13
Research Policy 41 (2012) 1729–1741 Contents lists available at SciVerse ScienceDirect Research Policy jou rn al h om epage: www.elsevier.com/locate/respol Mission-oriented biomedical research at the NIH Bhaven N. Sampat Department of Health Policy and Management, 600 West 168th Street, Room 404, Mailman School of Public Health, Columbia University, New York, NY 10032, United States a r t i c l e i n f o Article history: Received 16 May 2011 Received in revised form 16 May 2012 Accepted 17 May 2012 Available online 22 October 2012 Keywords: NIH Medical research Serendipity Mission-oriented research a b s t r a c t The NIH (National Institutes of Health) is the largest single funder of biomedical research in the world. This paper documents tensions between the agency’s health and science missions and considers how, in light of these, it has managed to sustain a level of bipartisan political support uncommon in U.S. health or research policy. It highlights the serendipity hypothesis, the presence of “safety valve” mechanisms that allow it to (on occasion) target research at particular diseases and priorities, and a broad and diverse set of constituencies as important to understanding the agency’s political success. Through an in-depth look at the NIH allocation process, the paper also provides insights into how demand-side considerations can affect the direction of scientific research. © 2012 Elsevier B.V. All rights reserved. 1. Introduction The NIH (National Institutes of Health) is the largest single fun- der of biomedical research in the world. The agency has managed to sustain a level of bipartisan political support uncommon in health policy, and in other mission-oriented R&D agencies. 1 Former NIH Director Harold Varmus has written that “few government agen- cies have the reputation, promise, and political appeal of the NIH” (Varmus, 2009, p. 147), and elsewhere cites the common charac- terization of the agency as “the jewel in the crown of the federal government” (Varmus, 2001, p. 1903). Reflecting this support, the agency saw an unprecedented doubling of its budget over the 1997–2002 period. Political scientists have characterized it as one of the “affluent agencies” (Wildavsky, 1961) that perennially com- mand significant Congressional support and funding. The NIH’s stated mission is “science in pursuit of fundamen- tal knowledge about the nature and behavior of living systems and the application of that knowledge to extend healthy life and reduce the burdens of illness and disability.” 2 Despite the NIH’s remark- able popularity, throughout its history there have been tensions between the health and science aspects of this mission. The sci- entific community has traditionally emphasized the “fundamental Tel.: +1 212 305 7293. E-mail address: [email protected] 1 In a recent Harris Poll asking U.S. adults to rank federal agencies, 75 percent of respondents rated the NIH as doing an “excellent” or “pretty good” job, ranking it third among the thirteen agencies listed. However, it is at the very bottom of the list in terms of the share of respondents that understands what the agency is or does (Harris Interactive, 2007). 2 http://www.nih.gov/about/almanac/. knowledge” aspect of the NIH’s mission; taxpayers, Congress, and disease advocates have generally been more focused on the health aspects. This same tension has been central in postwar science policy more generally, dating at least as far back as the debates between Vannevar Bush and Senator Harley Kilgore (D–W. Va) at the end of World War Two about whether federally funded research ought to focus on outcomes or science, and related issues about the appro- priate roles of politicians and scientists in the governance of science funding (Kevles, 1977; Brooks, 1995). In making the case for sci- entific autonomy in Science The Endless Frontier, Bush famously argued that research neither can nor should be planned, since much progress is from unexpected sources. 3 Though the Bush Report had a strong ideological imprint, post- war science funding has departed considerably from the framework it presented. In particular, mission-oriented agencies have grown to dominate science funding (Mowery, 1997). Bush envisioned a single major funding agency focused on supporting fundamental research: this ultimately became the National Science Foundation (NSF). Today, this basic science funding agency’s budget is less than a quarter that of the NIH (AAAS, 2012). This presents a puzzle: how has the NIH been so successful at generating support for science? One answer is that it funds “use- oriented” research, to uses Stokes’s (1997) language, rather than pure basic research. However, as I will argue below, the NIH allo- cation process in practice is much more oriented towards scientific 3 In the context of medicine specifically, he asserted that “Progress in the war against disease results from discoveries in remote and unexpected fields of medicine and the underlying sciences.” (Bush, 1945, p. 14). 0048-7333/$ see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.respol.2012.05.013

Mission-oriented biomedical research at the NIH

Embed Size (px)

Citation preview

Page 1: Mission-oriented biomedical research at the NIH

M

BD

a

ARRAA

KNMSM

1

dspDc(tga1om

tttabe

rti(

0h

Research Policy 41 (2012) 1729– 1741

Contents lists available at SciVerse ScienceDirect

Research Policy

jou rn al h om epage: www.elsev ier .com/ locate / respol

ission-oriented biomedical research at the NIH

haven N. Sampat ∗

epartment of Health Policy and Management, 600 West 168th Street, Room 404, Mailman School of Public Health, Columbia University, New York, NY 10032, United States

r t i c l e i n f o

rticle history:eceived 16 May 2011eceived in revised form 16 May 2012ccepted 17 May 2012

a b s t r a c t

The NIH (National Institutes of Health) is the largest single funder of biomedical research in the world.This paper documents tensions between the agency’s health and science missions and considers how, inlight of these, it has managed to sustain a level of bipartisan political support uncommon in U.S. health orresearch policy. It highlights the serendipity hypothesis, the presence of “safety valve” mechanisms that

vailable online 22 October 2012

eywords:IHedical research

erendipity

allow it to (on occasion) target research at particular diseases and priorities, and a broad and diverse setof constituencies as important to understanding the agency’s political success. Through an in-depth lookat the NIH allocation process, the paper also provides insights into how demand-side considerations canaffect the direction of scientific research.

© 2012 Elsevier B.V. All rights reserved.

ission-oriented research

. Introduction

The NIH (National Institutes of Health) is the largest single fun-er of biomedical research in the world. The agency has managed toustain a level of bipartisan political support uncommon in healtholicy, and in other mission-oriented R&D agencies.1 Former NIHirector Harold Varmus has written that “few government agen-ies have the reputation, promise, and political appeal of the NIH”Varmus, 2009, p. 147), and elsewhere cites the common charac-erization of the agency as “the jewel in the crown of the federalovernment” (Varmus, 2001, p. 1903). Reflecting this support, thegency saw an unprecedented doubling of its budget over the997–2002 period. Political scientists have characterized it as onef the “affluent agencies” (Wildavsky, 1961) that perennially com-and significant Congressional support and funding.The NIH’s stated mission is “science in pursuit of fundamen-

al knowledge about the nature and behavior of living systems andhe application of that knowledge to extend healthy life and reducehe burdens of illness and disability.”2 Despite the NIH’s remark-

ble popularity, throughout its history there have been tensionsetween the health and science aspects of this mission. The sci-ntific community has traditionally emphasized the “fundamental

∗ Tel.: +1 212 305 7293.E-mail address: [email protected]

1 In a recent Harris Poll asking U.S. adults to rank federal agencies, 75 percent ofespondents rated the NIH as doing an “excellent” or “pretty good” job, ranking ithird among the thirteen agencies listed. However, it is at the very bottom of the listn terms of the share of respondents that understands what the agency is or doesHarris Interactive, 2007).

2 http://www.nih.gov/about/almanac/.

048-7333/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.respol.2012.05.013

knowledge” aspect of the NIH’s mission; taxpayers, Congress, anddisease advocates have generally been more focused on the healthaspects.

This same tension has been central in postwar science policymore generally, dating at least as far back as the debates betweenVannevar Bush and Senator Harley Kilgore (D–W. Va) at the end ofWorld War Two about whether federally funded research ought tofocus on outcomes or science, and related issues about the appro-priate roles of politicians and scientists in the governance of sciencefunding (Kevles, 1977; Brooks, 1995). In making the case for sci-entific autonomy in Science The Endless Frontier, Bush famouslyargued that research neither can nor should be planned, since muchprogress is from unexpected sources.3

Though the Bush Report had a strong ideological imprint, post-war science funding has departed considerably from the frameworkit presented. In particular, mission-oriented agencies have grownto dominate science funding (Mowery, 1997). Bush envisioned asingle major funding agency focused on supporting fundamentalresearch: this ultimately became the National Science Foundation(NSF). Today, this basic science funding agency’s budget is less thana quarter that of the NIH (AAAS, 2012).

This presents a puzzle: how has the NIH been so successful atgenerating support for science? One answer is that it funds “use-

oriented” research, to uses Stokes’s (1997) language, rather thanpure basic research. However, as I will argue below, the NIH allo-cation process in practice is much more oriented towards scientific

3 In the context of medicine specifically, he asserted that “Progress in the waragainst disease results from discoveries in remote and unexpected fields of medicineand the underlying sciences.” (Bush, 1945, p. 14).

Page 2: Mission-oriented biomedical research at the NIH

1 Policy

cs

amreBsicsttbttfddath

poa1rt2mn(ewpthahm

aebenrtiferv

caamsatt(n

730 B.N. Sampat / Research

onsiderations than health, so this alone cannot explain the NIH’strong political support.

There are several reasons for the NIH’s success. First, thegency has skillfully employed the serendipity hypothesis—thatuch progress against specific diseases results from unplanned

esearch, or unexpectedly from research oriented towards differ-nt diseases—echoing claims about the limits of planning in theush Report. However, the salience of serendipity is not enough toolve the puzzle: after all, the same argument can be made (ands) for NSF-funded research. A second reason for the NIH’s suc-ess is that even though its main funding mechanisms focus oncience rather than health, there are also various “safety valves” inhe process that do allow it to be more targeted on occasion, andhis helps to diffuse pressures from Congress and constituents toe more directly responsive in particular cases. Third, NIH adminis-rators have been politically savvy in working with interest groupso increase their budgets, cultivating relationships with power-ul members of Congress, and developing a broad and geographiciverse constituency (academic medical centers, universities, andisease interest groups) to sustain support for increasing thegency’s budget. To be sure, there have been real tensions betweenhe health and science missions, but these factors together haveelped the agency thrive.

In addition to exploring the political economy of the NIH, thisaper connects to a small but important body of economic the-ry on how demand side considerations can and should affectllocation of public funding for scientific research (Schmookler,965). Several economists have argued that NIH funding shouldeflect both scientific considerations and disease burden, equatinghe marginal product of funding across disease areas (Lichtenberg,001, Zeckhauser, 1967). Others have applied “induced innovation”odels to publicly funded research, examining the responsive-

ess of funding and publications to shocks to disease burdenCutler et al., 2009; Bhattacharaya and Packalen 2011). How-ver, this work has been silent on the mechanisms throughhich demand side considerations might influence the direction ofublic research investments. This paper helps to fill this voidhrough detailed examination of the NIH allocation process, andow demand side considerations may affect funding decisions. Thenalysis also contributes to the broader science policy literature onow agencies balance scientific considerations and social needs inaking decisions about what research to support (Brooks, 1995).I begin in Section 2 by providing an historical snapshot of the

gency, and a description of its funding patterns. In Section 3, Ixamine the NIH budget process in detail, and consider channelsy which the agency might be responsive to health considerations,.g. disease burden. Based on this institutional description andew data on funding patterns, I argue that the potential levers foresponding to specific health priorities are, in fact, more limitedhan previous scholarship suggests: the bulk of NIH funding isnvestigator-initiated and funded through a peer review processocused on science. Nevertheless, I do highlight the presence of sev-ral mechanisms—including contracts and centers, and funding viaequests for applications (RFAs)—that can be targeted, the “safetyalves” discussed above.

In Section 4, I focus on several historical debates at the NIH con-erning the “big push” programs (the Total Artificial Heart Programnd the War on Cancer) in the 1960s and 1970s, and cross-diseasellocation patterns in the 1990s. These case studies illustrate, inore detail, the core tensions between the agency’s health and

cience missions, and the political balancing act that NIH represent-tives must occasionally perform before Congress to demonstrate

heir commitments to health. These cases also show the limits ofraditional funding mechanisms for addressing health prioritiesand the resulting importance of “safety valve” funding mecha-isms), as well as the centrality of the serendipity hypothesis in

41 (2012) 1729– 1741

bolstering arguments that the health versus science distinction isa false dichotomy.

2. The NIH: an overview

The NIH’s roots trace back to a bacteriological laboratory estab-lished in the Marine Hospital in Staten Island, NY, in 1887, whichaimed to bring the latest developments in bacteriology to thework of public health, most prominently sanitation and the treat-ment of infectious disease (Harden, 1986). This Hygenic Laboratoryeventually assumed responsibility as a clearinghouse for nationalbiomedical research activities, and the 1930 Randall Act trans-formed it into the National Institute of Health, which was givenauthority to give grants to universities to help jumpstart the econ-omy during the Depression.

In 1937, the National Cancer Institute bill turned the NIH intothe National Institutes of Health. However, its transformation intoa major science funding agency began after World War Two. Likeother mission-oriented agencies, the NIH benefitted from the newpublic commitment to federal funding of research, and hopes thatscience could promote peacetime goals (including health) as effec-tively as it helped win the war. One debate during the 1940s and1950s concerned the extent to which political targeting of researchto health goals—prominent during wartime—was appropriate inpeacetime (Strickland, 1972). This foreshadowed tensions betweenpoliticians and scientists that would occasionally re-emerge overthe next half-century.

Also important to its rapid growth after World War Two, andlater sustained political support, was the active involvement ofdisease advocates. These “Noble Conspirators” (Drew, 1967), mostprominently Mary Lasker and Florence Mahoney, made expan-sion of biomedical research funding their cause, drawing on tieswith powerful politicians, opinion leaders in the media, and themedical community to develop political support for the agency.Importantly, these efforts came after their attempts to pass nationalhealth insurance failed (Starr, 1984)—among other reasons, dueto opposition from the American Medical Association—and theseadvocates searched for less controversial means of improving pub-lic health. This suggests the importance of the “health” aspects ofthe NIH’s mission, dating back to its inception, in generating earlypolitical support.

Another important participant in the early growth of the NIHwas James Shannon, who served as NIH’s Director from 1955 to1968. Unlike many of his scientific colleagues (including VannevarBush), Shannon recognized the importance of disease-advocacyand an emphasis on the health mission for securing funds, andworked closely with advocates and politicians to increase budgetallocations (Cook-Deegan and McGeary, 2006). Even in these earlydays, however, the tightrope act of balancing the health rationalefor NIH funding—needed to sustain taxpayers, disease advocates,and Congressional support—with the desire to advance funda-mental science was clearly evident. There is an account, perhapsapocryphal, that after one member of Congress addressed the Direc-tor of the Institute of Microbiology, demanding to know “Who everdied of Microbiology?” the Institute’s name was changed to theNational Institute of Allergic and Infectious Diseases (Drew, 1967).Another early illustration of this mission tension surfaced in the1960s, when the agency was subjected to growing criticism fromdisease advocates (including “the Laskerites”) and Congress, basedon concerns that the agency was not generating sufficient healthpayoffs, privileged fundamental understanding over finding cures,

and lacked the ability to target specific, urgent health problems(Drew, 1967; Strickland, 1972).

While these tensions have continued to resurface periodicallythroughout the agency’s history, in general the NIH has enjoyed

Page 3: Mission-oriented biomedical research at the NIH

B.N. Sampat / Research Policy 41 (2012) 1729– 1741 1731

F FiscalD

rigpc1rbaepp2

iNc(galKfiRaerV

The bulk of the remainder reflects funding by private sector phar-maceutical, biotechnology, and medical device firms. However,there is a sharp division of labor, with NIH funding concentrated

5 At the same time, Varmus and other NIH directors have tried to emphasize thatan excessive focus on disease categories fragments basic research, and (invokingserendipity) that much progress occurs from unexpected sources, and has there-

ig. 1. NIH Funding 1938–2010. Notes: This chart shows total NIH appropriation byeflator inflation index. The raw data were obtained from the NIH Almanac.27

emarkable political support. Indeed, a well-known political gamen Washington is for the President to provide a “low-ball” bud-et request for the NIH, recognizing that Congress historically hasrovided funding beyond this to demonstrate its support for medi-al research. As a result, the NIH’s budget grew sharply over the938–2010 period, as Fig. 1 shows. The recent spike in fundingeflects the unprecedented campaign to double the NIH’s budgetetween 1997 and 2002, spanning the Clinton and George W. Bushdministrations.4 The NIH is currently the largest U.S. federal gov-rnment funder of research (excluding development). It is alsoarticularly important in funding university research, funding 65ercent of federally funded academic research and development in009 (NSF, 2010).

While the figures above seem to imply an overall NIH budget,n practice, Congress makes specific allocations to the individualIH Institutes and Centers that comprise the NIH, as I will dis-uss in more detail below. Some of these are organized by diseasee.g., the National Cancer Institute, the National Institute for Aller-ic and Infectious Diseases, the National Institute on Alcohol Abusend Alcoholism, the National Institute for Arthritis and Muscu-oskeletal and Skin Diseases, the National Institute of Digestive andidney Diseases), while others are more focused on organ systems,elds of science and technology, or health professions (U.S Nationalesearch Council and Institute of Medicine, 2003). That the NIHnd its funding are organized by Institutes, as opposed to basic sci-nce fields, is also reflective of its necessity to appeal to “health”ather than science for political support (Drew, 1967). Thus Directorarmus (2001) noted:

The NIH thrives politically and financially from the enthusi-asm of its supporters. This enthusiasm is enhanced when anew institute or center is founded in law, especially whenthe legislator-founders are prominent, the Administration gets

credit for its role, and the advocacy groups feel a loyalty to theNIH through “their” unit (p. 1903).

4 As the figure shows the budget has plateaued since the doubling.27 http://www.nih.gov/about/almanac/appropriations/index.htm.

Year from 1938 to 2010. The figures were converted to 2000 dollars using the GDP

He also noted there that the many different institutes anddiseases associated with the agency “cannot hurt in the budgetprocess” (Varmus, 2001, p. 1903).5

Across the Institutes, about 10 percent of the NIH budgetis allocated to intramural research, conducted within its ownlabs in Bethesda, Maryland,6 while the bulk of the rest is dis-bursed extramurally, mainly to researchers based at US academicinstitutions and medical centers. The geographic scope of itsfunding creates diverse constituencies across numerous statesand Congressional districts, and may also help explain its broadpolitical support (Cook-Deegan and McGeary, 2006; Hegde andMowery, 2007). Thus advocacy groups composed of life scien-tists and medical researchers emphasize, in their “talking points”calling for increased agency spending, that NIH funding createsjobs “at more than 3000 universities, medical schools, teach-ing hospitals, and other research institutions in every state.”7 Itis perhaps telling that while the NIH has generally been reluc-tant to provide detailed funding figures by disease (for reasonsI will discuss below) the agency’s website has long provideddata on awards by state, and, more recently, by Congressionaldistrict.8

Dorsey et al. (2010) report that in the post-doubling era, NIHaccounts for nearly a third of total US biomedical research funding.

fore pushed back against recent pressures from disease advocates and Congress toestablish additional institutes (Varmus, 2001).

6 Though I do not discuss it in detail below, the intramural program is anotherpotentially important “safety valve” in the allocation process. The intramural pro-gram has more flexibility to steer research, and shift foci quickly, than the extramuralprogram. Historically, intramural research has been important in helping the agencyrespond to several public health emergencies, including the HIV/AIDS crisis (Harden,2012). I thank Sejal Patel for bringing this point to my attention.

7 https://www.aamc.org/research/adhocgp/0211talkingpoints.pdf.8 http://report.nih.gov/award/trends/State Congressional/StateOverview.cfm.

Page 4: Mission-oriented biomedical research at the NIH

1 Policy

fTaa2

mNt2

3

Npoiaat(aBgf

aicTatatc

4pbraliadmeoatfbrpo

mmrsaaHg

732 B.N. Sampat / Research

urther upstream, on “basic” research than private sector funding.9

he biotechnology, pharmaceutical, and device industries that relynd build on NIH research (and hire NIH-trained researchers)re also typically enthusiastic supporters of NIH funding (Tauzin,008).

The role of private foundations and philanthropies in fundingedical research has also expanded recently (Wadman, 2007).onetheless, NIH medical funding exceeds that from all philan-

hropies and foundations combined, by a factor of six (Wadman,007; Moses et al., 2005).10

. Health in the NIH budget and allocation processes

In this section, I consider the channels through which theIH can be responsive to health considerations, i.e. where in therocess priority setting or targeting of specific diseases couldccur. Debates about cross-disease allocation patterns (surveyedn Section 4 below) assume the NIH has the ability to targetnd prioritize specific areas of disease over others; such debateslso tend to focus on how interest groups and politics may dis-ort funding away from the cost-effective, “first-best” allocationGross et al., 1999; Johnson, 1998). Several recent economics paperspply “induced innovation” models to the NIH (Cutler et al., 2009;hattacharya and Packalen, 2011; Lichtenberg, 2001), and also sug-est demand-side considerations can influence the direction ofunding.

However, none of these perspectives identifies where in thellocation process health (or political) considerations might enter,nstead treating the agency as a black box. To examine potentialhannels, I begin with an overview of the NIH budget process.he budget process also illuminates the relationship between thegency, Congress, and interest groups, important for understandinghe agency’s sources of political support. I also discuss how the NIHllocates funds once they are appropriated. Based on these insti-

utional descriptions, I then consider where in the process healthonsiderations might enter.

9 The NIH also funds “applied research and development” accounting for about0 percent of its expenditures (Moses et al., 2005). A subset of that is funding foratient-oriented “clinical” research: in FY2007 about $9 billion of its $27 billionudget was spent on these activities. The bulk of its funding, however, is catego-ized as for “basic” research. Debates about the definitions of basic research have

long history, and need not be repeated here. It is, however, important to high-ight that whether this basic research can be associated to specific diseases (e.g.s “use oriented” basic research in the language of Stokes, 1997) before its resultsre known is a long-running concern at the NIH, and re-surfaced in several of theebates surveyed in Section 4. Indeed, at the NIH (and perhaps more generally forission agencies) the definitions between basic and applied research are perhaps

ven more blurry than emphasized by Stokes. The very same research can be basicr applied from the perspective of the funder and a scientific investigator. Thusn NIH spokesperson (cited in Strickland, 1972, p. 30) observes that “in the sensehat the new knowledge is sought for the purpose of improving human health [NIHunding] is one of applied research, but many of the grantees consider their projectsasic.” Schmookler (1965) as part of his argument that even basic science may beesponsive to utilitarian objectives, observes that while “a scientist may pursue aroject because of its prospective contribution to knowledge ... his financiers, publicr private, expect and perhaps receive useful knowledge” (p. 48).10 However, because they lack congressional oversight, private funders may beore flexible than the NIH in their investments, and thus important actors in theedical innovation system. For example, philanthropies have a disproportionate

ole in funding research on neglected tropical diseases (Moran et al., 2009). This is noturprising, since the NIH relies on support from U.S. taxpayers who are not directlyffected by these diseases. Absent direct government accountability, foundationsnd philanthropies may also be able to take more risks: for example, the Howardughes Medical Institute (HHMI) offer considerably more research freedom to theirrantees than the NIH does (Azoulay et al., 2011).

41 (2012) 1729– 1741

3.1. The making of the NIH budget

The NIH budget process formally begins in summer, when heNIH Director works with other agencies within the Departmentof Health and Human Services (DHHS, the executive agency ofwhich NIH is a part), the HHS Director, and individual NIH Insti-tute and Center Directors to prepare a budget for next fiscal year(Shapley, 1992; Varmus, 2009).11 Generally, the meetings andnegotiations over the months that follow are about overall num-bers for each Institute and Center, and what is needed for pastand future commitments.12 Much of this stage of the budget pro-cess is anticipation of “what the traffic will bear” (Shapley, 1992),i.e. what the President and Congress will support. At this stage,and throughout the process, previous NIH commitments (throughmultiyear grants) dominate the budget, an important constraint onthe agency’s capacity to respond to demand side changes (e.g. theemergence of new diseases) in the short run (Varmus, 1997, 2009).

Another Executive Branch agency, the Office of Managementand Budget, gets involved later in the process, generally pressur-ing agencies to reduce budgets before they can be included in thePresident’s official budget request in January (Shapley, 1992).13 Thediscussion at this point in the process is mainly about overall budgetfigures for Institutes and Centers, and rough estimates of number ofnew grants and funding by different grant mechanisms, i.e. thereis still little targeting of specific health considerations or priori-ties (the President’s budget does, however, include broad areas ofemphases, which can include diseases and scientific fields, whichmay be used to justify privileging some Institutes over others).

In the spring, Congress begins hearings on the NIH budget. Whileseveral committees are involved, the most important are the Houseand Senate Appropriations Committees.14 Illustrating its impor-tance (concretely) many of the buildings on the NIH’s campus arenamed for chairs of these Appropriations Committees who wereparticularly supportive of the agency.15

As discussed above, the NIH is exceptional among other federalagencies in that Congress almost always increases the Presi-dent’s request, particularly when the President does not controlCongress.16 During Congressional hearings, the NIH Director, Insti-tute Directors, disease advocates, and representatives of scientificand medical societies testify, generally in support of higher budg-ets (and to provide cover for Congress to increase budgets over thePresident’s request).

Unlike most science funding agencies, NIH budgets are gener-ally free from “hard” earmarks from Congress that specify and set

aside funds for particular diseases or institutions. Varmus (2009)suggests this lack of earmarking “represent votes of congressionalconfidence in the NIH’s system of peer review” (p. 150). However,

11 In summer 2011 these discussions began for the budget for Fiscal Year 2013,which begins in October 2012.

12 The National Cancer Institute is an exception, as discussed below. Since theNational Cancer Act of 1971 this Institute has had “bypass authority” and goesdirectly to President its budget, bypassing DHHS.

13 The negotiation with the NIH is particularly interesting, given the “annual bal-let” (Shapley, 1992) in which Congress is expected to later increase the President’sbudget request for the agency, which I alluded to above.

14 Specifically, the Labor, Health, and Education subcommittee handles the NIH’sbudget.

15 These include the Lawton Chiles International House, the Lister Hill Center forBiomedical Communications, the Bumpers Vaccine Center, the Mark Hatfield Clin-ical Research Center, and the Warren Magnuson Clinical Center, the John PorterNeurosciences Building, the Claude Pepper Building, the Natcher Building, the SilvioConte Building, and the Louis Stokes Laboratories. The campus also has a buildingnamed for Mary Lasker, reflecting recognition of the crucial role of disease advocacyin addition to Congressional support.

16 Varmus (2009) observes “Medical research is politically popular, and the oppo-sition party can score points by saying the president is not doing enough for it” (p.144).

Page 5: Mission-oriented biomedical research at the NIH

B.N. Sampat / Research Policy 41 (2012) 1729– 1741 1733

Fig. 2. NIH Funding by Institute or Center: FY2010 versus FY1980. Notes: This chart shows total NIH appropriation by Institute/Center in FY2010 and FY1980. The figureswere converted to 2000 dollars using the GDP Deflator inflation index. The raw data were obtained from the NIH Almanac.28 There are 27 Institute and Centers. The 24Institutes and Centers represented on this chart include the National Cancer Institute (NCI), National Center for Complementary and Alternative Medicine (NCCAM), NationalCenter for Research Resources (NCRR), National Eye Institute (NEI), National Heart, Lung, and Blood Institute (NHLBI), National Human Genome Research Institute (NHGRI),National Institute on Aging (NIA), National Institute on Alcohol Abuse and Alcoholism (NIAAA), National Institute of Allergy and Infectious Diseases (NIAID), National Instituteof Arthritis and Musculoskeletal and Skin Diseases (NIAMS), National Institute of Biomedical Imaging and Bioengineering (NIBIB), Eunice Kennedy Shriver National Instituteof Child Health and Human Development (NICHD), National Institute on Deafness and Other Communication Disorders (NIDCD), National Institute of Dental and CraniofacialResearch (NIDCR), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) National Institute on Drug Abuse (NIDA), National Institute on EnvironmentalHealth Sciences (NIEHS), National Institute of General Medical Sciences (NIGMS), National Institute of Mental Health (NIMH), National Institute on Minority Health andH ke (NI( ormatt ere noN 000), N

ioo2d1

3

atfp

(eaArsAod(2dm

g

It is hard to avoid an unsettling conclusion: despite what NIHsays, it really has no settled priority-setting procedure at all(248).

ealth Disparities (NIMHD), National Institute on Neurological Disorders and StroNLM), and Fogarty International Center (FIC). The chart excludes the Center for Infhrough a central NIH management fund. Several of these Institutes and Centers wIH in 1999), NIMHD (1993), NHGRI (1989), NIAAA (1993), NIAMS (1986), NIBIB (2

ncreasingly since the 1990s Congress has included “soft” earmarks,r non-binding language in reports accompanying the bills urgingr encouraging research on specific diseases (Hegde and Sampat,011; Hegde and Mowery, 2007; Sampat, 2007). Section 4, below,iscusses controversies fueled by the rise of soft earmarking in the990s.

.2. The NIH allocation process

After the House and Senate versions of the bills are reconcilednd the President signs the budget, and each Institute and Cen-er gets a separate appropriation. Grant applications submitted forunding to the NIH go to a central office, and then are assigned toarticular review groups and Institutes/Centers based on fit.

The NIH established its “dual peer review” process in 1947Rettig, 1977). In the first stage, “study sections” made up of sci-ntists rank proposals based on scientific merit, assign scores,nd make funding recommendations. In a second stage, Institutedvisory Councils (generally two-thirds of whom are scientists)eview these applications and scores, and recommend a sub-et to the Institute Director for consideration. While in principledvisory Councils can take into account Institute relevance or pri-rities (IOM, 1998)—including “health” considerations—in practiceeviating from the scientific merit scores appears uncommonChubin and Hackett, 1990; U.S. Government Accountability Office,

009; Strickland, 1989). And while Institute Directors make finalecisions, their decisions tend to follow Advisory Council recom-endations (U.S. Government Accountability Office, 2009).17

28 http://www.nih.gov/about/almanac/appropriations/index.htm.17 Based on the funding available to the Institute in a given year and averagerant costs, overall “paylines” are determined indicating the scores (generally the

NDS), National Institute of Nursing Research (NINR), National Library of Medicineion Technology, Center for Scientific Review, and Clinical Center, which are fundedt part of the NIH in 1980 and thus have no funding in that year: NCCAM (added toIDA (1992), NIDCD (1988), and NINR (1986).

The peer review allocation process is mostly about identifyingand funding the best science. Peer-reviewed, investigator-initiatedgrants are viewed as being “the engine driving our biomedicalresearch machine” (Drazen and Ingelfinger, 2003)—the lifebloodof the NIH and its scientific constituency. In its own priority settingdocument, drafted in response to a 1998 IOM report, the agencyemphasized: “NIH’s highest priority is the funding of basic biomedi-cal research through research project grants.” Strickland (1989) hascharacterized these investigator-initiated grants as “living contin-uing proof of the government’s belief in the freedom of scientificinvestigation.”

For the very same reasons, the process has been criticized for notenough focus on health (IOM, 1998; Brooks, 1978). The scope forresponding to health considerations through investigator-initiated,peer reviewed funding is inherently limited. In the context ofdebates about whether NIH cross-disease funding patterns appro-priately reflect disease burden, Callahan (2003) observes:

percentile of the score distribution) above which applications can be funded. How-ever Institute Directors have discretion to not fund applications above the payline(“skips”) or fund applications below (“exceptions”). Though they have rarely over-ruled the Advisory Council’s recommendations historically, there is some evidencethat funding of low scoring applications has been increasing over time, likelyreflecting NIH’s efforts to promote grants to new investigators (U.S. GovernmentAccountability Office, 2009). Though exceptions appear to affect a small fractionof grants (there is surprisingly little data on this) they do reflect another poten-tial “safety valve” through which the allocation process might respond to healthconsiderations.

Page 6: Mission-oriented biomedical research at the NIH

1734 B.N. Sampat / Research Policy 41 (2012) 1729– 1741

Fig. 3. Distribution of annual percentage change in Institute/Center budgets, 1938–2010. Notes: This chart shows the distribution of annual percent changes in Institute orCenter budgets over the 1938–2010 period, for the 24 Institutes and Centers that receive their own appropriation. Percentage changes are calculated only for years after theI IA), wc l, i.e. ar 00 dol

h(srsiti

3

iWi

ptIga

(tiartracib

Moreover, the report notes that in contrast to investigator initi-

nstitute or Center’s first appropriation. Thus for the National Institute on Aging (Nalculated for all years between 1977 and 2010. There are 904 observations in totaaw data were obtained from the NIH Almanac.29 The figures were converted to 20

Other applications compete outside this peer review process,owever. In particular, Institutes can use requests for applicationsRFAs) to respond to the Congressional “soft” earmarks, or to targetpecific research areas for other reasons. These grants are ofteneviewed by ad hoc advisory councils rather than through thetandard peer review process (Gorden et al., 1993). As I discussn more detail below, in contrast to investigator-initiated grants,hese “solicited” grants also tend to use other grant mechanisms,ncluding research centers and contracts.

.3. Channels

Taking the process as a whole, from Congressional budget mak-ng to NIH allocations, where might health considerations enter?

here does the “health” portion of the NIH mission get operational-zed?

One possibility is in cross-Institute funding choices. For exam-le, if stomach cancer was viewed as an important health priority,he budget of the NCI (National Cancer Institute) could be increased.f research on the common cold were deemed less important in aiven year, the budget of the NIAID (National Institute of Allergicnd Infectious Diseases) could be decreased.

In practice, however Institute budgets increase “in lockstep”Varmus, 2009). Fig. 2 shows a strong correlation between Insti-ute funding levels in 1980 and three decades later, in 2010. Theres one outlier NIAID, which received funding increases in the 1980snd 1990s for AIDS research, and more recently for bioterrorismesearch. This exception may suggest some health priority set-ing based on cross-disease allocations. However, overall the charteveals considerable inertia. This is also seen in Fig. 3, which showsnnual percentage increases in Institute budgets. This distribution,

entered on zero, is a classic illustration of incremental budget-ng. The best predictor of an Institute’s budget in a given year is itsudget the year prior.

29 http://www.nih.gov/about/almanac/appropriations/index.htm.

hich received its first appropriation in fiscal year 1976, annual percent changes aren average of about 37.7 years of data for each of the 24 Institutes and Centers. Thelars using the GDP Deflator inflation index.

Cross-institute funding patterns would also be crude ways topromote or demote funding for specific diseases. Fig. 4 shows theshare of funding by Institute for the ten diseases with the mostfunding in 2008. While some diseases (e.g. cancer, heart disease, eyediseases) have a clear association with particular Institutes, otherstend to draw funding from many different Institutes.18

For all of these reasons priority setting via cross-Institute fund-ing choices would necessarily be very blunt. What about oncefunds are appropriated to each Institute? As I have described itabove, the standard grants process for investigator-initiated appli-cations appears to be driven mostly by scientific considerations.Though Advisory Councils and Institute Directors can tweak rank-ings from grants funded through study section, this is relativelyrare in practice. Accordingly, the capacity for investigator-initiatedgrant funding to reflect any health considerations would appearlimited.

However, a small but non-trivial share of NIH funding is notinvestigator initiated, but instead solicited via requests for appli-cations (RFAs). The IOM report observed, despite the NIH’s heavyemphasis on “basic” research,

Other mechanisms, such as grants for research centers, clinicaltrials, and R&D contracts, tend to be more directed in nature andmore tied to an institute’s mission. They are usually solicited byNIH through requests for applications (RFAs) rather than ini-tiated by extramural scientists, are reviewed by peer reviewcommittees in the Institutes rather than in the Center for Scien-tific Review, and are often funded as cooperative agreements inwhich NIH officials have more of a say in project direction thanthey do with [traditional investigator initiated grants]. They alsotend to be used more in problem-oriented research efforts (19).

ated grants, “Because of their directedness, such mechanisms tendto be specified by Congress in legislation or report language when

18 Not surprisingly, similar analyses of generality of funding by Institute (notshown) suggest that most of the Institutes fund multiple diseases.

Page 7: Mission-oriented biomedical research at the NIH

B.N. Sampat / Research Policy 41 (2012) 1729– 1741 1735

Fig. 4. Share of funds for a disease from each funding Institute or Center, by disease. Notes: This chart shows the distribution of FY2008 funding by Institute or Center, forthe ten diseases with the most total funding, as listed in the Research, Condition, and Disease Categories (RCDC) database. As discussed in detail in Sampat et al. (2011), a“ or Tenn

Cp

tccIcofbc

disease” is an RCDC category that can be reliably linked to categories in the Ninth

ecessarily mutually exclusive.

ongress concludes that NIH should move more quickly to attack aarticular disease or other problem” (19).

Both because they are more political and more applied—thewo are likely correlated—these mechanisms have generated con-erns among the scientific community that they will “crowd out”uriosity-driven, investigator-initiated research (Zerhouni, 2006;OM, 1998). For example, Rettig (1977) notes that the scientificommunity generally dislikes the use of contracts, the “instrument

f directed or targeted research programs” adding that this “stemsrom a difference of view about the most effective way to see thatiomedical research leads to results which can be applied to theonquest of major diseases” (14). He also contrasts these more

th versions of the International Classification of Diseases. RCDC categories are not

directed mechanisms to grants “where the determination of whatto support [is] basically left to the scientific community through theNIH peer-review system and where the direction of where researchshould go [is] largely a function of the prevailing views within thescientific community” (14).

Similarly, grants for centers have been characterized as impor-tant for promoting tackling problems “not being adequatelyaddressed by individual investigator-initiated grants alone” (IOM,

2004, p. 2). These grants are also viewed as popular amongCongress, disease advocates, and taxpayers since they bring“focus” and “visability” to research (p. 2). As with contracts, theresome concern that Centers—particularly when specified through
Page 8: Mission-oriented biomedical research at the NIH

1736 B.N. Sampat / Research Policy 41 (2012) 1729– 1741

Fig. 5. NIH funding by mechanism. Notes: This chart shows total NIH funding by mechanism for the FY2001–FY2010 period, based on data from the NIH Data Book.30 ResearchProject Grants are generally R01 grants, “the original and historically oldest grant mechanism used by NIH.” These grants are traditionally investigator-initiated, though notalways as the next figure shows.

F t showp ProjeT

Cr

ctetwfc

1

process focuses on science, rather than health considerations.

ig. 6. Share of Research Project Grants From Targeted Research. Notes: This chareriod that emanate from targeted research, and the total funding for new Researchargeted research is based on solicitations using requests for applications (RFAs).32

ongressional earmarks—can crowd out investigator-initiatedesearch (IOM, 2004).

Despite these criticisms, solicited grants, and the more targetedontract and center funding mechanisms, are plausible channelshrough which NIH funding choices could reflect health consid-rations. However, they too are limited in practice. Fig. 5 showshe share of NIH funding over the 2001–2009 period associated

ith different funding mechanisms. On average, contracts account

or about 9 percent of total NIH funding over this period, andenters about 10 percent. Fig. 6 shows the share of targeted

30 http://report.nih.gov/NIHDatabook/Charts/Default.aspx? showm=Y&chartId=53&catId=1.

s the share of new Research Project Grants awarded during the FY2000–FY2010ct Grants associated with targeted research. The data are from the NIH Data Book.31

grants—emanating from RFAs—in all new research grants over thesame period. Over this period of time, about 12 percent of all newgrants were targeted, accounting for about 20 percent of NIH fund-ing.

Thus the bulk of new grants and funding were for investigator-initiated, non-solicited grants. By and large, the NIH allocation

However the more targeted mechanisms, and research solicita-tions, may nevertheless have an important, under-appreciated role

31 http://report.nih.gov/NIHDatabook/Charts/Default.aspx? showm=Y&chartId=25&catId=2.

32 http://www.niaid.nih.gov/researchfunding/glossary/pages/t.aspx.

Page 9: Mission-oriented biomedical research at the NIH

Policy

atioC

4

rmtIitfi1lcs

4

1m“iPi

u(CNdmppn

ahirA

CaSpor

and more layperson representation in research allocation choices,which advocates hoped would be more feasible with a separateagency (Strickland, 1972).

19 In addition to concerns about the science and the technical feasibility of theproject, questions also were raised about the appropriateness of support of activities

B.N. Sampat / Research

s “safety valves” that, on occasion, allow the agency to pursue moreargeted inquiry than is possible through standard peer-reviewednvestigator-initiated grants. As I show in the next section, this isften in response to pressure to do so from disease advocates andongress.

. Tensions in mission-oriented biomedical research

Debates about the relative value of “basic” and “applied”esearch, and over serendipity versus planning, have been in com-on in the NIH’s history, as was evident in the broad overviews of

he agency’s and funding processes presented above. In this section, show that these themes are also present in important episodesn the history of the NIH I detail below: “big push” funding forechnology development and disease-specific research (the Arti-cial Heart Program and the War on Cancer) during the 1960s and970s, and the controversy during the 1990s surrounding disease

obbying, Congressional earmarking, and NIH priority setting. Thesease studies also illustrate the roles of different interest groups andtakeholders in shaping and sustaining support for the agency.

.1. The artificial heart program

The NIH’s Artificial Heart program was launched in the mid-960s to develop a totally implantable device (one with no externalachinery) for patients with advanced heart disease. Like other

big-push” mission oriented R&D programs (Mowery et al., 2010)ncluding the War on Cancer discussed below, the Artificial Heartrogram was motivated in part by the success of the Apollo programn successfully landing a man on the moon.

Before the public sector effort began, research was alreadynderway on artificial heart technology in the private sectorMorlacchi and Nelson, 2010) and lobbying by researchers ledongress to thrust a program (with considerable funding) onto theational Heart, Lung, and Blood Institute (NHLBI). This occurredespite concerns by NIH establishment that the level of funda-ental understanding at the time may not have warranted such

rodigious funding (Straus, 1984). Much of the motivation for therogram was that heart disease was then (as it is now) the nation’sumber one killer.

The artificial heart program was targeted towards a deliver-ble outcome: artificial heart devices for clinical use. It is unclearow the agency could have achieved this goal through standard

nvestigator-initiated peer-reviewed grants. Thus the project had toely on of the “safety valve” mechanisms, contracts. A 1991 Nationalcademy of Sciences publication noted:

As with other component of the National Institutes of Health(NIH), the principal mechanism of achieving NHLBI’s over-all mission is the funding of extramural research throughinvestigator-initiated, nontargeted (“R01”) grants. Most of theinstitutes that make up the NIH do not fund later developmentalstages of medical technologies, focusing instead on fundamentalor basic research . . . the NHLBI artificial heart program is, how-ever, a notable exception to this generalization. Historically, thefunding mechanism for R&D [for artificial heart research] hasbeen contracts, issued following requests for proposals (21).

The case also reveals the complex relationship betweenongress, interest groups, and the agency. In 1967, when Congress

llocated more funding to the program than NIH Director Jameshannon believed was desirable, he responded by re-labeling therogram the “Artificial Heart-Myocradial Infarction Program” andpened up a substantial share of the funds for other heart-relatedesearch (Straus, 1984). And when, in the mid-1980s, the NIH tried

41 (2012) 1729– 1741 1737

to shut down the program in light of disappointing progress,19

the efforts were blocked by powerful Senators (Ted Kennedy andOrrin Hatch) whose states received considerable funding throughthe program (“Senators Doctor Kennedy and Hatch”, 1988; U.S.National Academy of Sciences, 1991).20 In response to the concernsabout limited progress, eventually the NIH began to shift fund-ing away from the total artificial heart towards ventricular assistdevices (Morlacchi and Nelson, 2010).

The case also exposes real differences of opinion in the valueof basic versus targeted research. It also shows the value thatthe “safety valve” mechanisms can and do have in respondingto political pressures for more directed research, which are diffi-cult to accommodate through standard investigator-initiated peerreviewed grants. Similar dynamics are seen in the War on Cancer,another targeted research program, discussed below.

4.2. The war on cancer

The NIH had been involved in cancer research since early in thetwentieth century. The National Cancer Institute (NCI) was foundedin 1937 and became part of the NIH, although its research effortswere interrupted by the Second World War. Beginning in the 1950s,disease advocates including Mary Lasker began strategizing for anew War on Cancer (Mukherjee, 2010), and momentum for such aninitiative intensified in the 1960s. As with the artificial heart pro-gram, here too advocates were influenced by the Apollo program,and called for a “moon shot” for cancer.

Several forces combined to create momentum for new cancerresearch, including claims that the disease was underfunded rela-tive to its burden and scientific opportunity (Strickland, 1972) andgrowing popular fears about the disease (Rettig, 1977). Anticipat-ing more recent controversies about cross-disease funding patternsthat I will discuss below, the War on Cancer created tensionsbetween disease groups, with cardiovascular disease advocatesquestioning why cancer should get a major infusion of new funds,when heart disease was the nation’s primary source of mortality(Rettig, 1977).

Another major impetus was concern, among Mary Lasker andother disease advocates, that the NIH peer review process wasnot equipped to target a specific disease (Strickland, 1972; Rettig,1977). Previous difficulties encountered by NCI administratorsattempting to employ the contract mechanism, instead of grants,to target specific research, helped fuel these concerns that thepeer review process “boxed in” NCI bureaucrats who wanted torapidly tackle the disease (Strickland, 1972, p. 268). Specifically,there was a perception that “NIH was more concerned with medicalresearch than with finding cures” (Rettig, 1977, p. 306). Accord-ingly, the interest groups pushing for increased cancer funding alsorecommended establishing a new organization separate from theNIH—“a goal driven institute that would decisively move towardfinding a cancer cure” (Mukherjee 2010, p. 117)—to be in chargeof cancer research. Echoing themes from the Bush-Kilgore debatesa quarter century prior, there were further calls for more directaccountability of cancer researchers to politicians and taxpayers,

that the private sector already undertaking (NAS, 1991), and about the cost-effectiveness of the new technology relative to other heart disease treatments.

20 Specifically, the Senators threatened to curtail other funding for the agency,which led the New York Times to accuse the legislators of “terrorizing” the agencyand interfering with internal NIH priority setting “in a way that lards their own porkbarrels.” From “Senator Doctors Kennedy and Hatch” July 15, 1988.

Page 10: Mission-oriented biomedical research at the NIH

1 Policy

rimompsbr1toi

rlfadehcwwrir

IDpSoermiboDMNc

sfictdaedato

t(l

Ao

738 B.N. Sampat / Research

As Rettig (1977) has chronicled, initiatives for increasing canceresearch in a new agency were embraced by powerful legislatorsncluding Edward Kennedy (D–MA). In part to seize the political

omentum away from Kennedy, President Nixon in his 1971 Statef the Union address famously announced executive branch com-itment to the War on Cancer, invoking the successes of other big

ush projects (splitting the atom, and putting a man on the moon) inupport of the argument for increased funding. Nixon, influencedy NIH officials and his science advisor, did not support a sepa-ate agency, but rather a major boost in funding to the NCI (Rettig,977). This spurred a debate between the Executive and Legisla-ive branches (and also between important members of the Housef Representatives and the Senate) on whether the NIH was the bestnstitutional locus for waging a war on cancer.

Perhaps unsurprisingly, the proposal for a separate canceresearch agency drew opposition from the NIH. On one hand, NIHeadership was concerned that a decoupling of cancer researchrom the agency would reduce the agency’s funding (Cohn, 1970)nd would be followed by similar initiatives for the other majoriseases (Rettig, 1977). And without housing major disease ori-nted research, they may have realized, funding for the NIH wouldave dwindled. Strickland (1972) notes: “If the Conquest of Can-er Agency were established, a Conquest of Heart Disease Agencyould be next, other separations would surely follow, and whatas left of the National Institutes of Health would shrink into

elative insignificance” (278). It would be difficult to generate polit-cal support or command significant funding for pure biomedicalesearch, absent explicit links to the health mission.

The serendipity hypothesis was also important in these debates.n making a case for keeping cancer research within the agency, NIHirector Robert Marston argued “cancer research profits from inter-lay with other research in the diseases of man” (Cohn, 1970, A2).enator Gaylord Nelson (D-Wisconsin) also defending the existingrganizational structure, similarly claimed “All of the major discov-ries, including ones which fell out inadvertently from non canceresearch, have occurred largely because of the present broad-based,ulti-disciplinary system of federally supported research embod-

ed in the NIH” (quoted by Rettig, 1977, p. 211). The NIH alsorought to Congress numerous prominent scientists who testifiedn the importance of serendipity in their own cancer research.avid Baltimore, who in 1975 would win the Nobel Prize inedicine, discussed how virus research originally funded by theational Institute for Allergic and Infectious Diseases (NIAID) wasrucial for breakthroughs in cancer.21

In part due to the strength of this argument (but also a con-equence of other political forces chronicled by Rettig, 1977) thenal legislation, the National Cancer Act of 1971, housed new can-er research at the NCI rather than a separate agency. However,he Act required more direct reporting by the agency to the Presi-ent, created a National Cancer Advisory Board (NCAB) which wasppointed by the President to help the NCI Director in planning, andstablished a “bypass” budget through for the NCI to be submittedirectly to the President, rather than as part of the NIH budget as

whole. Between 1970 and 1980 NCI funding grew nearly threeimes in real terms, twice the rate of increase for the rest of the NIHver the same period.22

The Act also aimed to address concerns, discussed above, about

he ability of traditional grant mechanisms to target researchRettig, 1977). The Committee Report for the legislation estab-ishing the Act noted that “the extraordinary demands placed on

21 Illustrating the importance of the agency’s broad constituencies, the Americanssociation of Medical Colleges, among other academic societies, was also active inpposing major changes to the existing NIH structure (Strickland, 1979).22 http://www.nih.gov/about/almanac/appropriations/index.htm.

41 (2012) 1729– 1741

this new cancer program may require the creation of new mecha-nisms of peer review.” Thus the Act also allowed and encouragedthe NCI to use funding mechanisms that were more targeted thanextramural research grants, including cancer centers and researchcontracts. Despite this relatively more targeted approach, however,fundamental research allocated through peer-reviewed grants didreceive substantial additional funding associated with the War onCancer. This in turn triggered the classic debates about whethermore “basic” or “applied” mechanisms were more effective attackling disease, and whether science can or should be planned(Schmidt, 1977).

4.3. Priority setting

Similar themes are present in discussions about the appropri-ateness of NIH cross-disease allocation patterns, which date backto its early history (Strickland, 1979; Rettig, 1977). Callahan (2003)recalls that, during the 1948–1968 period, “the standing joke wasthat Congress supported research on diseases that killed old men,just like those in Congress” (237).

The most recent episode in this saga occurred in the early1990s, with concerns in Congress, in the scientific community,and among some disease advocates that AIDS, Breast Cancer andother high-profile diseases got more than they should given theirsocietal burden, because of the strength of the lobbies for these dis-eases (Johnson, 1998). Related to this, other disease organizationshad begun to lobby aggressively for “soft” earmarks in the mid-1990s (as noted above, these earmarks are generally in the formof report language attached to appropriations bills, often aiming toget RFAs issued for particular diseases and to bypass the traditionalpeer review process). The debates over earmarking reflected thetensions between science and politics described above, includingwhether the peer review process was narrowly focused on science(Sampat, 2007), and whether research can or should be targeted toparticular diseases. The debate also raised important equity issues(e.g. whether the peer review process is more fair than politicallyguided allocation, if the latter would necessarily cater to powerfuldisease interest groups).

All of these concerns led Congress to commission a report fromthe Institute of Medicine (IOM) on the NIH’s internal priority set-ting procedures. The resulting IOM report, issued in 1998, focusedspecifically on concerns that the NIH did not pay sufficient atten-tion to the burden of disease, and cared more about science. Italso addressed the related concern that certain salient diseasescommanded more resources than they ought to. Despite being gen-erally enthusiastic about the NIH, the IOM report observed (in itssummary Introduction): “There is a sense that NIH has evolvedmechanisms for judging scientific opportunity and merit that sur-pass its capabilities for assessing and being influenced by publichealth needs” (8). These concerns also led the NIH to prepare awhite paper on priority setting (NIH, 1997) and prompted the NIHDirector to provide testimony on NIH priority setting before theHouse Appropriations Committee, (Varmus, 1999).

In a statement echoing Vannevar Bush’s report more than ahalf-century prior, the Director emphasized serendipity to deflectconcerns about the agency’s cross-disease allocation choices, argu-ing that “There are legitimate limits to our ability to plan science.Because science attempts to discover what is unknown, it is inher-ently unpredictable.” He continued, “History has repeatedly shownthe benefits of allowing a significant portion of our research activityto be governed by the imagination and productivity of individualscientists” (Varmus, 1999).

In addition to planning, the Director argued that serendipityalso complicates disease categorization, and questioned the dataused in calculations suggesting a mismatch between funding lev-els and disease burden. Specifically, he noted in previous testimony

Page 11: Mission-oriented biomedical research at the NIH

Policy

(aowsca

wfiI(aaa

stCdt

5

nrfha

pioChmobtsbibaftsit

fiqpCr

iu

iN

and tensions between the health and science missions—presentthroughout the agency’s history—have continued force.

B.N. Sampat / Research

Varmus, 1997) “much of the research done by each IC [Institutend Center] may be difficult or impossible to explain as a partf a research plan against a specific disease. Nevertheless, suchork—on, for example, the pathways cells use to interpret chemical

ignals, the three-dimensional structures of proteins, or the pro-esses of cell death—could ultimately form the basis for practicaldvances against any of several diseases.”23

Varmus (1997) also emphasized the limited channels throughhich the NIH might be responsive to health priorities in ane-grained way, noting that Congressional choices about cross-

nstitute allocations, the considerable inertia in the processincluding funds already committed due to previous grants), and

commitment to investigator-initiated research “leave only a rel-tively small fraction of each year’s appropriation that can beffected by changes in funding policies.”

These arguments appear to have been persuasive. Despite thetrong tensions between the science and health missions of the NIHhat were evident in these debates, the NIH, powerful members ofongress, disease and science advocacy groups were successful inoubling the agency’s budget over the five years that followed thisestimony, as noted above.

. Conclusion

I began with a puzzle: how has the NIH been so successful in gar-ering political and public support, and large budgets, for scientificesearch? The arguments that the NIH’s own allocation process isocused more on science than health, and that the tensions betweenealth and science missions have been common throughout thegency’s history, complicate this puzzle further.

My analysis suggests three factors that may explain this sup-ort. First, there are various “safety valves” in the allocation process,

ncluding RFAs, that allow it to target a small but significant portionf its budget in response to specific health considerations (and/orongressional directives, which may or may not be correlated withealth priorities). Cross institute funding decisions by Congressay also play a role in aligning funding patterns with health pri-

rities, though for reasons discussed above these are necessarilylunt. Second, while it has occasionally come under question, overhe long run the serendipity hypothesis appears to have been per-uasive in convincing Congress and taxpayers that the dichotomyetween health and science is a false one: that the best route to

mproved health is fundamental science.24 Third, the NIH has aroad range of constituents—patient groups, private sector drugnd device companies, and universities and hospitals benefittingrom NIH support—scattered across states and Congressional dis-ricts, crucial to sustaining support for the agency. While I cannotay for sure that these factors are absent at other science fund-ng agencies, I believe they have been particularly important athe NIH.

Certainly another explanation for the unique political supportor the NIH is a widespread belief that the model works, i.e.mproves human health. At the same time, there has been recurrent

uestioning of the returns to publicly funded biomedical research,erhaps most prominent in discussions of the whether the War onancer was successful (Brennan et al., 2010). These concerns areeflected in numerous attempts throughout the agency’s history

23 Elsewhere in this testimony he provided a more succinct version of the serendip-ty argument: “Research aimed in one direction frequently provides benefits in annexpected direction” (Varmus, 1997).24 Though the idea of serendipity has been persuasive politically, there is surpris-ngly little large-sample evidence on the frequency of serendipitous discovery inIH-funded research.

41 (2012) 1729– 1741 1739

to estimate these returns (e.g. Mushkin, 1979; Sampat, 2011).25

There are related questions about whether technological advancein medicine in “worth it” given the high costs of new medicaltechnologies, recognizing they are the main sources of long-runincreases in health care costs (Callahan, 2003).

These concerns have occasionally generated calls to re-examinethe effects of upstream funding choices at the NIH on patternsof technology development and health costs, and to use NIHresearch to promote development of cost-reducing (rather thancost-increasing) technologies (Mushkin, 1979). However, if myanalysis above is correct, this is unlikely under the agency’scurrent grant structure, since it is much better positioned to influ-ence the rate rather than the direction of technical change inmedicine.

Support for the NIH may also reflect uniquely American atti-tudes towards medical technology. Survey research (Kim et al.,2001) suggests that about two thirds of Americans are “very inter-ested in” new medical discoveries, compared to 44 percent acrossEurope. However, these attitudes may themselves reflect the largescale and scope of the effects of NIH funding, i.e. the causal arrowmay go the other way as well. Indeed, the U.S. spends more ongovernment health R&D as a share of national income than anyother OECD country (OECD, 2007). Nonetheless, even in one of thecountries whose citizens are less sanguine about medical tech-nology, the United Kingdom, similar questions about health andscience (and basic research versus targeting) have been asked ofits major government funder of medical research, the MedicalResearch Council (Mowery et al., 2010; Bryder, 1989). These ten-sions may well be inherent to mission oriented health research.

While much of the discussion above was historical, the verysame themes are being debated today. For example the NIHRoadmap unveiled by Director Zerhouni in 2003 aimed to tacklespecific diseases and technical problems, relying heavily on RFAcontracts, and other “targeted” mechanisms to steer the researchagenda (Zerhouni, 2003). In response, prominent members of thescientific community objected, on the ground that this would dis-tort resources away from higher productivity investigator initiatedgrants on which basic science depends, and which, it was argued,are the surer path to medical progress (Check, 2006). Zerhouniand his successor Francis Collins (who became NIH Director in2009) also have emphasized funding for so-called “translational”research, which aims to bridge basic research and clinical appli-cation. In response to concerns from the scientific communitythat these efforts would reduce funding for investigator-initiatedresearch, Collins argued that these more applied initiatives “get a lotof traction with the Congress, with the public” but also emphasizedhis desire “to protect the basic science foundation.. because that isour future” (Kaiser et al., 2010).26 The serendipity versus outcomesdebate, the tradeoffs between funding basic and applied research,

25 Sampat (2011) argues that collectively, there is strong evidence that the NIHhas had an important role in drug and device development. The NIH’s applied activ-ities (funding of trials, consensus development conferences) appear to have beenimportant in the cardiovascular arena, the disease where there has been the mostprogress in the postwar era. There is less convincing direct evidence on the effects ofits fundamental research activities on population health outcomes. This may reflecta range of difficulties in evaluating the returns to “basic” research in particular,including long lags, difficulties in linking inputs to outcomes, and the absence ofclear counterfactuals.

26 Recently Collins spearheaded the establishment of the National Center forAdvancing Translational Scientists (NCATS) to help speed drug development.Here again important voices in the scientific community expressed concerns thatthese efforts were inappropriate for the NIH and could undermine basic research(Weissman, 2011).

Page 12: Mission-oriented biomedical research at the NIH

1 Policy

A

tMdI

R

A

A

B

B

BB

B

B

C

CC

C

C

C

D

D

D

G

G

H

HH

H

H

I

I

J

K

K

740 B.N. Sampat / Research

cknowledgments

I thank Larry Brown, Dominique Foray, Sherry Glied, David Gus-on, David Mowery, Richard Nelson, Sejal Patel, Daniel Sarewitz,

ichael Sparer, and Jennifer Washburn for comments on a previousraft. This work was funded by a Robert Wood Johnson Foundation

nvestigator Award in Health Policy Research.

eferences

merican Association for the Advancement of Science (AAAS), 2012. AAAS ReportXXXVII: Research and Development FY2013 (downloaded on 17 September2012 from http://www.aaas.org/spp/rd/rdreport2013/).

zoulay, P., Graff Zivin, J., Manso, G., 2011. Incentives and creativity: evidence fromthe life sciences. Rand Journal of Economics 42, 527–554.

hattacharya, J., Packalen, M., 2011. Opportunities and benefits as determi-nants of the direction of scientific research. Journal of Health Economics 30,603–615.

rennan, R., Federico, S., Dyer, M., 2010. The war on cancer: have we won the battlebut lost the war? Oncotarget 1, 77–83.

rooks, H., 1978. The problem of research priorities. Daedalus 107, 171–190.rooks, H., 1995. The evolution of U.S. science policy. In: Smith, B.L.R., Barfield,

C.E. (Eds.), Technology, R&D, and the Economy. The Brookings Institution andAmerican Enterprise Institute, Washington, DC, pp. 15–48.

ryder, L., 1989. Tuberculosis and the MR. In: Austoker, J., Bryder, L. (Eds.), His-torical Perspectives on the Role of the MRC. Oxford University Press, Oxford,pp. 1–22.

ush, V., 1945. Science The Endless Frontier: A Report to the President. GovernmentPrinting Office, Washington, DC.

allahan, D., 2003. What Price Better Health? Hazards of the Research Imperative.University of California Press, Berkeley and Los Angeles, CA.

heck, E., 2006. Biomedical research: facing the opposition. Nature 441, 17–19.hubin, D., Hackett, E., 1990. Peerless Science: Peer Review and U.S. Science Policy.

SUNY Press, Albany, NY.ook-Deegan, R., McGeary, M., 2006. The jewel in the federal crown? History, poli-

tics, and the National Institutes of Health. In: Stevens, R.A., Rosenberg, C.E., Burns,L.R. (Eds.), History and Health Policy in the United States: Putting the Past Back.Rutgers University Press, Piscataway, NJ, pp. 176–201.

ohn, V., 1970. Multi-billion cancer war by a new agency urged; study urges separatecancer agency. Washington Post, December 5, 1970: A1–A2.

utler, D., Meara, E. Richards, S., 2009. Induced Innovation and Social Inequality:Evidence From Infant Medical Care. NBER Working Paper 15316.

razen, J.M., Ingelfinger, J.R., 2003. Grants, litics, and the NIH. The New EnglandJournal of Medicine 349, 2259–2261.

rew, E., 1967. The health syndicate: Washington’s noble conspirators. AtlanticMonthly 220, 75–82.

orsey, E.R., de Roulet, J., Thompson, J.P., Reminick, J.I., Thai, A., White-Stellato,Z., Beck, C.A., George, B.P., Moses III, H., 2010. Funding of US biomedi-cal research 2003–2008. Journal of the American Medical Association 303,137–143.

orden, P., Cyphers, D.F., Feld, C., 1993. Budget trends and issues affecting biomedicalresearch: a perspective from the National Institute of Diabetes and Digestive andKidney Diseases. Hepatology 18, 677–687.

ross, C.P., Anderson, G.F., Powe, N.R., 1999. The relation between funding by theNational Institutes of Health and the burden of disease. The New England Journalof Medicine 340, 1881–1887.

arden, V., 1986. Inventing the NIH: Federal Biomedical Research Policy 1887–1937.Johns Hopkins University Press, Baltimore, MD.

arden, V., 2012. AIDS at 30: A History. Potomac Press, Washington, DC.arris Interactive. 2007. CDC, FAA, NIH, FDA, FBI and USDA Get the Highest Ratings

of Thirteen Federal Government Agencies (downloaded on 17 September2012 from http://www.harrisinteractive.com/vault/Harris-Interactive-Poll-Research-Government-Agencies-2007-02.pdf).

egde, D., Mowery, D., 2007. Politics and funding in the U.S. public biomedical R&Dsystem. Science 322, 1797–1798.

egde, D., Sampat, B., 2011. The Political Economy of Publicly Funded Biomed-ical Research: Evidence from NIH Funding for Rare Diseases. Workingpaper (downloaded on 17 September 2012 from http://politics.as.nyu.edu/docs/IO/21318/Hegde.pdf).

nstitute of Medicine, 1998. Scientific Opportunity and Public Needs: ImprovingPriority Setting and Public Input at the National Institutes of Health. NationalAcademies Press, Washington, DC.

nstitute of Medicine, 2004. NIH Extramural Center Programs: Criteria for Initiationand Evaluation. National Academies Press, Washington, DC.

ohnson, J.A., 1998. Disease Funding and NIH Priority Setting. Congressional ResearchService Report for Congress, Report 97–917. Congressional Research Service,

Washington, DC.

aiser, J., Marshall, E., Zahn, L., 2010. An interview with Francis Collins at the NIH.Science 328, 1090–1091.

evles, D., 1977. The national science foundation and the debate over postwarresearch policy. Isis 68, 4–26.

41 (2012) 1729– 1741

Kim, M., Blendon, R.J., Benson, J.M., 2001. How interested are Americans in newmedical technologies? A multicountry comparison. Health Affairs 20, 194–201.

Lichtenberg, F.R., 2001. The allocation of publicly funded biomedical research.In: Berndt, E., Cutler, D. (Eds.), Medical Care Output and Productiv-ity. Studies in Income and Wealth University of Chicago Press, Chicago,pp. 565–589.

Moran, M., Guzman, J., Ropars, A.L., McDonald, A., Jameson, N., Omune, B., Ryan, S.,Wu, L., 2009. Neglected disease research and development: how much are wereally spending? PLoS Medicine 6, e1000030.

Morlacchi, P., Nelson, R.R., 2010. How medical practice evolves: the case of the leftventricular assist device. Research Policy 40, 511–525.

Moses III, H., Dorsey, E.R., Matheson, D.H.M., Their, S.O., 2005. Financial anatomyof biomedical research. Journal of the American Medical Association 294,1333–1342.

Mowery, D.C., Nelson, R.R., Martin, B.R., 2010. Technology policy and global warm-ing: why new policy models are needed (or why putting new wine in old bottleswon’t work). Research Policy 3, 1011–1023.

Mowery, D., 1997. The Bush report after fifty years: blueprint or relic? In: Barfield,C. (Ed.), Science for the Twenty-First Century: The Bush Report Revisited. AEIPress, Washington, DC, pp. 24–40.

Mukherjee, S., 2010. The Emperor of All Maladies: A Biography of Cancer. Simon andSchuster, New York.

Mushkin, S., 1979. Biomedical Research: Costs and Benefits. Ballinger Publishing,Cambridge, MA.

National Institutes of Health, 1997. Setting Research Priorities at the National Insti-tutes of Health, NIH publication no. 97-4265. National Institutes of Health,Bethesda, MD.

National Science Foundation (NSF), 2010. Science and Engineering Indicators2010 (downloaded on 17 September 2012 from http://www.nsf.gov/statistics/seind10/pdf/seind10.pdf).

OECD, 2007. Science, Technology, and Industry Scoreboard 2007: Innovation andPerformance in the Global Economy.

Rettig, R.A., 1977. Cancer Crusade: The Story of the National Cancer Act of 1971.Princeton University Press, Princeton, NJ.

Sampat, B., 2007. The dismal science, the crown jewel, and the endless frontier.In: Foray, D. (Ed.), The New Economics of Technology Policy. Edward Elgar,Cheltenham, UK, Northampton, MA, pp. 148–162.

Sampat, B., 2011. The impact of publicly funded biomedical and health research:a review. In: Olson, S., Merrill, S. (Eds.), Measuring the Impacts of Fed-eral Investments in Research. National Academy Press, Washington, DC,pp. 153–192.

Sampat, B., Buterbaugh, K., Perl, M., 2011. New Evidence on the Allocation of NIHFunds Across Disease Areas. Working Paper.

Schmidt, B., 1977. Five years into the national cancer program. The Yale Journal ofBiology and Medicine 50, 237–244.

Schmookler, J., 1965. Catastrophe and utilitarianism in the development of basicscience. In: Tybout, R. (Ed.), Economics of Research and Development. Ohio StateUniversity Press, Columbus, OH, pp. 19–33.

“Senators Doctor Kennedy and Hatch,” Opinion Page, New York Times July 15, 1988.Shapley, W., 1992. The Budget Process and R&D. Carnegie Commission on Science,

Technology, and Government, New York.Starr, P., 1984. The Social Transformation of American Medicine: The Rise of a

Sovereign Profession and the Making of a Vast Industry. Basic Books, New York.Stokes, D.E., 1997. Pasteur’s Quadrant: Basic Science and Tech-

nological Innovation. Brookings Institution Press, Washington,DC.

Straus, M.J., 1984. The political history of the artificial heart. New England Journalof Medicine 310, 332–336.

Strickland, S.P., 1972. Politics, Science, and Dread Disease: A Short History ofUnited States Medical Research Policy. Harvard University Press, Cambridge,MA.

Strickland, S.P., 1989. Story of the NIH Grants Program. University Press of America,Lanham, MD.

Tauzin, B., 2008. The next 50 years of medical innovation has alreadystarted. Pharmaceutical Research and Manufacturers of America, Washington,DC.

U.S. Government Accountability Office, 2009. National Institutes of Health: Comple-tion of Comprehensive Risk Management Program Essential to Effective Over-sight (downloaded on 17 September 2012 from http://www.gao.gov/assets/300/295085.pdf).

U.S. National Academy of Sciences, 1991. The Artificial Heart: Prototypes, Policiesand Patients. National Academy Press, Washington, DC.

U.S. National Research Council and Institute of Medicine, 2003. Enhancing the Vital-ity of the National Institutes of Health. National Academy Press, Washington,DC.

Weissman, G., 2011. Is drug development too slow? NIH to the rescue! The FASEBJournal 25, 1119–1122.

Varmus, H., 1997. Testimony on Setting Research Priorities at NIH before theHouse Committee on Appropriations, Subcommittee on Labor, HHS, Educa-tion.(downloaded on 17 September 2012 from http://www.hhs.gov/asl/testify/

t970610a.html).

Varmus, H., 1999. Statement on Funding Allocation for Disease Research Beforethe Senate Committee on Appropriations, Subcommittee on Labor, HHS, Educa-tion (downloaded on 17 September 2012 from http://www.hhs.gov/asl/testify/t990506a.html).

Page 13: Mission-oriented biomedical research at the NIH

Policy

V

VW

W

B.N. Sampat / Research

armus, H., 2001. Proliferation of National Institutes of Health. Science 291,1903–1905.

armus, H., 2009. The Art and Politics of Science, 1st ed. W.W. Norton, New York.adman, M., 2007. Biomedical philanthropy: state of the donation. Nature 447,

248–250.ildavsky, A., 1961. Political implications of budgetary reform. Public Administra-

tion Review 21, 183–190.

41 (2012) 1729– 1741 1741

Zeckhauser, R., 1967. Some Thoughts on the Allocation of Resources to Bio-Medical Research. U.S. Department of Health, Education, and Welfare, Office

of the Assistant for Planning and Evaluation, Occasional Paper No. 4,November 9.

Zerhouni, E.A., 2006. NIH in the Post-Doubling Era: Realities, Strategies. Science,1088–1090.

Zerhouni, E., 2003. The NIH Roadmap. Science 302, 63–72.