
reduce reliance on soft money, especially for more senior researchers. The recent reduction (in
fiscal year 2012 the level is set at $20,000 less than the previous year) may provide information
about how implementing this recommendation would affect different types of institutions.
Finally, one could tailor limits on soft money contributions for each institution by taking into
account existing institutional costs and revenue streams by building faculty salary rate caps into
typical indirect cost negotiations. Those negotiations are periodically required for all
institutions that participate in federal grant programs and they already include factors that
could help determine reasonable federal contributions to faculty salaries.
b. Shift the balance of NIH resources among researchers at various career stages and at different
funding levels
Background: The average age of biomedical researchers and the share of older biomedical
researchers have both increased, in part because of the elimination of mandatory retirement. At
the same time the average age at which researchers receive their first R01 and tenure-track jobs
have both increased. Although research shows that older researchers can be highly innovative,
the aging of the biomedical research workforce ties up positions and funding that could assist
young and mid-career researchers.
In addition, the distribution of resources at NIH is highly unequal. In general, 20% of the
Principal Investigators (PIs) receive 50% of RPG funding. While it is socially optimal to allocate
66
There was a diversity of opinion among the BMW-MSC members as to the appropriate numbers. Note that the
proposal Bruce Alberts published in Science in September 2010 was more aggressive than that suggested here,
suggesting a goal of 50% in 10 years.
Final Report 87
Appendix E Biomedical Workforce Working Group Report
more resources to the most productive researchers (who produce some of the most innovative
and important research and best trainees), NIH allocations should equalize the research output
produced by an additional investment in each researcher. Also, highly-funded researchers are
likely to rely on research and administrative staff more heavily than less-funded researchers.
Impact: Gradual measures to redistribute some resources from late-stage to early- and mid-
career researchers would complement other recommendations in this document designed to
increase the attractiveness of the biomedical research career. Insofar as senior researchers
have higher earnings and rely more heavily on research staff, this shift will also reduce costs and
the need for research staff.
In addition, a moderate equalization of resources to less-well funded researchers would
generate more positions for independent researchers both directly and by reducing reliance on
research and administrative staff. Lower salaries among less-well funded researchers increase
the impact of such equalization.
Implementation: NIH already has first-time investigator and early-career investigator initiatives.
These appear to have at least reversed the previous declining trend of proportional support for
more junior researchers. It will be important to evaluate the effect of these programs both on
new investigators and on the resulting mid-career investigators as time progresses. NIH also
might consider funding investigators at specific career stages for a specific period of time (e.g. 7
years) instead of funding specific projects, similar to the HHMI model.
Initiatives to review funding levels of investigators have been utilized or proposed by NIH. For
example, NIGMS has a policy of additional scrutiny of awards to any principal investigator with
existing funding of $750,000. The review is conducted by the NIGMS Advisory Council. In
addition, in the FY 2013 President’s Budget, NIH proposed to establish a similar process with a
threshold of $1.5 million or more in NIH grants. It may make sense to institute a threshold level
that evolves over time, adjusting as the cost of research increases.
c. Gradually reduce the dependence on trainees to staff labs and develop a more permanent
staffing model
Background: The biomedical research enterprise requires a large corps of bench researchers.
Currently this corps is largely populated by students and postdoctoral researchers, many of
whom remain in training positions well beyond the period necessary for effective training. Most
people feel that the duration of these experiences largely reflects an inadequate number of
good opportunities for positions as independent researchers.
Impact: Creating a cadre of less temporary scientists would reduce the need to staff labs with
graduate students and particularly postdoctoral researchers, allowing for clearer career paths.
The cost implications of such scientists are unclear. On the one hand, such scientists likely will
be paid more than postdoctoral researchers and would receive the normal employee benefits.
On the other hand, staff scientists may be more experienced and productive than postdoctoral
researchers and especially graduate students. Moreover, staffing labs with staff scientists may
reduce training costs borne by NIH, institutions, and trainees if it reduces the number of
students admitted to graduate school. These gains can be multiplied if the staff scientist position
appeals to people who might otherwise leave the biomedical research enterprise. Such
positions also will provide a better work-life balance and shorter lags between degree
completion and a reasonably compensated career with benefits, all of which will likely improve
retention.
Final Report 88
Appendix E Biomedical Workforce Working Group Report
Implementation: We note that permanent scientist positions currently are most viable in large
labs. However, if funds are available, such scientists could be very effective in a small lab. The
NIH could consider programs that would offset a portion of the costs of such employees to
encourage the development of such scientist programs. It is possible that the marginal costs for
such positions will be relatively small in comparison to postdoctoral positions and it is possible
that such a program could have a limited duration. One possibility could be to allow those
marginal costs and potentially transition funds for a scientist position to be built into research
grant budgets. Thus if grant funding runs out on one grant there will be funds set aside to
continue to pay the scientist for a short period.
d. Push to stabilize science funding
Background: Funding for NIH has grown rapidly, but in fits and starts. This episodic growth
generates large fluctuations in entry (driven at least in part by staffing needs) and in
construction of research facilities. Given that both training and construction are long, slow
processes, the fluctuations in funding mean that when trainees and facilities are ready, the
initial demand spurt has frequently passed.
Impact: A more stable funding stream would moderate boom-and-bust cycles that
disadvantage trainees and destabilize institutions by encouraging building or remodeling of
research facilities leading to increased pressure to generate proposals as funding growth
declines. The cost in lost stimulus support would be made up by greater stability.
Implementation: The BMW-MSC recommends that NIH and the research community should
push to have funding set as a percentage of GDP over the past 5 years.
Final Report 89
Appendix E Biomedical Workforce Working Group Report
These data should be made available to the public in uniform, aggregated (and de-identified)
form as well as provided to NIH for analysis.
These data, when combined with the data infrastructure recommended below, will allow NIH to
generate statistical summaries of outcomes for trainees overall and for particular programs.
Understanding this kind of information will permit the NIH to more easily identify training
programs that work and to properly adjust its support for graduate and postdoctoral training,
and also enable institutions to more effectively determine the optimal size of their biomedical
training programs.
b. Establish an NIH-hosted job posting clearinghouse
Background: There is no single source of information on job openings for biomedical researchers
as there are for researchers in many other fields.
Impact: A centralized job website would facilitate job searches and provide NIH and other policy
makers with concrete, up-to-date information on labor market demand.
Implementation: As part of the analysis of the biomedical workforce the subcommittee is
obtaining data from WANTED Technologies that collates job posting information from online
resources including Science Careers and Chronicle Careers. NIH could find a way to collate such
data on an ongoing basis. Alternatively, NIH could cross-list jobs from select vendors such as
Science Careers. These data should be summarized on an annual basis to provide information
about the types and locations of jobs.
c. Develop and securely distribute data infrastructure on the biomedical research enterprise
Background: Much of the data necessary for rigorous research on the biomedical research
enterprise exist, but they are not integrated and are difficult for the research community to
access. Additional data could be harvested from grant applications and/or required in reporting.
Other data are either lacking or insufficiently reliable to support a comprehensive analysis of the
enterprise. The lack of data limits our understanding of the biomedical research enterprise. It
also hampers endeavors like that of the present Biomedical Research Workforce Taskforce,
meaning that the modeling subcommittee is forced to quickly draw together a patchwork of
data from multiple sources.
Impact: Developing and securely distributing large-scale data infrastructure on the biomedical
research workforce will not only assist future efforts such as the present one but will also
support research on biomedical research more broadly that can inform NIH policy.
Implementation: NIH should support the large-scale combination of its internal data (from grant
applications, including biosketches) and their linkage to publicly available data and make these
data available to the research community securely. One model is the nascent policy on access to
internal NIH data for research purposes, which should be supported. Valuable components
include data on: postdoctoral fellows; people trained outside of the US; numbers and support in
academic positions; demand, especially outside of academia; and non-PhD MDs.
d. Support ongoing research on the biomedical research workforce and enterprise
Background: As a research agency, NIH routinely seeks evidence-based information to optimize
and quantify its performance including, but not limited to, periodic studies of the biomedical
research workforce, economic impact analyses, and evaluations of programs and initiatives.
These efforts typically operate on a tight timeframe and in the face of the data fragmentation
Final Report 90
Appendix E Biomedical Workforce Working Group Report
described above, limiting the analyses that are possible and greatly increasing the cost of such
studies.
At the same time, the science of science and innovation is an increasingly dynamic field,
meaning that the extramural research community has an ever-expanding ability to provide
rigorous, research-backed answers to NIH’s pressing policy questions. Unfortunately, NIH does
not have a good mechanism for supporting such research – proposals submitted to standing
study sections at the various institutes, for instance, are often viewed as not directly relevant to
the interests of that institute even if they viewed as important for NIH as a whole.
Impact: An ongoing effort would ensure that future efforts such as the present one would have
in place state-of-the -art infrastructure for more thorough analyses and simulations. It could
assist the periodic outside assessments of the biomedical workforce mandated in current law,
and would also be able to address the broader range of policy questions that NIH seeks to
answer. It could also produce preliminary or “flash data” indicators of current conditions in the
biomedical workforce, along the lines of such economic indicators routinely released by the BLS.
The data obtained through such an ongoing effort also will be useful for implementing the
recommendation to provide systematic information to people contemplating careers in
biomedical research.
Implementation: NIH should provide a continuing flow of resources for research on the
biomedical research enterprise as well as access to the necessary data (described above). This
research could be guided by an entity in the office of the NIH Director, combining NIH staff with
a small number of extramural advisors with expertise in labor economics supplemented by
members of the biomedical research community who would work together to identify and
implement projects that address NIHs policy needs using cutting-edge economic methods. One
model for such an effort could be the President’s Council of Economic Advisors.
Final Report 91
Appendix F Biomedical Workforce Working Group Report
• Cathee Johnson Phillips, M.A. - Executive Director and Zoe Fonseca-Kelly, Ph.D. - Chair, Board of
Directors, National Postdoctoral Association
• Roger Chalkley, D.Phil, Chair, NRC Study on Research Training and the Biomedical, Behavioral,
and Clinical Research Sciences, Senior Associate Dean for Biomedical Research Education and
Training, Vanderbilt University Medical Center
• Phillip Pizzo, M.D., Dean, Stanford University School of Medicine
• Lawrence Brass, M.D. Ph.D., Professor of Medicine, Associate Dean, Combined Degree and
Physician Scholars Program, Director, Penn MSTP, University of Pennsylvania
• Peter Henderson, Ph.D., Director, NRC Report on Expanding Underrepresented Minority
Participation, Director, National Research Council’s Board on Higher Education and Workforce
• Ann Bonham, Ph.D., Chief Scientific Officer, Association of American Medical Colleges
• Susan Amara, Ph.D., Detre Professor and Chair, Dept. of Neurobiology, University of Pittsburgh
School of Medicine, and President, Society for Neuroscience
• Howard Garrison, Ph.D., Director of the Office of Public Affairs, Federation of American Societies
for Experimental Biology
• Jennifer Poulakidas, Vice President, Congressional and Governmental Affairs, Association of
Public and Land-grant Universities
• Carrie D. Wolinetz, Ph.D., Associate Vice President for Federal Relations, Association of
American Universities
Final Report 92
Appendix G Biomedical Workforce Working Group Report
January, 2012
This report covers the findings and conclusions of the comment analysis on the NIH Request for Information on
the future of the biomedical workforce. The analysis was done by Ripple Effect Communications Inc. under
contract # HHSN276200800275U.
Appendix G Biomedical Workforce Working Group Report |
Executive Summary
This report provides a summary of the comments received in response to the Request for Information (RFI):
“Input into the Deliberations of the Advisory Committee to the NIH Director Working Group on the Future
Biomedical Research Workforce” (NIH Guide Notice NOT-OD-11-106).
The RFI provided a list of eight issues that had been identified as important to consider when developing a
model of the future biomedical research workforce. Information was requested in response to three questions
related to the eight issues (or other unidentified issues).
The comments received from 219 commenters were parsed into 498 “quotations” representing unique ideas,
with an average of 2.3 quotations per commenter. Those quotations were key-word coded for sorting purposes.
Only 20% of the commenters replied on behalf of an organization, while 75% of the commenters provided
personal input; the remaining 5% of the commenters were NIH staff. The organizations represented in the 20%
were a broad cross section of NIH stakeholders, including NIH-funded investigators and research institutions.
Feedback was received on 1) how identified and unidentified issues affect institutions, scientists, or both; 2)
what issue(s) are most important for the working group to address and why; and 3) how these issues should
affect NIH policies or processes. The feedback was categorized into 12 primary issues, with 7 overlapping issues.
Primary Issues
The analysis process identified four primary issues in addition to the original eight primary issues included in the
RFI. The distribution of primary issues, as cited by commenters, is shown in the graph.
Final Report 94
Appendix G Biomedical Workforce Working Group Report |
In cases where the primary and secondary issues are similar, the secondary issue covers only certain aspects of
the primary issue. For example, the secondary issue of Career Appeal covers the specific issue of working
conditions; whereas the primary issue of Biomedical Research Career Appeal encompasses all issues related to
the attractiveness of biomedical research careers (e.g. salary, availability of research funding, working
conditions).
Funding. Uncertainty and lack of funding, distribution of funding, restricted paylines, success rates, and
excessive competition
Multi-disciplinary. Need for multi/ inter/ trans-disciplinary research training to prepare individuals for a
wide range of academic and non-academic career opportunities
Salary. Inadequate compensation and benefits
Length of Training. Amount of training time too long to be feasible for majority
Non-US Citizens. Foreign students and post-doctoral fellows
Career appeal. Working conditions (e.g. heavy workload, perception of being perceived as cheap labor,
long work hours)
Mentoring. Quality of career development and the need for pre-college preparation
Diversity. Under-represented minority post-doctoral, fellows and junior faculty
The comments received are summarized within this report in a variety of ways to provide multiple options for
the NIH ACD Working Group to review and utilize the information in their recommendations.
Final Report 95
Appendix G Biomedical Workforce Working Group Report |
Contents
EXECUTIVE SUMMARY ................................................................................................................................ 94
Primary Issues 94
Overlapping (Secondary) Issues 94
BACKGROUND ............................................................................................................................................. 98
Methodology 99
About the Data 99
Coding Scheme 100
Analysis Process 101
Affected Parties 101
Priority Analysis 102
ISSUE PRIORITY ......................................................................................................................................... 103
QUALITATIVE ANALYSIS ........................................................................................................................... 104
Supply and Demand, 97 quotations [19%] 104
PhD Characteristics, 84 quotations [17%] 105
Post-doctoral Fellow Training Characteristics, 62 quotations [12%] 107
Biomedical Research Career Appeal, 53 quotations [11%] 109
Clinician Characteristics, 44 quotations [9%] 110
Staff Scientist Career Track, 36 quotations [7%] 111
Diversity, 32 quotations [6%] 112
Effects of NIH Policies, 29 quotations [6%] 113
Mentoring, 24 quotations [5%] 114
Training to Research Grant Ratio, 19 quotations [4%] 116
Early Educational Interventions, 11 quotations [2%] 117
Industry Partnership, 7 quotations [1%] 118
APPENDIX .................................................................................................................................................. 120
Additional Data 120
Primary Issues and Descriptions 124
Secondary Issues and Descriptions 125
Final Report 96
Appendix G Biomedical Workforce Working Group Report |
Final Report 97
Appendix G Biomedical Workforce Working Group Report |
Background
In April 2011, the National Institutes of Health (NIH) Advisory Committee to the Director (ACD) formed a
Working Group to examine issues related to the future of the biomedical research workforce in the United
States and make recommendations to the ACD that would help promote a diverse and sustainable biomedical
and behavioral research workforce. As part of the process, the Working Group was tasked with gathering input
from the extramural community, including students, post-doctoral fellows, investigators, scientific societies, and
grantee institutions to consider various aspects of the future workforce.
The Working Group identified eight (8) issues to consider in the development of the future biomedical research
workforce model:
The balance between supply, including the number of domestic and foreign trained PhDs and post-
doctoral fellows, and demand, i.e. post-training career opportunities.
Characteristics of PhD training in biomedical research, including issues such as
o The length of the PhD training period.
o Recommendations for changes to the PhD curriculum.
o Training for multiple career paths (including bench and non-bench science).
Characteristics of clinician-research training including issues such as
o The balance between MDs and MD/PhDs
o Career development of clinician-researchers.
o Recommendations for changes to the curricula for training clinician-researchers.
Length of Post-doctoral training.
The ratio of PhD students and post-doctoral fellows on training grants to those supported by research
grants.
Possibilities for professional/staff scientist positions and the level of training required for such positions
(e.g. PhD or MSc degrees).
Issues related to the attractiveness of biomedical research careers (e.g. salary, working conditions,
availability of research funding)
The effect of changes in NIH policies on investigators, grantee institutions and the broader research
enterprise.
NIH issued a Request for Information (RFI) to the community to provide input into the deliberations of the ACD
Working Group. From August 17 through October 7, 2011 the extramural community submitted input to NIH on
the identified issues (and other unidentified issues), the importance and effects of these issues on institutions
and scientists, and how the issues should affect NIH policies and procedures.
Final Report 98
Appendix G Biomedical Workforce Working Group Report |
Methodology
About the Data
The primary type of submission was via an online form, but comments also were received via e-mail and postal
mail. Responses from three commenters were received more than once; these duplicative comments were only
analyzed once. There were a total of 219 commenters. Comments from the 219 commenters were parsed into
498 quotations, which correspond to an average of 2.3 quotations per commenter. Commenters responding on
behalf of organizations provided an average of 3.4 quotations each, while commenters responding on behalf of
themselves provided an average of 2.0 quotations each.
Each of the 498 quotations corresponded to one primary issue, according to the following distribution.
Final Report 99
Appendix G Biomedical Workforce Working Group Report |
Coding Scheme
The coding scheme evolved from the bottom up, by utilizing the eight issues identified in the RFI, and analyzing a
sample of the responses to generate the scheme in an iterative fashion. Through this process, we identified four
additional issues that were suggested by commenters. This bottom up approach was consistent with the key
aspect of the RFI design which stated that all ideas and suggestions were welcome. The final issue categories
and their descriptions are available in the Appendix).
The following is a list of the 12 primary issues (including 4 new* issues):
1. Supply and Demand
2. PhD Characteristics
3. Post-doc Training Characteristics
4. Biomedical Research Career Appeal
5. Clinician Characteristics
6. Diversity*
7. Staff Scientist Career Track
8. Mentoring*
9. Effects of NIH Policies
10. Training to Research Grant Ratio
11. Industry Partnership*
12. Early Educational Interventions*
Note that the issue “Post-doc Training Characteristics” is not identified as a new issue; however, it was
broadened from the original primary issue, “ Length of Post-doctoral Training,” to be more inclusive of all the
aspects of post-doctoral training that were identified by commenters.
During the coding process, we discovered that specific aspects of primary issues were appearing across all
comments. To capture these “secondary” themes, overlapping issues were developed. For example, a
comment on the primary issue of Supply and Demand may cite other interlocking issues such as Career Appeal
or Funding.
The following is a list of the secondary issues and their descriptions:
Funding. Uncertainty and lack of funding, distribution of funding, restricted paylines, success rates,
indirect costs, excessive competition
Multi-disciplinary. Need for multi/ inter/ trans-disciplinary research training to prepare trainees for a
wide range of academic and non-academic career opportunities
Length of training. Amount of training time too long to be feasible for majority
Career appeal. Working conditions, i.e. heavy workload, perception of being used as cheap labor, long
work hours.
Mentoring. Quality of career development and the need for pre-college preparation
Analysis Process
Both the primary and secondary issues were used as a starting point and expanded through successive
analysis/coding iterations by a coding team. The following process was followed:
1) First, a random selection of comments was assigned to each team member to scan for meaningful
quotations that addressed issues related to the future of the biomedical workforce;
2) Team members analyzed the quotations to define the code categories (starting with the three questions
identified by the RFI and the 8 issues identified by ACD) into which the quotation might belong (first
cut);
3) Team members assigned the quotations into one of the existing issue code categories that best matched
the quotation (second cut);
4) Finally, team members revised the code structure by creating new issues to categorize the quotations
that did not fit existing issue categories.
Affected Parties
Each quotation was analyzed for affected party (investigators, institutions, or both), as cited by the commenter.
Of the 498 quotations, only 319 (64%) were identified with an affected party. The distribution of affected party
(for the overall data set and by affiliation) is shown in the graph below.
Priority Analysis
Commenters were asked to indicate the most important issue(s) for the working group to address. To capture
and analyze responses to this question, cited issues for each commenter were given a ranking number 67. For
example, if a commenter mentioned three issues in her response, the first issue received rank 1, the second
issue received rank 2, and the third issue received rank 3. The total count per issue was summed across all
comments to determine the overall priority. This method allowed us to distribute the appropriate weight per
issue, when commenters mentioned more than one issue. The theoretical maximum priority score for an issue
was 219, and would have occurred if the issue received rank 1 from every commenter. Priority scores for each of
the twelve issues, in descending order, are provided in the table below.
As expected, the overall priority of issues followed a similar pattern to the frequency counts by issue.
67
If issue priority was not explicitly stated by the commenter, it was assigned by the order in which the issue appeared
within each comment.
Final Report 102
Appendix G Biomedical Workforce Working Group Report |
Issue Priority
Commenters were asked to indicate the most important issue for the working group to address. To capture and
analyze responses to this question, cited issues were given a ranking number 68 by commenter. The total count
per issue was summed across all responses to determine overall priority.
As described above, the overall issue priority was similar to the overall frequency of issues. However, when
considered by affiliation (self and organization), the pattern differed, as shown below.
Issue Priority Score
Issue Priority Score
(Self)
(Organization)
Supply and Demand 68
PhD Characteristics 29
PhD Characteristics 55
Clinician Characteristics 21
Biomedical Research Career Appeal 47
Post-doc Training Characteristics 19
Post-doc Training Characteristics 36
Supply and Demand 18
Clinician Characteristics 30
Biomedical Research Career Appeal 18
Effects of NIH Policies 21
Staff Scientist Career Track 15
Diversity 20
Diversity 12
Mentoring 18
Effects of NIH Policies 12
Staff Scientist Career Track 13
Training to Research Grant Ratio 12
Industry Partnership 6
Early Educational Interventions 2
Early Educational Interventions 5
Mentoring 1
Training to Research Grant Ratio 3
Industry Partnership 1
The order of issues as determined by frequency (below) is the same whether considering either the Organization
or the Self perspective, but it does not correlate with the order of issues as determined by priority score for
either Organization or Self.
Issue Frequency (Organization) Frequency (Self)
Supply and Demand 17% 20%
PhD Characteristics 15% 18%
Post-doc Training Characteristics 11% 13%
Biomedical Research Career Appeal 10% 10%
Clinician Characteristics 9% 9%
Staff Scientist Career Track 8% 6%
Diversity 7% 6%
Effects of NIH Policies 7% 5%
Mentoring 5% 5%
Training to Research Grant Ratio 5% 3%
Early Educational Interventions 3% 2%
Industry Partnerships 1% 2%
68
Priority was assigned based on the order of issue appearance within each separate comment response.
Final Report 103
Appendix G Biomedical Workforce Working Group Report |
Qualitative Analysis
The issues identified by the Working Group and commenters are discussed below in descending order
by comment frequency, beginning with the issue that received the most comments, Supply and
Demand.
SECONDARY ISSUES
Most (64%) of the quotations identified a secondary issue. Those with a secondary issue were
categorized as follows:
Secondary Issues for Supply and Demand
Non-US Citizens 35%
Funding 32%
Career Appeal 11%
Salary 10%
Multi-disciplinary 6%
Length of training 3%
Mentoring 2%
IMPACT
Almost half (43%) of the commenters felt that Supply and Demand was an issue for both institutions and
investigators, while 34% felt that Supply and Demand was an investigator issue and 23% felt that Supply
and Demand was an institutional issue.
hope that the NIH would “redefine success” for training grant reviews to include non-academic
appointments. Beyond a lack of information regarding non-academic careers, several noted that interest
in such careers may actually be discouraged by faculty.
Length of training period. Only a few commenters suggested that PhD training was too long. Several
warned that imposing a reduction or cap on length of training could have a negative effect on the
quality of doctoral graduates.
SECONDARY ISSUES
A strong majority (81%) of the quotations identified a secondary issue. Those with a secondary issue
were categorized as follows:
Secondary Issues for PhD Characteristics
Multi-disciplinary 55%
Length of training 16%
Funding 14%
Salary 4%
Mentoring 4%
Career appeal 3%
Non-US Citizens 1%
Diversity 1%
IMPACT
Almost half (46%) of the commenters felt that PhD Characteristics was an investigator issue; slightly less
(42%) felt that it was an issue for both institutions and investigators, and only 12% felt that PhD
Characteristics was an institutional issue.
Content of training. Some respondents called for more structure in the post-doctoral training
experience. There was no consensus on what training ought to include or not include, but many
expressed that post-doctoral fellows do not receive enough experience in non-research skills, such as
teaching and grant writing. Many believed that a structured competencies-based approach, supported
by NIH, would be preferable to a strict limitation on number of years.
Career development and mentoring. Most commenters seemed to agree that post-doctoral fellows do
not engage in enough career development. This was especially true for post-doctoral fellows with non-
academic or non-research career interests. A few commenters stressed the need for an increase in the
training areas of lab management and teaching to create more self-sufficient researchers. Many others
pointed out the mentoring challenges that PIs face in the current funding climate. With more time
needed to write winning grants, PIs appeared to be neglecting their mentoring duties of new scientists.
Overall, individual respondents felt that mentoring and career development was generally inadequate; a
majority of institutions did not comment on this aspect of post-doctoral training.
Lifestyle of a post-doctoral fellow. Several commenters explained that the lifestyle of a post-doctoral
fellow (hours worked, workload, and relationships with PIs) is not amenable to family life. Thus, many
women of child-bearing age chose to place their career on hold at this stage or leave academia
altogether. This issue appeared more often in comments from individuals than institutions.
SECONDARY ISSUES
A strong majority (82%) of the quotations identified a secondary issue. Those with a secondary issue
were categorized as follows:
Secondary Issues for Post-doctoral Training Characteristics
Length of training 27%
Funding 25%
Salary 20%
Multi-disciplinary 12%
Career appeal 8%
Mentoring 6%
Non-US Citizens 2%
IMPACT
More than half (54%) of the commenters felt that Post-doctoral Training Characteristics was an
investigator issue; about one-third (36%) felt that it was an issue for both institutions and investigators,
and just 10% felt that Post-doctoral Training Characteristics was an institutional issue.
SELECTED PUBLIC RECOMMENDATIONS FOR NIH ACTION
Increase the availability and length of transition funding for senior post-doctoral fellows.
Raise the NRSA post-doctoral stipend and mandate that all NIH-supported post-doctoral fellows
(whether directly or indirectly supported) receive this amount.
Require better documentation and monitoring of training progress and career planning.
SECONDARY ISSUES
More than three-quarters (79%) of the quotations identified a secondary issue. Those with a secondary
issue were categorized as follows:
IMPACT
Almost half (46%) of the commenters felt that Biomedical Research Career Appeal was an issue for both
institutions and investigators, while 33% felt that it was an investigator issue and 21% felt that it was an
institutional issue.
SECONDARY ISSUES
Slightly more than half (57%) of the quotations identified a secondary issue. Those with a secondary
issue were categorized as follows:
Secondary Issues for Clinician Characteristics
Funding 36%
Multi-disciplinary 32%
Salary 16%
Length of training 8%
Career appeal 8%
IMPACT
More than half (53%) of the commenters felt that Clinician Characteristics was an issue for both
institutions and investigators, while 33% felt that it was an investigator issue and 13% felt that it was an
institutional issue.
SECONDARY ISSUES
Only 36% of the Staff Scientist Career Track quotations identified a secondary issue. Those with a
secondary issue were categorized as follows:
Secondary Issues for Staff Scientist Career Track
Funding 46%
Multi-disciplinary 15%
Length of training 15%
Salary 15%
Career appeal 8%
IMPACT
Half (50%) of the commenters felt that Staff Scientist Career Track was an issue for both institutions and
investigators, while 44% felt that it was an investigator issue and only 6% felt that it was an institutional
issue.
these barriers were well-known. Commenters proposed a diverse list of recommendations to resolve
this issue including pre-college mentoring, use of social media tools to mentor, recruitment and tracking
of candidates, and support of diversity specific mechanisms.
SECONDARY ISSUES
Less than half (44%) of the quotations identified a secondary issue. Those with a secondary issue were
categorized as follows:
Secondary Issues for Diversity
Funding 50%
Career appeal 21%
Non-US Citizens 14%
Multi-disciplinary 7%
Length of training 7%
IMPACT
A plurality (40%) of commenters felt that Diversity was an institutional issue; about one third (35%) felt
that Diversity was an investigator issue, and only 25% felt that it was an issue for both institutions and
investigators.
NCRR dissolution. Several commenters submitted similar responses, which expressed concern that
funding and extramural support for animal-model biomedical research would be reduced with the
dissolution of NCRR. These commenters urged the NIH to ensure that NCRR’s commitment to animal-
model research would continue following this reorganization.
Institution affiliation. Currently, scientists must be associated with an institution or hold a certain job
title to apply for certain types of funding at NIH. One commenter felt that this policy was
counterproductive in the current economic climate, especially for unemployed scientists who are trying
to re-enter to the workforce. This policy could also be a burden for post-doctoral applicants who are
unable to get institutional support.
SECONDARY ISSUES
More than half (59%) of the quotations identified a secondary issue. Those with a secondary issue were
categorized as follows:
Secondary Issues for Effects of NIH Policies
Funding 88%
Salary 6%
Career appeal 6%
IMPACT
A majority (61%) of the commenters felt that Effects of NIH Policies was an investigator issue, while 26%
felt that it was an issue for both institutions and investigators, and only 13% felt that it was an
institutional issue.
both the mentor and mentee. The progress of such plans should be included in the annual reports to
NIH for funded trainees.
Conflict interest for mentors. An additional concern for many commenters was the conflict of interest
that mentors experience. Funding systems, such as the NIH, and academic promotion structures reward
discovery and publications, metrics which are largely fueled by student and post-doctoral fellow
productivity. This reliance on student and post-doctoral labor creates an environment where
productivity is prioritized over career development, especially when pursued outside of the lab. Because
mentoring is an uncompensated activity, several commenters suggested that making mentoring a part
of funding reviews might provide the incentive needed to address these conflicts.
Funds for institutional program management staff. Several commenters recommended that training
grants provide support for program management of the training grants, including salary support for
program directors, staff, and significant faculty. Also requested on training grants were funds for
purchasing training technologies.
Non-academic mentoring. Of commenters addressing non-academic career paths, most agreed that
training for these paths is inadequate. Given that most current mentors were “raised” on the academic,
tenure-track path, commenters expressed an absence of mentoring for non-academic careers, which
has a downstream effect on the supply and demand issue.
SECONDARY ISSUES
More than half (54%) of the quotations identified a secondary issue. Those with a secondary issue were
categorized as follows:
Secondary Issues for Mentoring
Multi-disciplinary 43%
Funding 36%
Length of training 7%
Salary 7%
IMPACT
A strong majority (73%) of the commenters felt that Mentoring was an investigator issue and none (0%)
felt that it was an institutional issue. The remaining 27% felt that Mentoring was an issue for both
institutions and investigators.
SECONDARY ISSUES
More than half (58%) of the quotations identified a secondary issue. Those with a secondary issue were
categorized over a broad group of the following secondary issues:
Secondary Issues for Training to Research Grant Ratio
Funding 42%
Mentoring 17%
Salary 8%
Career appeal 8%
Multi-disciplinary 8%
Non-US Citizens 8%
Diversity 8%
IMPACT
Commenters were equally split (40% each) between those that felt Training to Research Grant Ratio was
an investigator issue versus an issue for both institutions and investigators; only 20% felt that it was an
institutional issue.
SECONDARY ISSUES
Most (64%) of the quotations identified a secondary issue. Those with a secondary issue were
categorized as follows:
Secondary Issues for Early Educational Interventions
Funding 43%
Mentoring 29%
Multi-disciplinary 14%
Length of training 14%
IMPACT
A strong majority (67%) of the commenters felt that Early Educational Interventions was an issue for
both institutions and investigators; the remaining 34% were equally split (17% each) as to whether they
felt Early Educational Interventions was an investigator issue or an institutional issue.
SECONDARY ISSUES
Only two of the quotations (29%) identified a secondary issue. In both cases, the secondary issue was
Mentoring.
IMPACT
Three quarters (75%) of the commenters felt that Industry Partnership was an issue for both institutions
and investigators, while 25% felt that it was an investigator issue; none (0%) classified it as an
institutional issue.
SELECTED PUBLIC RECOMMENDATIONS FOR NIH ACTION:
Set up or encourage partnership agreements between private industry and individual scientists;
partnerships would define focus of academic partner (discovery) and focus of industry
(commercialization).
Adopt a ”net-benefit policy” in which institutions receiving NIH funds must secure a commitment
from private US employers to hire an equal or greater number of scientists than those supported
by the NIH monies.
Promote partnership programs for post-doctoral fellows to provide them with a better
understanding of how science is carried out in industry, which would help prepare them for
career paths outside of academia.
Appendix
Additional Data
Comment Coding Status
Counts
TOTAL Duplicates: 3
COMMENTS 224
Non-
RECEIVED 2
Responsive:
Unique and Responsive Comments Received: 219
Quotations Coded To Date: 498
Mean Quotations Per Submission: 2.3
Submission Method
Count Percent
Web Form: 213 96%
Email: 7 3%
Postal Mail: 1 1%
Fax: 0 0%
Affiliation Catergory
Count Percent
Self (Private Citizen) 165 75%
Organization 44 20%
NIH Staff 10 5%
Issues by Affiliation
0
Institutions Investigators Both
Investigators
Institutions
Both
Issue Description
Biomedical Research Career Issues related to the attractiveness of biomedical research careers (e.g. salary, working conditions,
Appeal availability of research funding)
Effects of NIH Policies The effect of changes in NIH policies on investigators, grantee institutions and the broader research
enterprise.
Industry Partnership Problems related to relationships between academic research and commercial industry research.
Examples are:
• industry use of academic discovery
• difficulty of industry scientists returning to academia
• partnering with industry to train new scientists
Mentoring The need to improve the quality of career development at institutions. Guidelines and monitoring of
mentorships is needed as there is a lack of non-research science skills being taught.
PhD Characteristics Characteristics of PhD training in biomedical research, including issues such as:
• The length of the PhD training period.
• Recommendations for changes to the PhD curriculum.
• Training for multiple career paths (including bench and non-bench science).
Staff Scientist Career Track Possibilities for professional/staff scientist positions and the level of training required for such
positions (e.g. PhD or MSc degrees).
Supply and Demand The balance between supply, including the number of domestic and foreign trained PhDs and post-
doctoral fellows, and demand, i.e. post-training career opportunities.
Training to Research Grant The ratio of PhD students and post-doctoral fellows on training grants to those supported by
Ratio research grants.
Overlapping Descriptions
Funding Uncertainty and lack of funding, distribution of funding, restricted paylines, success rates, indirect
costs, excessive competition
Multi-disciplinary Need for multi/ inter/ trans-disciplinary research training to prepare trainees for a wide range of
academic and non-academic career opportunities
Length of Training Amount of training time too long to be feasible for majority
Career appeal Working conditions (e.g. heavy workload, perception of being used as cheap labor, long work
hours)
Mentoring Quality of career development and the need for pre-college preparation
December 6, 2011
69
The Biomedical Workforce (BMW) Working Group is a subcommittee of the NIH Advisory Committee to the
Director (ACD)
Final Report 127
Appendix H Biomedical Workforce Working Group Report |
Introduction
The NTW Committee 70 explored different scenarios for NIH’s role in the support of training the future
biomedical research workforce (biomedical, unless otherwise noted, includes the biomedical,
behavioral, social and clinical sciences). The scenarios developed in this document are intended to
provide a framework for discussion about how NIH supports research training and possible avenues for
change. The biomedical research workforce comprises not only independent research scientists, but
also a much larger component, the trainees themselves. This obvious point is made to emphasize that
any scenario that results in a reduction in the size of the trainee pool also reduces the size of a vital
component of the workforce, with significant financial consequences.
NIH support of research training and education is focused at the predoctoral, postdoctoral, and early
career stages of a biomedical scientist’s career. NIH also supports diversity-promoting programs at all
these career stages, as well as at the high school and undergraduate levels. Although the committee did
not explore possible changes to these diversity programs, it did consider the impact of these scenarios
on NIH’s diversity efforts. NIH also supports short-term and science outreach and education programs
that encompass the K - 12 grade levels. Although rigorous science education at these early grades is
absolutely essential to creating a well-prepared pipeline of individuals for NIH’s programs at the
graduate level and beyond, NIH activities at the K - 12 stage generally provide for short-term
interventions and thus constitute a smaller contribution to science education at these earlier grade
levels.
A recurring theme throughout these scenarios is the importance and effectiveness of NIH formal
research training programs (namely, institutional training, fellowship and career development awards)
versus training supported by research grants. The formal training programs are peer reviewed not only
for the excellence of the science proposed, but also for the excellence of the proposed training and the
training potential. Formal training awards can also be targeted to emerging scientific disciplines and to
areas of national need. Of particular note, formal training programs have played a vital role in creating
incentives for increasing the diversity of the workforce pipeline. Research grants play little, if any, such
role in developing a more diverse biomedical workforce. Given the important role that formal training
programs play in research training, it is a concern that less than half of the individuals earning a research
doctorate in the biomedical sciences in 2009 were ever supported by an NIH formal training
mechanism. 71
The committee made NO distinctions among the careers of individuals engaged in biomedical research
in academia, industry, government, or any other sector; these were NOT considered ‘non-traditional’ or
‘alternative’ careers. The committee felt it was important not to confuse these careers with truly
alternative or non-traditional careers, such as science policy, law, finance, and teaching at K-12 levels or
at institutions where there is little or no research activity.
The committee made the distinction between the career paths of the majority of biomedical research
scientists, namely PhD recipients, and the much smaller population of clinician scientists (individuals
holding the MD, DO, DDS, DVM, DN, or equivalent clinical doctorate degrees who are research
scientists). Clinician scientists generally do not receive support from the NIH for their clinical training,
except for those enrolled in formal combined degree programs. The committee also notes the
70
Background and membership of the NTW Committee is in the Appendix.
71
NIH Supported Ph.D. Recipients in the NIH Data Book, at:
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=268&catId=21
Final Report 128
Appendix H Biomedical Workforce Working Group Report |
differences in immediate career options for clinicians, who may fall back on clinical practice for their
livelihoods, and PhD researchers, whose options are more limited. On the other hand, clinician scientists
face demands on their time for research activities that are not generally encountered by PhD scientists.
The committee did not take a position on whether we are producing too many or too few scientists,
although the number of PhD graduates in the biomedical sciences (excluding behavioral, social and
clinical sciences) increased by nearly 40% in the past decade (to well over 7,000 in 2010). It is difficult to
argue that during this same period opportunities for research doctorate scientists in the biomedical
workforce grew at this pace. Many argue that this is not a major concern because the individuals who
were trained acquired skills that can be of utility in a broad range of careers. Others argue that,
although true, many of these non-traditional careers don’t necessarily require a research doctorate in
science, nor the requisite investment of time in training. Thus, the committee essentially developed
scenarios from two opposing positions:
A) NIH should support primarily the training of independent research scientists; or
B) NIH should support training for a wide range of research and non-traditional careers.
The committee proposed policy changes that could be considered consistent with Position A or B.
Relevant to Position B, the recent paper by Furhmann 72 et al. provides insight into what sorts of careers
current graduate students at a top tier graduate institution are actually considering, and at what point
during their training they are making such choices. These choices are being made relatively early in their
graduate training, and include science education, science policy, the business of science, law-related
occupations, and science writing, among others. The NTW committee developed several scenarios
based on these findings.
The biomedical sciences workforce pipeline is a continuum, but the focus of the scenarios in this report
is on three distinct training and career stages of biomedical scientists, namely:
A) Predoctoral training
B) Postdoctoral training
C) Early career development
There are more policy options and leverage at the beginning of the career continuum, since changes
made early in the training stages have greater potential to align the numbers of individuals in training to
existing opportunities, especially in the long term. Conversely, policy and programmatic changes later in
the training pipeline offer fewer opportunities for realignment; in addition, they cannot easily address
the current population of individuals already at the Early Career stage. In the Early Career stage, the
committee struggled with how to retain highly trained individuals in the research workforce without
impacting the research grant budget. For this reason, the committee chose not to reduce the large
postdoctoral pool directly but to make changes earlier in the training pipeline and allow workforce
supply and demand to self-correct.
The scenarios in each career stage represent moderate to significant policy changes required of the NIH,
depending on the desired outcome. There were varying levels of support for the different scenarios
from the full NTW committee; however, most disagreement was manifested in the scenarios for the
Early Career stage, for the reasons mentioned above. In addition, committee members noted that
72
Improving Graduate Education to Support a Branching Career Pipeline: Recommendations Based on a Survey of
Doctoral Students in the Basic Biomedical Sciences. (Fuhrmann, et al, CBE—Life Sciences Education, Vol. 10, 239–
249, Fall 2011) http://www.lifescied.org/content/10/3/239.full
Final Report 129
Appendix H Biomedical Workforce Working Group Report |
solutions that work for one biomedical discipline may not work for another, and that particular NIH
Institutes and Centers might have unique issues. Finally, for all scenarios, there was concern that some
of the policy changes could impact efforts to create a more diverse workforce, but there was agreement
that adjustments in policies were possible to minimize or eliminate these effects.
Assumptions
• The NIH currently supports the graduate research training of many more individuals than there are
opportunities for independent investigators in academia, industry, and government.
• Most NIH support for graduate students is provided by research grants rather than by formal
training. This approach affords the NIH little control over the number of students trained, their fields
of study, the quality of research training, or the diversity of the trainees.
• The imbalance in the supply and demand of research scientists has led to longer periods ‘in training’
beyond the receipt of the PhD degree, as individuals wait for opportunities to materialize.
• This situation is discouraging some of the most talented college graduates from pursuing a
biomedical research career. Graduate school admissions decisions are driven by incentives other
than future workforce needs, including demand for graduate researchers and teaching assistants
and the availability of extramural research funds for their support.
• Institutions are responsible for a small fraction of a graduate student’s financial support; the balance
comes from research grants and, to a lesser extent, from training grants and individual fellowships.
Scenario Overview
• Significantly increase the proportion of PhD recipients that move into careers as independent
research scientists, and significantly reduce the number of biomedical graduate students.
Intended Outcome
Prepare the most talented students for careers as independent research scientists in numbers that are
more aligned with current and projected demand.
Supporting Data
• Total graduate enrollment and first-time, full-time enrollment (NIH Data Book)
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=227&catId=19
• Primary mechanisms of graduate support (NIH Data Book)
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=235&catId=19
• Career Outcomes of NRSA Trainees/Fellows vs. Other Biomedical PhDs (National Research Council.
2011. Research Training in the Biomedical, Behavioral, and Clinical Research Sciences. Washington,
DC: The National Academies Press.)
http://grants.nih.gov/training/Research_Training_Biomedical.pdf
• Comparison of PhD graduates to NIH awardees (Survey of NIH Support of Physician Scientists Post-
Graduate Years to First Major Research Project Grant – A Report to the NIH Deputy Director for
Extramural Research, September 20, 2011)
o 8,021 individuals graduated with PhDs in biological/biomedical sciences in 2009
o 2,214 (27.6%) of the PhD graduates had received NIH training grant or fellowship support
during the course of their training
o Only 650 Early-Stage Investigators (ESIs) with PhD degrees received R01 research awards in
FY2010
• Expected funding growth and projected workforce size (National Research Council. 2011. Research
Training in the Biomedical, Behavioral, and Clinical Research Sciences. Washington, DC: The National
Academies Press.) (http://grants.nih.gov/training/Research_Training_Biomedical.pdf)
o “For 2006, total U.S. biomedical research funding, from government, industry, and
foundations, was $93.4 billion, or $262,000 per scientist . . . Real growth in funding, from
2003 to 2007, was 3.4 percent annually. If the growth rate stays at this level (or declines
because of recession), funding growth will be slower than the projected growth of the . . .
workforce.”
• Life sciences doctorate holders that ever held a postdoctoral position (NSF Science and Engineering
Indicators 2010, Chapter 3) http://www.nsf.gov/statistics/seind10/pdf/c03.pdf
o The percentage of life sciences doctorate holders that ever held a postdoctoral position has
increased from 46% (for those who received their doctorate before 1976) to 60% (for those
receiving their doctorate in 2002-2005)
Assumptions
• The number of graduate students pursuing PhD degrees in the biomedical sciences continues to
increase, although opportunities for these individuals to pursue doctoral-level research in academia,
industry, academia, and government do not.
• Graduate students play an important role in the biomedical research workforce, regardless of
whether they ultimately pursue research careers.
• In many instances, graduate students make decisions about ‘non-traditional’ careers early in their
PhD training.
• Doctoral education in the biomedical sciences provides important and transferable skills that can be
applied to a broad range of research and ‘non-traditional’ careers.
• Despite their high-level skills, students in the biomedical sciences often feel that they are trained too
narrowly to readily enter the workforce and transition to ‘non-traditional’ careers.
Scenario Overview
• Incentivize graduate school programs to monitor the career intentions of students in their early
years of graduate study, and to be responsible for supporting students that have interests in ‘non-
traditional’ careers.
Intended Outcome
Provide the opportunity for individuals early in their graduate school training to select a career option
other than biomedical research.
Supporting Data
• Graduate enrollment in the biomedical sciences, by citizenship (NIH Data Book)
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=252&catId=19
• National Research Council. 2011. Research Training in the Biomedical, Behavioral, and Clinical
Research Sciences. Washington, DC: The National Academies Press.)
http://grants.nih.gov/training/Research_Training_Biomedical.pdf
• Improving Graduate Education to Support a Branching Career Pipeline: Recommendations Based on
a Survey of Doctoral Students in the Basic Biomedical Sciences. (Fuhrmann, et al, CBE—Life Sciences
Education, Vol. 10, 239–249, Fall 2011) http://www.lifescied.org/content/10/3/local/complete-
issue.pdf
Assumptions
• The number of graduate students pursuing PhD degrees in the biomedical sciences continues to
increase, although opportunities for these individuals to pursue doctoral level research science in
industry, academia and government do not.
• Foreign students now constitute close to 30% of graduate school enrollment in the biomedical
sciences.
• Graduate students play an important role in biomedical research, in part by providing a low-cost
workforce, regardless of whether they ultimately pursue research careers.
• Doctoral education in the biomedical sciences provides skills that can be applied to a broad range of
careers.
• Despite their high-level skills, students in the biomedical sciences often feel that they are trained too
narrowly to readily transition to non-research careers.
Scenario Overview
• Align biomedical graduate school training and educational experience with the reality that students
will pursue a range of career outcomes aside from research science, and provide them with a
broader set of skills to prepare for such careers.
Intended Outcome
Students would earn research doctorate degrees in biomedical sciences, but would be better prepared
to embark on careers related to science.
Supporting Data
• Total graduate enrollment and first-time, full-time enrollment (NIH Data Book)
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=227&catId=19
• National Research Council. 2011. Research Training in the Biomedical, Behavioral, and Clinical
Research Sciences. Washington, DC:The National Academies Press.
http://grants.nih.gov/training/Research_Training_Biomedical.pdf
Assumptions
• The NIH is currently supporting the postdoctoral training of more individuals than there are
opportunities for independent research scientists in academia, government, and industry.
• Retention of a very large pool of scientists ‘in training’ for a prolonged time makes biomedical
research a less attractive career choice to the best and brightest domestic students .
• The numbers of postdoctoral scientists are dependent on the rate of U.S. PhD production and, to a
larger extent, the recruitment of foreign postdoctoral scientists who, because of cultural and
language differences, may not be fully contributing to or benefiting from the intellectual
environment.
• The time of postdoctoral training is indefinite and can be as long as, or even longer than, the
predoctoral training.
• The compensation of biomedical postdoctoral scientists does not reflect their years of professional
training and experience, and lags behind the compensation of postdocs in other science areas.
Scenario Overview
• Target NIH formal training support to individuals most likely to continue in careers as research
scientists with the emphasis on high quality career development experiences
• Shift the balance in Kirschstein-NRSA postdoctoral training so that a greater proportion of PhD
training occurs through fellowships rather than training grants
o Limit the number of formal institutional training positions for postdoctoral scientists,
encouraging their use in emerging, specialty, multidisciplinary, or interdisciplinary fields and
shifting other postdoctoral training positions to individual postdoctoral fellowships.
o Increase the starting postdoctoral trainee and fellowship stipend to levels comparable to
those provided to postdoctoral researchers in other science fields.
o Expand the use of transitional career development awards for advanced postdoctoral
fellows.
o Require that postdoctoral scientists supported by research grants are within five years of
receipt of their terminal doctoral degree.
o Limit the support of postdoctoral scientists on research grants by adopting one or more of
the following policies:
Require individual development plans of all postdocs supported by NIH funding.
Mandate full employee benefits for all individuals supported on research grants that
are at or above the postdoctoral level.
Reduce the indirect cost rate to 8% for postdoctoral positions on research grants.
Intended Outcome
With an emphasis on high quality research experiences, provide for the professional development of
postdoctoral individuals so that they are fully prepared for research careers.
Supporting Data
• Postdoctoral Population in the US in Biomedical Sciences (NIH Data Book)
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=255&catId=20
• Funding success of Early-Stage Investigators with PhD Degrees. (Survey of NIH Support of Physician
Scientists Post-Graduate Years to First Major Research Project Grant – A Report to the NIH Deputy
Director for Extramural Research, September 20, 2011, Data from IMPAC II)
o 650 Early-Stage Investigators with PhD Degrees received R01 funding in FY2010.
• Expected funding growth and projected workforce size (National Research Council. 2011. Research
Training in the Biomedical, Behavioral, and Clinical Research Sciences. Washington, DC:The National
Academies Press.) http://grants.nih.gov/training/Research_Training_Biomedical.pdf
o “For 2006, total U.S. biomedical research funding, from government, industry, and
foundations, was $93.4 billion, or $262,000 per scientist . . . Real growth in funding, from
2003 to 2007, was 3.4 percent annually. If the growth rate stays at this level (or declines
because of recession), funding growth will be slower than the projected growth of the . . .
workforce.”
• Starting Postdoctoral Stipends in 2009 (Doctorate Recipients from U.S. Universities: 2009, NSF)
http://www.nsf.gov/statistics/nsf11306/data_table.cfm#44
o Note: Kirschstein-NRSA Stipend, 2009: $37,368 (NOT-OD-09-075)
Kirschstein-NRSA Kirschstein-NRSA
postdoctoral fellows postdoctoral trainees
Application rate for a subsequent R01 48.3% 32.5%
research award
Funding rate for subsequent R01 research 63.3% 50.0%
award
Average time to receive subsequent R01 5.3 years 6.5 years
research award
Assumptions
• There are more postdoctoral individuals in training than there are opportunities for employment as
research scientists.
• These numbers are directly dependent on the rate of U.S. PhD production, and the recruitment of
foreign postdocs.
• As evidenced by their large numbers, postdocs fill an important need in the biomedical research
workforce, although their compensation does not align with their level of training and experience.
• A small proportion of postdocs will become research scientists; transition awards facilitate their
transition to independence.
• Many postdocs will not become independent or lead scientists, but may play valuable roles in the
support of the research establishment, either directly as research staff scientists, or in careers that
relate to or support the biomedical research enterprise.
• Postdoc experiences should align with this broad range of possible career options.
Scenario Overview
• Align the postdoctoral experience to the reality that not all individuals will move on to careers as
independent research scientists
Intended Outcome
Preserve postdoctoral fellows as a large and important component of the biomedical workforce, but
ensure that they are better prepared to move on to both research and ‘non-traditional’ careers.
Supporting Data
• Postdoctoral Population in the US in Biomedical Sciences (NIH Data Book)
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=255&catId=20
• Funding success of Early-Stage Investigators with PhD Degrees. (Survey of NIH Support of Physician
Scientists Post-Graduate Years to First Major Research Project Grant – A Report to the NIH Deputy
Director for Extramural Research, September 20, 2011, Data from IMPAC II)
o 650 Early-Stage Investigators with PhD Degrees received R01 funding in FY2010.
• Expected funding growth and projected workforce size (National Research Council. 2011. Research
Training in the Biomedical, Behavioral, and Clinical Research Sciences. Washington, DC:The National
Academies Press.) http://grants.nih.gov/training/Research_Training_Biomedical.pdf
o “For 2006, total U.S. biomedical research funding, from government, industry, and
foundations, was $93.4 billion, or $262,000 per scientist . . . Real growth in funding, from
2003 to 2007, was 3.4 percent annually. If the growth rate stays at this level (or declines
because of recession), funding growth will be slower than the projected growth of the . . .
workforce.”
• Starting Postdoctoral Stipends in 2009 (Doctorate Recipients from U.S. Universities: 2009, NSF)
(http://www.nsf.gov/statistics/nsf11306/data_table.cfm#44)
o Note: Kirschstein-NRSA Stipend, 2009: $37,368 (NOT-OD-09-075)
Pros Cons
incurred by postdocs who are undergoing
protracted ‘training’ that limits their long-term
earning power
• Does not address the unattractive career
outlook for postdocs (i.e., no retirement fund,
health benefits, etc.) when compared to other
professionals at comparable career stages
• Fails to increase the chances for postdocs to
become independent or lead investigators
unless there are substantial limits placed on
the number of PhDs awarded
• Increasing stipends will raise overall research
costs
Assumptions
• A large proportion of the work on research grants is currently carried out by graduate students and
postdoctoral fellows, who are sources of inexpensive temporary labor.
• External factors, such as the state of the economy and the amount of public support for research,
currently drive career opportunities for postdoctoral researchers.
• At the moment, a sizeable group of very strong researchers are trapped in a career stage just
beyond the postdoctoral pool.
• The nation cannot afford a “lost generation” of researchers. NIH needs to provide the opportunity to
allow the best and brightest of the postdoc pool to advance.
Scenario Overview
• Create a research scientist career award mechanism aimed at recent postdoctoral individuals with
the intention to provide a large fraction of the resources needed for institutions to establish a
faculty appointment:
o The award would be renewable, if the recipient has obtained independent research support
by the time of renewal.
o Provides a way to move postdoctoral researchers into positions where they can direct their
own research.
• The concept is counter-cyclical; therefore, when faculty openings are abundant, NIH would limit the
number of awards. When faculty openings are very limited, NIH would increase the funding of this
program.
Intended Outcome
This award mechanism would help to retain highly trained research scientists in the biomedical
workforce during difficult budget periods by moving them to awards with lower indirect costs, thereby
reducing the number (and cost) of postdoctoral employees supported by research grants.
Supporting Data
AAMC Data Book (https://www.aamc.org/data/databook/)
Assumptions
• There is a cadre of highly-trained scientists emerging from post-doctoral training experiences that
have exquisitely-developed technical skills and a scientific conceptual framework and depth that
could play an invaluable role as career staff scientists and creative team members, beyond the
capabilities of the best trained technicians.
• Career staff scientists make very substantive contributions to the biomedical research enterprise,
generally serving more than one research project grant or subprojects of center grants.
• Career staff scientists are generally not suitable or competitive as Principal Investigators on their
own research grant applications.
• Several NIH Institutes utilize the Research Core Center grant mechanism, i.e. P30, P50, P60, which
typically support one or more research-serving cores providing centralized research resources,
services and facilities to an aggregate of funded research grants with common technical needs.
Scenario Overview
• Support the hiring of staff scientists in core services laboratories with NIH research grant support.
The career staff scientists will bring new technical competencies and scientific conceptual thinking
to investigative teams that will promote the “cutting-edge” capabilities and productivity of
individual projects.
Intended Outcome
Provide an NIH-supported, viable career pathway for individuals who might otherwise leave the
biomedical workforce.
Clinician Scientists
The situation regarding research careers for clinician scientists 73 is very different than the situation
discussed for research doctorates in the previous scenarios. While there are many PhD holders seeking
limited research jobs (the domestic plus foreign postdoctoral pool is estimated to be more than 35,000 74
and growing), the number of clinician scientists at the postdoctoral level (post-residency) is much
smaller. Additionally, only a small fraction of medical students express interest in careers as
independent research scientists. This committee concentrated on the challenges to becoming a clinician
scientist and considered the critical need for a comprehensive workforce infrastructure to support
clinical-translational research as beyond the scope of this exercise.
The clinician scenario includes actions that NIH could take to support the careers of clinician scientists at
three career stages: medical school, residency, and early career. For medical students, NIH could
support more programs for short-term research experiences during medical school, and create more
awards for ‘year out’ programs. At the residency level, NIH could create awards that support ‘fast-
tracking’ in residency programs, thereby allowing clinicians to begin research careers sooner. At the
early career stage, the NIH Loan Repayment Program could be extended to more clinicians by
broadening the eligibility criteria. Finally, the NIH could create incentives for more team approaches to
research that involve collaborations between clinician scientists and PhDs. Combined, or in part, these
measures could contribute to increasing the numbers of clinicians who become scientific leaders by
making careers more attractive and attainable.
3
For the sake of clarity, we use the term ‘clinician scientist.’ This could be any clinical doctorate
(physician, DN, DDS, DVM, etc.) who also performs scientific research, including bench science,
epidemiological work, outcomes research, etc. However, most of the available data pertain to MD
degree holders and those who hold both MD and PhD degrees.
74
Postdoctorates, by citizenship and gender (NIH Data Book)
http://report.nih.gov/NIHDatabook/Charts/Default.aspx?showm=Y&chartId=255&catId=20
Final Report 151
Appendix H Biomedical Workforce Working Group Report |
Background Information
The group was instructed to place some boundaries around what issues would be addressed regarding
clinician (or equivalent degree holders) scientists’ involvement in biomedical and clinical research.
While it was acknowledged that the clinical-translational research workforce encompasses a wide array
of disciplines and skill sets, this group will concern itself with a more narrowly defined group: the
scientific leaders in the clinical-translational research workforce, whether in industry, private research
foundations, academia, or government.
With the data at hand, it is difficult to determine if we are preparing a sufficient number of clinicians as
independent scientists for the biomedical workforce of the future. There is abundant anecdotal
evidence of shortages in specific fields. Unlike the situation with the PhD pool of scientists, where the
supply of research doctorates exceeds the demand, the number of clinician scientists (and certainly the
legendary ‘triple threat’ individual: clinic; bench; classroom) has been declining over the past several
decades. The ‘demand’ for these individuals is in fact driven primarily by NIH research dollars, as
‘positions’ on medical school faculty are only created when an individual can demonstrate the ability to
successfully compete for extramural research dollars. Thus, given the current system’s dependence on
NIH research support of clinicians to pursue research programs, this group intends to steer clear of the
debate about too many versus too few, and instead concentrate on how such research careers can be
made more attractive to the most capable individuals.
Assumptions
• MDs, MD/PhDs and PhDs have similar success rates when applying for NIH research grants, with
MD/PhDs holding a slight advantage. This has remained unchanged for a number of years.
• In this respect, MD/PhD programs have been very successful in seeing a much greater proportion of
their graduates being successful at earning independent research support than PhD programs.
• While the number of MDs applying for NIH research grants has remained fairly constant over several
decades, the number of PhDs applying continues to increase.
• The perception, and perhaps the reality, is that there are insufficient numbers of clinician scientists
on NIH review panels that review applications from clinician scientists.
• Salary and lifestyle issues play a significant role in clinicians opting out of a career in research.
Lifestyle issues especially impact the decisions of women to pursue careers as clinician scientists.
This is important as the proportion of women medical students has steadily increased over the
years.
• It is worth repeating that ‘bedside-to-bench’ is as important as ‘bench-to-bedside’, and that the
isolation noted above threatens both, but especially the former. Both scientists and clinician
scientists are becoming increasingly isolated from the clinicians who interact with patients at
academic medical centers.
• Possible career stages at which the NIH policies/programs could influence clinicians’ interest in and
preparation for a career in research include medical school, residency, and post-residency career
stages.
Final Report 152
Appendix H Biomedical Workforce Working Group Report |
• Recent evaluations of NIH’s K awards to early career clinician (MD and MD/PhD) scientists (post-
residency stages) show a significant positive impact on their ability to obtain independent research
support.
• There remains insufficient data on academic medical center faculty appointments that actually
provide the opportunity to develop a career as a clinician scientist. These seem to be essentially
derivative; the absolute number of these positions is dependent on continued success at acquiring
NIH research grant (or other major grant) support, and thus ultimately dependent on the NIH
research grant budget.
• The economic pressures on the academic medical centers push them to rely increasingly on ‘soft’
money from the NIH or other funding organizations for the support of their research enterprises.
Scenario Overview
• Enhance existing NIH programs directed at medical students, residents and early career clinician
scientists
DISCUSSION
The following considerations led to the policy changes recommended.
• The Loan Repayment Programs restrict the areas of research and lead to the exclusion of clinician
scientists pursuing more fundamental areas of biomedical science. This is unfortunate as it excludes
a group of scientists who could serve as links between the basic and clinical sciences.
• There is a demand for more support for ‘year out’ programs for medical students. These programs
attract very talented medical students to research activities, and in many cases to research careers.
These could be designed to restrict support to a stipend and health insurance costs, with no tuition
reimbursement. Should NIH ICs offer more of these? Should NCATS fund these programs?
• The T35 short term research experiences for medical students are also quite popular, although not
offered by all ICs, and suffer from the limitation that one medical school may not be able to enroll
sufficient numbers of students in a specific IC mission-oriented scientific area. Should NCATS
support these programs?
• Although NIH support of faculty salaries may require downward adjustment, should NIH’s career
awards policies be more flexible with salary reimbursement for these early career clinicians?
• Should NIH consider more dual degree individual fellowships for those medical students who
develop a strong interest in earning a research degree to complement their clinical degree?
• Should the limit on the number of years of support from an NIH career award be relaxed so as to
provide more protected time to be in a position to successfully compete for independent research
support?
• The notion of fast-tracking a resident to a research career should be seriously considered, although
this would involve decisions beyond the NIH, namely the buy-in of residency programs and
professional accrediting organizations. Select residency programs still value the presence of
research-oriented clinicians in their classes. In fact, select residency programs continue use the
presence of research-oriented physicians as a recruitment incentive to attract the best of the
medical graduating classes to their specialty.
• The usual issues of procedure-intensive specialties precluding sufficient time to establish a research
career came up, with the usual alternatives: relax effort requirements on career awards and/or
extend the length of career awards to clinicians, although many thought this was not a promising
way to get sufficient research experience and prove one’s independence. For individuals in those
specialties, some ventured that if one is to have a research career, less time needs to be spent
staying proficient in procedures, and more time must be spent engaging in research.
Intended Outcome
Increase the number of clinicians who become scientific leaders by making clinician-research careers
more attractive and attainable.
Supporting Data
• The Vanishing Physician Scientist? Andrew I. Schafer, ed. Cornell University Press, 2009.
• AAMC Faculty Databases
• Funding success of Early-Stage Investigators in FY2010. (Survey of NIH Support of Physician Scientists
Post-Graduate Years to First Major Research Project Grant – A Report to the NIH Deputy Director for
Extramural Research, September 20, 2011, Data from IMPAC II)
o PhD degrees: 650
o MD degrees: 132
o MD/PhD degrees: 115
• AAMC Survey of Medical Students
• NIH Individual Career Development Awards Program-Evaluation 2011:
http://grants.nih.gov/training/K_Awards_Evaluation_FinalReport_20110901.pdf
Co-chairs:
• Rod Ulane, OD Director, Division of Scientific Programs, Office of Extramural Programs, Office
of Extramural Research
• Richard Baird, NIBIB Director, Division of Interdisciplinary Training, NIBIB
Full Membership:
Name and IC Title/Position
Amy Adams, NIDCR Director, Office of Science Policy & Analysis
David Armstrong, NIMH Chief, Scientific Review Branch
Robin Barr, NIA Director, Office of Extramural Activities
Ken Bridbord, FIC Director, Division of International Training and Research
Alison Cole, NIGMS Assistant Director for Research Training
Sandra Colombini-Hatch, NHLBI Medical Officer
Associate Director, Division of Neuroscience & Basic Behavioral
Nancy Desmond, NIMH
Science
Irene Eckstrand, NIGMS Supervisory Health Scientist Administrator
Katrin Eichelberg, NIAID Scientific Review Officer
Robert Finkelstein, NINDS Associate Director for Extramural Research
Valerie Florance, NLM Director, Division of Extramural Programs
Greg Germino, NIDDK Deputy Director, NIDDK
Mimi Ghim, NDA Deputy Coordinator for Research Training
Ann Graham, CSR Program Analyst
Lindsey Grandison, NIAAA Project Director
Anthony Hayward, NCRR Director, National CTSA Program
Milton Hernandez, DLR Director, Loan Repayment Program