Myers, Elizabeth R. PhD; Alciati, Marianne H. PhD; Ahlport, Kathryn N.; Sung, Nancy S. PhD
Editor’s Note: Commentaries on this article appear on pages 1464 and 1466.
The health research enterprise comprises a continuum extending from basic biologic discovery through studies of pathogenesis to development and clinical testing of diagnostic or therapeutic interventions to evaluation in populations and health systems. Knowledge flows in both directions along this continuum, with clinical insights and population outcomes feeding back to inform hypotheses in basic discovery research.1,2 In the United States, the National Institutes of Health (NIH) supports the vast majority of the basic and translational aspects of this enterprise, whereas industry supports applied research leading to commercialization of interventions or therapeutic products.
Philanthropic organizations complement these government and corporate funders, providing research support unencumbered by the need to respond to fluctuating political situations or return profits to shareholders. Strategies used by philanthropy include targeting specific and sometimes-neglected disease areas, supporting vulnerable researchers (e.g., early-career scientists, members of underserved communities), and supporting high-risk projects. In many cases, philanthropic funders have tested new grant mechanisms or opened new fields, which are later taken to scale by federal funders.3 Despite their relatively modest means, philanthropic organizations have provided essential risk capital to the health research enterprise.
The medical research community currently has no detailed or aggregated source of information on health research funding across philanthropic organizations. However, the Health Research Alliance (HRA), a consortium of private, nonprofit funders of biomedical research and training, has created and maintains an online database of research grants awarded by its member organizations (gHRAsp: Grants in the HRA Shared Portfolio). In the present report, we provide the first analysis of these data with the aims of (1) characterizing funding by HRA members from 2006 to 2008, (2) identifying trends in the data, including comparisons with federal government funding, and (3) establishing a baseline for future comparisons.
Thirty-two HRA member organizations* uploaded information into gHRAsp about their biomedical research and training grants that had start dates in 2006, 2007, and 2008. Membership in the alliance is limited to nonprofit, nongovernment funders of peer-reviewed health research and/or training that publish their awarded grants as well as their process for vetting grant applications. The grants whose data are captured in the gHRAsp database are estimated to account for approximately 40% of philanthropic health research funding.4 (See Supplemental Digital Table 1 at http://links.lww.com/ACADMED/A103, which provides a listing of the mission statements and links to research award program descriptions for each of the 32 member organizations that submitted 2006–2008 awards data to gHRAsp.) One member organization, the Howard Hughes Medical Institute (HHMI), a nonprofit medical research organization, provides its research support both in the form of grants and as laboratory support for HHMI investigators who are classified as HHMI employees, not grantees. Only HHMI grants are included in gHRAsp, although nongrant research support is listed in Table 1 to provide a more complete picture of the overall scope of nonprofit funding.
We collected data about each HRA member organization’s grants (start and end dates, duration, award amount, type of grant), investigator (degrees earned, fields of training, gender, race, ethnicity), and recipient institution (name, location). Characteristics of funding organizations were obtained from the Web site of each organization (fundraising status) and from www.Guidestar.org (funder type).
To capture the focus of each grant regarding disease or topic, research condition and disease categorizations (RCDCs) were assigned by Collexis, a company that develops semantic technology to categorize scientific publications for the NIH and others. RCDC software uses a text-mining process to provide a consistent categorization of grants into diseases and conditions. HRA grants were analyzed using selection criteria provided by the NIH so that the categories would be comparable.5 RCDC codes were assigned to HRA grants using grant title and abstract. For grants tagged with multiple codes, the full dollar amount of the grant was attributed to every assigned category. The same process of assigning a grant’s full dollar amount to every tagged category is used by the NIH. As described by the NIH, each category stands on its own: “RCDC approaches each research category as a separate inquiry, answered to the best of its ability.”5 There were 211 RCDC codes possible using the software version available at the time of this analysis. Currently, the NIH has expanded the list of possible RCDC codes to 233.5
We analyzed number of grants and dollars by grouping grant-associated variables, including year, type of granting institution, type of grant, grantee institution location, and RCDC code. Investigator-associated variables included gender, race, degree, and field of research training. We compared nonprofit funding with NIH funding using publicly available sources.5–8
Some variables were not routinely collected by all organizations, and some data were unavailable at the time of data collection. In two cases where omitting these data would have skewed results, missing total dollars and numbers of grants were either obtained from the organization or extrapolated from other years’ totals provided in the organizations’ IRS 990s.
A total of 32 HRA member organizations collectively awarded $2,712,418,254 in research and training support, reflecting 9,934 grants with start dates in calendar years 2006, 2007, and 2008. Grant funding across HRA member organizations rose by $211,939,427 (26%) from 2006 to 2008, unadjusted for inflation (see Table 1), and the number of individual grants increased by 185 (6%), reflecting growth in both the size and number of individual grants. NIH funding increased just 3% during this same time frame (see Table 1), according to annual spending levels set for the NIH by congressional appropriations.8 Investments across the broad sector of U.S. philanthropic foundations and voluntary health organizations (beyond the HRA membership) increased by an estimated 20%.4
When HHMI nongrant research support is added to the grants captured in gHRAsp, total research funding by HRA member organizations increased from $1.35B to $1.68B (25%) across three years (see Table 1). This includes costs associated with the HHMI research facilities at Janelia Farm as well as for HHMI investigators nationwide. Whereas the total provided by HRA members (including HHMI) is only about one-twentieth of the NIH’s total research funding, HRA member organizations collectively would have ranked sixth among the NIH’s 27 institutes and centers in terms of the amount of funding in 2008.8
As shown in Table 2, more than 80% of the 32 member organizations reporting data in gHRAsp in 2006–2008 have a disease-specific focus, and, of these, half have a cancer-specific focus. HRA includes many of the larger health research funders, and the grants captured in the gHRAsp database from 2006 to 2008 ($2.71billion) are estimated to account for approximately 40% of major philanthropic health research funding over the same time frame ($6.64M)4 (see Table 1). Furthermore, the vast majority of the 32 member organizations are public charities (n = 26; 81%), and the remainder are private foundations. Public charities are established and sustained with funding received from the public and/or other organizations. Private foundations, in contrast, are usually established with funding in the form of an endowment, and they do not raise public funds. Nearly 88% of research dollars recorded in gHRAsp during the study period came from disease-focused organizations that raise funds from the public to support their missions (data not shown).
More than half of all dollars awarded across HRA member organizations supported research project grants (56%), accounting for 43% of grants awarded (see Table 3). Career grants, which include training fellowships and early faculty awards to an individual, represented the next-largest investment, reflecting almost one-quarter of the dollars and one-third of the grants awarded. Institutional training grants, in which the institution selects trainees, represented 21% of grants awarded but only 6% of research dollars invested, reflecting the relatively small size of these grants. Conversely, consortia and center grants were few in number but larger in the dollar amounts awarded. Comparison with the NIH revealed that HRA members invested much more in career and training grants than did the NIH6 as a proportion of total funding from 2006 to 2008 (30% versus approximately 7%).
HRA members collectively provided grant funding to institutions in 49 of the 50 U.S. states from 2006 to 2008. The states receiving the largest amounts of funding were California, Massachusetts, New York, Texas, and Pennsylvania, mirroring the geographic pattern of NIH funding to a great extent.7
Ninety percent of all HRA member funding was awarded to institutions in the United States. The remaining 10% was distributed to institutions in 35 non-U.S. countries. This is most likely an underestimate of the extent of HRA funding abroad, however, because gHRAsp only captures the geographic location of the principal investigator’s home institution. Many HRA organizations fund clinical trials that accrue patients both in the United States and abroad, or fund U.S.-based investigators working in international settings or with international collaborators. During 2006 to 2008, the non-U.S. countries receiving the greatest share of direct research funding from HRA member organizations were Australia, Canada, and the United Kingdom. Most international grants were research project grants (71%).
Of the 211 automated RCDC codes used by the NIH to categorize its grants, 174 were also assigned to gHRAsp grants, indicating that the scope of research funded by HRA members is nearly as broad as that of the NIH. Ninety-five percent of the grants in gHRAsp were assigned to at least one of the codes that was among the 15 RCDC codes most frequently assigned to HRA member grants, as shown in Table 4.
Although NIH grant funding was more than 18 times that of HRA members during the study period, a comparison of the proportional distribution of funding across HRA’s top 15 RCDC categories in 2008 (shown in Figure 1) illustrates the relative priorities placed on these categories by the NIH and by the philanthropic community, as represented by HRA members. A higher proportion of NIH grant dollars was allocated to categories such as biotechnology, neuroscience, and behavioral and social sciences compared with HRA dollars, whereas HRA member organizations funded proportionally more research addressing specific diseases or conditions which include several of the leading causes of U.S. deaths,9 such as cancer (28.8%), diabetes (21.0%), and cardiovascular disease (9.3%). This reflects the fact that the donor-driven philanthropic community in biomedical research, which makes up the bulk of HRA’s membership, tends to support specific disease areas more than research in basic or social sciences.
In 2006, 2007, and 2008, HRA member organizations supported a total of 8,849 unique investigators, indicating that 1,085 of the 9,934 grants were awarded to investigators already holding a grant from one of the HRA member organizations. Approximately 36% of the unique investigators were women, but women accounted for increasingly larger proportions in cohorts of investigators who had received their degrees more recently (Figure 2). Of those investigators whose most recent doctoral degrees were received between 2000 and 2009, 41% were women. This represents a large increase: Among those receiving doctoral degrees between 1970 and 1979, only 25% were women, and before 1970, only 15% of that group were women. When the data are broken down by type of grant, women made up 38% of HRA career grant awardees and 33% of HRA investigators with research project grants, which are usually awarded at a later career stage (data not shown). For NIH-supported grants during the same period, women represented approximately 43% of career awardees and 27% of those with research project grants.6 It is notable that there is more of a gap in the representation of women between NIH-funded career versus research project awards (16%) than there is for HRA-funded awards (5%). Future tracking of gender data across the different grant types, however, will be necessary to determine whether there are significant differences between philanthropic and federal funding sources in retention of female investigators as they advance from early career support to independent research funding.
Submission of information about investigator race and ethnicity to gHRAsp was optional, but race was included for approximately two-thirds of investigators, among whom 66% were white and 32% were Asian. Blacks (i.e., African Americans) accounted for less than 2% of investigators. Changes over time in the racial demographics of HRA-funded investigators were evident. There was an increasing proportion of Asian investigators among those earning doctoral degrees after 1990 (Figure 3). This likely reflects the influx of Chinese students and postdoctoral trainees in the early 1990s. Although the presence of blacks and other underrepresented minority investigators is vanishingly small among those investigators earning doctoral degrees before 2000, they begin to appear among those earning doctoral degrees more recently.
Information about terminal degrees was included in gHRAsp for 93% of investigators. Among those, nearly all had an advanced degree, and among those with advanced degrees, 97% held doctorates. Half of all investigators held a PhD or equivalent degree, 17% held a medical doctorate only (doctor of medicine, veterinary medicine, dental medicine, or medical science), and 21% held both a medical degree and PhD or equivalent doctorate. The majority of investigators on research project grants held PhDs or equivalent (59%) compared with medical (20%) or dual degrees (20%). The predominance of PhDs was lower for principal investigators of career development grants: 51% held PhDs or equivalent, 21% held medical degrees only, and 26% held dual degrees. An analysis of trends in characteristics of first-time applicants for NIH R01 grants from 1964 to 2004 was completed in 2006.10 Whereas a majority held doctoral degrees, only about 15% held dual degrees—slightly lower than the proportion of dual degree investigators funded by HRA member organizations.
HRA member organizations were encouraged, but not required, to include the research training field for all funded investigators, and it was available for 64% of the investigators in gHRAsp. Reflecting the increasing interdisciplinary focus of research training, multiple training areas (up to three) were submitted for 35% of these investigators. The most well-represented fields of research training among HRA-funded investigators were those traditionally associated with basic biomedical research: cell biology, biochemistry, neuroscience, physiology, genetics, and immunology.
Despite the fact that the vast majority of biomedical research in the United States is funded by industry ($61B for research and development in 2008) and the NIH ($28B for research in 2008),11 nonprofit organizations play a critical role in advancing biomedical research and sustaining the research enterprise. Foundations and voluntary health organizations are in an opportune position to fill gaps in research funding by supporting innovative and sometimes-risky research, exploring alternative approaches to supporting the research enterprise, accelerating the development of new interventions, and fostering implementation of more effective practice models.12,13 Our comprehensive analysis of research funding across HRA member organizations in the present report represents an important milestone in efforts to describe nonprofit, nongovernment contributions to biomedical research. It begins to chart a course for understanding how nonprofit organizations invest research dollars.
Other analyses have documented steady increases in philanthropic investments in biomedical research over the past two decades.11,12 The analyses presented here focus on calendar years 2006 through 2008, during which time funding for biomedical research by HRA member organizations rose substantially, increasing by almost 10% between 2006 and 2007 and by more than 15% between 2007 and 2008, unadjusted for inflation. During this same time period, NIH funding increased by a modest 3%,8 after doubling between 1998 and 2003, and overall investments by major U.S. philanthropic foundations and voluntary health associations increased by an estimated 20%.4 Thus, over this time frame, the increased support from the philanthropic sector represents an important, potentially stabilizing, source of research support. The Foundation Center, a leading source of information about philanthropy worldwide, reported that overall giving by large and midsize private foundations addressing all philanthropic areas declined in 2009 but remained steady in 2010.14 Therefore, our analysis represents an important baseline for biomedical research funders before the global recession that accelerated in late 2008.
The scope of funding across HRA member organizations addresses basic biologic discovery as well as disease-specific research, but HRA funding tends to focus on areas where the burden of disease is quite high, such as cancer, diabetes, and cardiovascular disease. The majority of HRA members are public charities whose missions focus on the prevention and treatment of specific diseases, whereas the NIH mission is to seek fundamental knowledge about the nature and behavior of living systems as well as the application of that knowledge to enhance health, lengthen life, and reduce the burdens of illness and disability.15
Slightly more than half of the grant dollars awarded by HRA member organizations during the study period supported research projects. HRA members also recognize the importance of nurturing the next generation of scientists and invest proportionally more in career development and training than does the NIH.6 These investments are most likely underestimates for both HRA and the NIH because many trainees are supported under research project grants for at least part of their training.16 And, of course, the absolute amount of NIH funding for career development and training is much greater than that from HRA members. Our findings do highlight, however, that support for training and early career development is a high priority and a distinct niche for many philanthropic organizations, and is a way they take on risk.
The future of the research workforce is of vital importance to HRA members and, indeed, is one of the factors that led to the formation of the HRA. Many HRA member organizations offer award programs targeted to physician scientists as a critical and vulnerable part of the research enterprise.17 In addition, our analysis highlights important trends in attracting more women and minorities into biomedical research careers. Women who obtained their degrees more recently make up a larger portion of privately funded investigator cohorts than was true in the past. Less encouraging evidence was provided for the involvement of racial minorities; no enrichment was observed in the younger cohorts until the most recent cohort—largely predoctoral research fellows. Diversity is clearly an area requiring further monitoring and effort within the research community. A recent study of race and NIH research awards18 found that the likelihood of obtaining NIH funding for a black applicant was only two-thirds that for a white applicant, even after adjusting for potentially confounding influences such as institution and past scientific achievements. In light of such results, both the NIH19 and HRA members are committed to exploring ways to remedy the situation. Ensuring an appropriately diverse research community is critically important to the understanding of those diseases with a disproportionate impact on racial and ethnic minorities and to the participation of minorities as research study subjects.16,20,21
There are several limitations to this analysis of philanthropic funding. First, whereas the results are representative of 32 HRA member organizations in 2006 to 2008, the grants analyzed ($2.7 billion) represent just 40% of the estimated $6.6 billion4 spent by foundations, charities, and other private philanthropic funds in 2006–2008. Although this is a robust start toward characterizing the details of philanthropic support for health research in the United States, the sample presented here may be biased because it includes only peer-reviewed grants for research and training. A second limitation results from the notable gaps in existing gHRAsp data such as ethnicity and training information. Finally, another potential limitation is the text-mining process used to categorize grants into RCDC codes. Although this analysis used the same methodology as that used by the NIH, the underlying sources of information were not identical. The NIH’s coding of grants is based on automated text searches of the title, abstract, specific aims, and public-health-relevance sections of the grant. In our study, only the title and abstract were available for text searching to code gHRAsp grants.
A major strength of the study is that it does not rely on samples of aggregated or extracted annual report data but, rather, on analysis of an underlying database of grants. Previous summaries of U.S. private nonprofit funding have relied on sampling to describe trends in research grant funding because of the lack of a comprehensive data source.4,11,12,22,23 These previous analyses did not categorize support by type of funder, type of grant, geographic area, or disease category, and no information has been reported on the characteristics of investigators of projects supported by philanthropic organizations.
Demands for increased accountability, in terms of describing both research investments and their impact, have risen dramatically over the past two decades. The U.S. government has responded with increased standardization of descriptive grant metrics across government agencies, richer data about its research investments, more analysis and compilation of funding information, and greater accessibility to data and reports. NIH agencies also are conducting and sharing analyses to reflect the impact of research funding24 and are echoing the call for accountability.25 Nonprofit funders are under no less pressure for accountability. Our inaugural analysis of the gHRAsp database begins the process of standardizing metrics across nonprofit funders, compiling and analyzing funding information, and ultimately making this information available to the research community, organizations’ constituents, and the public.
Acknowledgments: The authors thank Dr. T.J. Koerner of the American Cancer Society for his early vision for the shared database. The authors thank the many staff of HRA member organizations who contributed time and intellectual effort to implementing the project, as well as Kyle Brown and the staff at Innolyst, who built the gHRAsp database and provided considerable technical support. The authors also are grateful to Elaine Gallin, John Burris, and Lorraine Egan for critical reading of the manuscript, and thank Lisa Bain for her editorial assistance and Annette Huetter of HRA for assistance in preparing the manuscript.
Funding/Support: Dues paid by all HRA member organizations supported the creation of the gHRAsp database and analyses of the data. In addition, the Avon Foundation for Women under the direction of Dr. Marc Hurlbert provided a grant to support the gHRAsp database. The Alzheimer’s Association provided funding to support reporting the results of the analyses. The National Science Foundation provided support in the form of the time spent by Dr. Sung in preparing this manuscript while she was a program director there in 2011–2012.
Other disclosures: None.
Ethical approval: Not applicable.
Previous presentations: Data were presented at a closed meeting of the Health Research Alliance membership in March 2011. A portion of the data was presented at a closed meeting of the Institute of Medicine’s Forum on Drug Discovery, Development and Translation in June 2011 and at a closed meeting of the International Cancer Research Partners.
First published online
* Membership in the Alliance has grown significantly since the initial data submission, resulting in more growth for the gHRAsp database in future data submissions (see http://www.healthra.org/membership.profiles.html for a complete list of HRA member organizations). The HRA member organizations that contributed data to the gHRAsp database on awards starting in calendar years 2006, 2007, and 2008 are AACR Foundation for the Prevention and Cure of Cancer, Alliance for Cancer Gene Therapy, Alzheimer’s Association, American Cancer Society, American Diabetes Association, American Heart Association, Arthritis Foundation, The ASCO Cancer Foundation (now the Conquer Cancer Foundation of ASCO), Autism Speaks, Avon Foundation for Women, Burroughs Wellcome Fund, Cancer Research Institute, Inc., Children’s Tumor Foundation, Damon Runyon Cancer Research Foundation, The Donaghue Foundation, Doris Duke Charitable Foundation, Food Allergy & Anaphylaxis Network, Foundation Fighting Blindness, Inc., The Gerber Foundation, Howard Hughes Medical Institute, Juvenile Diabetes Research Foundation, Intl., W. M. Keck Foundation, Susan G. Komen for the Cure, Leukemia & Lymphoma Society, Lymphoma Research Foundation, March of Dimes Foundation, The Medical Foundation (a division of Health Resources in Action), Multiple Myeloma Research Foundation, Muscular Dystrophy Association, Prevent Cancer Foundation, Simons Foundation, and Sjögren’s Syndrome Foundation. Cited Here...