The biomedical workforce report (TRR-I)

Important? Yes. Flawed? Yes. Must we do better? Yes. So let’s give it a try.

Long-awaited, the NIH-sponsored report on US biomedical

Let’s give it a try! (Picture from Raisintoast)

research training finally appeared in June (1). Late in 2010 Francis Collins, director of NIH, gave marching orders to a panel of experts—the Biomedical Research Workforce Working Group (BRWWG), which is part of the NIH Advisory Committee to the Director. Replacing that ugly acronym with the names of the panel’s co-chairs (2), I shall call it the Tilghman-Rockey Report, or TRR. Collins asked the panel to develop “a model for a sustainable and diverse biomedical research workforce,” to help determine how many and what kinds of people the US should train to “advance science and promote health.” On the basis of this model, but “recognizing that there are limits to NIH’s ability to control the training pipeline,” the panel was asked to recommend actions “NIH should take to support a future sustainable biomedical research infrastructure” (1).

Many of us hoped the TRR would answer grave questions about PhD and postdoctoral training and outline a clear path to a better future. This report does no such thing. Parts of it are useful, but on the whole the report proves to be something of a disappointment, despite generous arrays of facts and recommendations. Many sets of critical facts, the report scrupulously admits, are unavailable, incomplete or wildly unreliable. Of TRR’s recommendations, the most crucial often come across as timid, feckless, or aimed at the wrong target.

These deficiencies reflect no lack of experience, intelligence, or diligence on the part of the report’s framers. Some stem from troubling data gaps, especially with respect to the population size and career paths of groups whose trajectories are not adequately tracked—for instance, postdocs who are not US citizens and graduate students not supported by NIH training grants. Other deficiencies result from tight constraints. Thus Collins’s charge confined the recommendations to actions within NIH’s mandate, and the working group explicitly chose not to detail how its recommendations should be implemented, pleading lack of adequate time and resources to tackle such complex issues.

To my mind, the TRR suffers from two more serious flaws: (i) failure to connect dots between proposed recommendations for change and the root causes of problems those proposals are designed to fix; (ii) too narrow a focus on training, despite its charge to recommend ways to make the entire biomedical research workforce sustainable. Blurring cause-effect connections weakened the report’s impact by permitting key recommendations to pull their punches, and the narrow focus led to ignoring obvious additional connections between training difficulties and other problems that bedevil research centers, research administrators, funding agencies, PIs, and working laboratories. Ignoring the latter connections not only obscures rationales for proposed remedies, but also systematically limits those remedies to unilateral actions by the NIH—despite the fact that every stakeholder in the biomedical research enterprise shares partial responsibility for its workforce problems, which cannot be fixed without cooperation among all the stakeholders (3-5).

By way of clearing the deck before we turn to specific concerns and recommendations of the TRR, let me remind readers of certain workforce issues that are too important to ignore. Conversations with PIs, administrators, and leaders of funding agencies, often revolve about this short litany of 21st century problems afflicting US biomedical research:

  1. Ever-fiercer competition among; scientists (both well established and beginners) wrangling for publications, positions, promotions, trainees, and grants.
  2. Similarly intense—and increasing—competition among research institutions vying for prestige, money, and scientists.
  3. Arbitrary evaluation of individuals, research findings, ideas, and proposed projects, increasingly focused on marginal rather than core qualities: e.g., indirect costs vs. creative science; fancy journals vs. original work; reams of corroborating data vs. new data that disturbs hoary assumptions (6).
  4. Progressive graying of PIs. Beginning PIs apply for their first grant in their forties, while their elders control large resources and distribution of rewards.
  5. A growing tendency of the best college seniors to opt for scientific careers outside biomedical research—especially distressing because it contrasts sharply with so many opportunities for new discoveries and real-world applications.

Each of these troubles is too important to ignore, because all of them profoundly affect PIs, trainees, and all other members of the US biomedical research workforce. Troubles 1-3, however, do not figure in the TRR and numbers 4 and 5 receive short shrift there. But all five can be traced, at least in part, to a single cause, addiction of the US biomedical research enterprise to rampant expansion (4). Expansionism was a natural response to steady increases in NIH funding (averaging 8% per year, 1970-1998, higher in 1999-2003; 7). Its first and best-known consequence was a long spate of exciting discoveries and clinical advances that benefitted scientists, patients, and the entire nation. At the same time, biomedical investigators and research institutions became gradually addicted to expansion, and in turn that expansion exacerbated our litany of troubles. The TRR fails to mention expansionism, and—with a single exception, noted below—ignores its most disturbing consequences. In contrast, Biomedwatch’s discussions of TRR recommendations will often remind us of the role expansionism plays in causing the problem targeted by a particular recommendation.

The TRR must have been hard to write, because it tackles many problems, all multi-faceted, interconnected, and replete with unavoidable details that must be explained to most audiences—including biomedical PIs and research administrators. In analyzing the 156-page TRR, Biomedwatch faces the same difficulties. Because the problems, their potential solutions, and the report itself are too important and complex for a single post, I shall divide my comments into separate parcels. Most succeeding posts will detail specific problems with the biomedical workforce and its training, analyze the TRR’s recommendations with respect to training, and amend or extend certain recommendations to make them more effective. Our next foray into TRR-land will focus on the exception noted above: soft-money salaries for research faculty, a major contributor to expansionism and a problem badly in need of a remedy.


1. Biomedical Research Workforce Working Group Report. Pdf here.

2. The working group’s co-chairs were Shirley Tilghman, a molecular biologist and president of Princeton, and Sally Rockey, NIH’s deputy director for extramural research. The document refers to itself as the Biomedical Research Workforce Working Group Report, or BRWWGR.

3. HR Bourne and MO Lively, Iceberg Alert for NIH, in Science Express (for pdf, see the “Must read” section on this blogsite’s main page. This editorial will be printed in the July 27 issue of Science.

4. Why ignore those icebergs? (I).

5. Why ignore those icebergs? (II)

6. For a thoughtful critique of evaluations in academic science, see RD Vale, Evaluating how we evaluate, President’s Column, ASCB Newsletter, 1 May 2012. Pdf here.

7. D Korn, et al., The NIH Budget in the “Postdoubling” Era, Science 296, 1401 (2002).


Why ignore those icebergs?(II)

How do expansionism and research training stimulate one another? And why, pray, do we still sit on our hands?

Addictive expansionism in biomedical research centers (1) is one important element of a conceptual framework that may help us to understand the myriad troubles that afflict biomedical research in the US. Now let’s take a look at how that expansionism connects to a second element: the training of young biomedical scientists.

Before expansion of biomedical research really took off, in the 1960s, training PhD students and the few available postdocs could be pretty much catch-as-catch-can. Scientists chose a few young people to work in their labs and paid them meager stipends, sometimes from their own pockets. Because training varied greatly in quality, from lab to lab and institution to institution, in the early 1970s NIH began to offer institutional training grants, which paid uniform stipends to students and postdocs (2). Training standards probably improved, but available training grants supported too few trainees, so PIs and their institutions tapped into research grants to pay postdocs and many graduate students (dubbed research assistants).

Paying trainees from research grants brought important benefits: (i) a large supply of bright, highly-motivated lab workers; (ii) a substantial financial advantage, because trainees command lower salaries than do permanent staff scientists with PhDs; (iii) a boost in creative innovation, kindled by mixing established mentors with ambitious young trainees less burdened with pre-conceived ideas and unafraid of tackling new questions; (iv) production of many well-trained, skilled scientists for jobs in academia and industry, at a time when such jobs were relatively more plentiful than now. Finally, generations of mentors and trainees discovered that solving scientific problems together can be exciting, gratifying, and productive.

Later, BiomedWatch will paint a larger picture, illustrating how multiple powerful positive feedbacks and weaker negative feedbacks connect the NIH, rampantly competitive expansion of research centers, and the essential role of trainees in the investigative workforce. Here I present a less fine-grained account of two positive feedbacks, located close to the system’s center, each of which stimulates the other.

One exceptionally strong positive feedback loop, discussed earlier (1), is driven by the strong competitive urge of biomedical research centers and their learned addiction to regular increases in NIH funding. Like a knife-wielding gourmand used to consuming slices of an ever-growing pie, each research center fiercely competes to wangle a disproportionately bigger helping. In the recent era of flat-line NIH budgets, however, each center usually captures a slice approximately equal to last year’s or, owing to inflation, even smaller.

For institutions to reap greater return from indirect cost payments and feed their addiction to rampant competitive expansion, they need research labs to grow in size, number, or both. This need drives a positive feedback exerted by both research centers and PIs on training. Academic institutions and their faculties expand PhD training programs and hire more postdocs because relatively cheap trainee labor produces the publications necessary to attract more grant money and indirect cost payments.

Positive feedback also works in the other direction, albeit in a more stutter-start fashion. More new PhDs become more new postdocs, who soon find themselves in a brimming “holding tank” of frustrated senior postdocs, each seeking interviews for a few open positions for running an independent research lab in academia or industry. In addition to furnishing highly skilled labor for established labs, the ready availability of postdocs in this holding tank makes it easier for institutions to hire competent researchers on soft-money faculty salaries, thereby filling newly built labs and reaping further expansion money from indirect cost payments. Ready availability of excess cannon-fodder probably also makes it easier for departments and institutions to discard soft-money faculty researchers whose grant applications are turned down.

We cannot refrain from asking the obvious questions. Why are these and other positive feedbacks not opposed by the negative feedbacks we might expect? Why don’t institutions, faced with decreasing endowments and flat-line NIH budgets, decisively curtail bricks-and-mortar expansion to put more money into faculty salaries? Why don’t PhD programs reduce the size of entering classes? If jobs are really tight, why do young people persist in applying for biomedical training?

I shall return to these questions in later blogs, because the answers differ from case to case and feedback to feedback. For the moment, I submit that the problem reflects a pervasively common human frailty: our inveterate tendency to resort to “double-think” when we want to avoid difficult problems by resorting. For example, a PI may argue that it is his (or his institution’s) duty to attract and train more students, for a host of reasons—and then, on a different day, admit that his students aren’t as good as they used to be, his lab couldn’t possibly handle more students, and his postdocs are having a harder time finding positions. NIH personnel, research trainees, and university administrators are no less susceptible to double-think than PIs. Indeed, double-think is much more common than hypocrisy, malice, or even stupidity—but also more dangerous, because shape-shifting sincerity convinces while it thoroughly stymies real choice.

         Together, rapid expansion of US biomedical research and the creative efficiency of the mentor-trainee relationship furnished an indispensable fulcrum for leveraging investment in 20th-century biomedical research into discoveries and therapies that immeasurably improved the lives of billions. Those triumphs, however, were accompanied by less fortunate effects, including the every-day icebergs (1) that plagued PIs in the late 20th century and persist as grave menaces today, plus the most recent iceberg, which right now is also the scariest—that is, the looming cataclysm created by Congress’s repeated yearly failures to increase the NIH budget, beginning in 2004 and still continuing (1). In combination with inflation, in eight years flat-line NIH funding has already decreased real support of NIH grantees by 18%. Public universities and some private universities find themselves in dire financial straits.

Despite these trends, and their likely persistence for years to come, universities and medical schools continue building new labs, hiring more research faculty on soft-money salaries, and reducing “hard” salary support for faculty already on board. Similarly, NIH continues to receive and process huge numbers of grant proposals, while entering classes of PhD graduate programs and postdoc numbers remain stable or even increase. PIs complain, as we always have, and spend increasing time and effort massaging and polishing more and more papers and grant proposals while competing to attract more postdocs and students. All the auguries have produced laughably negligible responses from the NIH, which shuffles decreasing funds to divide a shrinking pie among a few extra grantees and mires itself in controversy about less pressing issues like translational medicine.

Why do institutions, biomedical scientists, and the NIH continue to sit on their hands, instead of banding together to prepare for or avert looming disaster? We have already suggested two possible reasons. One, surely, is simple inertia, reinforced by adaptation and eventual addiction to federal funding, which increased steadily from the 1960s up to 2003. Inertia and addiction, abetted by double-think, helped prevent decisive responses to the daily icebergs we lived with for decades (1).

Another contributing cause, I suspect, is the hoary American belief—shared by PIs, students, postdocs, universities, research institutes, medical schools, journals, editors, and peer reviewers—that competition solves all problems. But this belief just ain’t necessarily so! Competition is certainly essential in science, but in excess it may sap or even destroy scientific creativity. The problem, as many scientists recall from their own sad mistakes, is that competition distracts us from real questions while it tends to make us believe we must be right and are bound to win. For scientists, competition works better when self-generated than when imposed by adverse financial circumstances. Indeed, truly good scientists often do their best work when they are left alone (see 4).


1. HR Bourne, Why ignore those icebergs? (I).

2. The NIH started the National Research Service Awards (NRSA) program in the mid-1970s to improve training of graduate students and postdocs. At the beginning Congress required the program to set trainee stipends, but these were not increased as cost of living rose in the late 1970s and early 1980s; as a result, stipends remained low at the beginning and students and postdocs became relatively cheap labor. At virtually the same time a few schools—and eventually all—realized that they could pay trainee stipends from research grants as well, at the (low) rates set by NRSA grants.

3. This was certainly true of the scientists whose discoveries I described in HR Bourne, Paths to Innovation: Discovering recombinant DNA, oncogenes, and prions, in one medical school, over one decade, University of California Medical Humanities Consortium, distributed by UC Press (2011). The four scientists included Herbert Boyer (recombinant DNA), Michael Bishop (oncogenes), Harold Varmus (oncogenes), and Stanley Prusiner (prions).

Is your alma mater really a procuress?

In one perverse recommendation, a prestigious NRC panel shows what it really thinks of biomedical research

“The Procuress,” a 17th century painting by Jan Vermeer.

In the course of one perversely ignorant recommendation, a recent National Research Council (NRC) report (1) talked about agencies that “procure” academic biomedical research. From the report’s grand title, “Research Universities and the Future of America,” you may suspect “procure” was a way to avoid the crass verb, “buy.” Instead, I flashed on “procuress”—crasser yet, but redeemed by this magnificent painting, by Jan Vermeer.

By implication, albeit not explicitly, the NRC panel’s recommendation (number six out of ten) characterizes biomedical research faculty as Vermeer’s skilled and attractive harlot (in yellow, her right hand about to receive a coin), the NIH as her lusty john (red jacket, one hand on the harlot’s breast, the other proffering gold), and the research university—which we used to call alma mater, or “nourishing mother”—as the procuress (black hood, satisfied smile). The alma mater/procuress (research university) takes money from the john (NIH) to procure the harlot (researchers) and to rent a suitable room (think research lab). The simpering fop (left)—in black silk and lace collar, raising a glass to the romantic couple—has remained an enigma for 350 years. Later I shall unmask his 21st century identity.

In 2010, four members of Congress asked the NRC to recommend the “top ten actions” government and research universities should undertake in order “to help the United States to compete, prosper, and achieve national goals for health, energy, the environment, and security in the global community of the 21st century.” Few will argue against some of its recommendations—for instance, that we should: stabilize federal funding for university research and graduate education; raise total US research and development to 3% of GDP; increase cost-effectiveness and productivity of university-based research; remove regulations that increase costs and impede creativity; improve PhD graduate programs; offer education to all Americans, regardless of race; and ensure US benefit from international students and scholars in our research enterprise. A few additional recommendations sound too vague, like long-sought secret gardens. The NRC may know where the garden keys are hidden, but doesn’t tell us (2).

Of the report’s recommendations, number six reeks of ignorance so rank it will take your breath away. Quiet but deadly, it boils down to this: “The federal government and other research sponsors should strive to cover the full costs of research projects and other activities they procure from research universities in a consistent and transparent manner.” For non-cognoscenti, a word of orientation is in order. Research costs are often divided into “direct” and “indirect” categories. “Direct” costs (for researcher salaries, materials, equipment, etc.) are those necessary to get the research project done, while “indirect” costs denote “facilities and administration”—F&A, which includes research administration, lab maintenance and utility bills, and research environment. Because funding agencies pay only part of the “full” indirect costs, the NRC report contends, universities cover the shortfall by taking money from student tuition fees and clinical care revenues. In contrast, if the funding agency fully covered indirect costs, universities could “allocate . . . resources more strategically for their intended purpose” (1).

That sounds fine, but here’s the poison: “Federal coverage of a higher portion of indirect costs would, at the margins, shift part of federal research funding from direct to indirect costs, so there will be no net change in cost to the federal government.” In other words, funds for performing actual research would be reduced. Full agency funding is critical, according to Charles Holliday, chair of the NRC panel, who identified the losers with crystalline clarity: if funding agencies can’t pay more, he said, “we’d rather see more of [their money] go to pay the full cost [of supporting research]. . . . Yes, if that were the only option, then cut the amount of money available for research” (3).

What might this mean in practical terms? My former employer, the University of California, San Francisco (UCSF), receives indirect cost payments on NIH grants at a negotiated rate: 56% of direct costs. According to a UCSF official (4), however, the “full” indirect costs (as calculated by UCSF) would amount to 77% of direct costs. Thus, if the NIH paid the same total amount to the university, the NRC’s recommendation would reduce money available to the labs that perform the research by about 21%. Is that a good idea? Would diverting one fifth of all research grants to students or clinicians be a good bargain for society? Investigators already use that money to perform first-class research. Is that purpose really less worthy than those of the university’s students or doctors?

Moreover, Holliday’s committee completely ignored a serious problem with their analysis: of the indirect costs paid from university coffers, a substantial proportion (as much as 54%; 4) relates neither to NIH nor to other federal grants. Instead, much of the shortfall relates to grants from foundations and other private sources, most of which pay indirect costs at substantially lower rates than the federal government (4). Is it likely that the American Cancer Society, the American Heart Association, and other private funders will heed the NRC’s delicately phrased request that they “strive to pay the full costs of research projects . . . they procure from research universities”? No. Will universities refuse their grants, even when the indirect cost reimbursements remain inadequate? No.

In any case, how can the NRC committee assert so unequivocally that sources outside the university must pay all research costs? Shouldn’t a research university contribute some of its own money? Indeed it should, for excellent reasons:

  • The prowess and accomplishments of a research university’s investigators substantially enhance its reputation, attracting students and alumni donations.
  • Creating, imparting, and implementing knowledge are closely connected, each activity increasing the others’ quality. Universities derive great benefits from connecting teaching, research, and community service, such as medical care.
  • Universities already make every possible effort to enhance efficiency and reduce actual indirect costs. Paying indirect costs from funds previously allocated to laboratories will certainly diminish research, without increasing efficiency.

Finally, how did the NRC panel manage to come up with this perverse, short-sighted recommendation? The simple answer: recommendation six directly reflects the ignorance and prejudices of this panel’s 20 members, who included a retired Du Pont CEO as chair, plus seven other business people; ten university administrators (presidents, chancellors, etc.); and two working scientists (5). With only two of 20 members engaged in actual science, the panel bills the indirect cost issue as a business calculation to clinch a lucrative bargain for universities: “If the government covers the full costs of research it procures, universities will be able to hold steady or reduce the amount of research funding they contribute from other sources, such as tuition revenue or patient clinical fees.” But championing this calculation lays bare a shameful fact: the NRC panel doesn’t know where research money comes from or how low indirect cost reimbursements paid by private funders of research can be. Among prestigious university presidents and captains of industry, such ignorance betrays serious deficits of both curiosity and business sense. Luminaries like these should know better.

To explore the NRC panel’s prejudices, I’d like to hear it answer (honestly) these questions: Are business and academic research essentially equivalent activities, best managed in identical ways? Does research funding simply purchase a product that can easily be predicted and quantified? Why did the panel ignore the cherished benefits research bring to universities, the non-government private sources that support academic research, and the impossibility of getting more research for less money?

Now let’s return to Vermeer’s painting. Of course the artist cannot have known that his 17th century procuress, harlot, and john would reincarnate in the 21st century as key players in the US biomedical enterprise: respectively, as research universities, academic biomedical scientists, and the NIH. And in his wildest dreams Vermeer could not have imagined the identity of the reincarnated fop, with his lifted glass and foolish simper. His painting suggests a brothel tout in cahoots with the procuress, a tout who surely shares her cynical disregard of what either harlot or john will get from their transaction. Now, in the 21st century, his fop has unmistakably reincarnated himself as . . . none other than the NRC panel responsible for recommendation six! Precisely the kind of person, we now see, who could berate a spendthrift procuress for paying part of the rent for the brothel room from her own funds, and insist instead that the entire rent come out of the john’s fee, which (so he says) will not increase. He would think it’s fine for the poor harlot to get less money for her performance, because the procuress gets extra dollars to spend on something else. Besides, a harlot’s heart of gold allows her to perform just as well, whether she is paid or not. Are scientists’ hearts not made of similar stuff?

Shame on the NRC for appointing this panel! Somewhere, surely, it could have found 20 individuals less ignorant and disdainful of the purpose, practice, and funding of biomedical research—and, by extension, all science? If not, our society’s movers and shakers are hardly better informed than its poorest untutored dregs.


1. National Research Council report, Research Universities and the Future of America: Ten Breakthrough Actions Vital to Our Nation’s Prosperity and Security, National Academy Press, 2012. Pdf available on web.

2. One of these is a “strategic investment program” to fund initiatives “in areas of national priority.” The reader will also wonder exactly how can we persuade states to make their public universities autonomous and agile enough “to navigate an extended period with limited state support,” or get business and the academic research enterprise to interact more effectively.

3. J Mervis, The Real Cost of Research—and Who Pays, ScienceInsider. The blog includes quotes from an interview with Charles Holliday, chair of the NRC committee responsible for the report discussed in this post. Blog posted on the web 15 June 2012, 2:16 pm.

4. Interview with Eric Vermillion, Vice Chancellor of Finance at UCSF, 2 July 2012. Federal grants and contracts pay for about 58% of UCSF’s research (assessed in direct costs), but the federal government pays about 74% of the portion of indirect research costs that are reimbursed by all funding sources put together. The federal government pays indirect costs at a rate (in proportion to direct costs) slightly more than twice the average rate paid by other funding sources. As a result, UCSF itself foots the F&A bill to the tune of about 35 cents for every federal direct research dollar expended by investigators, on the average, but pays 57 cents for every non-federal direct research dollar. Consequently, non-federal sources pay for 42% of UCSF’s research (in direct costs), but research funded by these sources accounts for a much larger percentage (54%) of the total F&A cost UCSF must pay to keep all its externally supported research going. Bottom line: following recommendation six of the NRC report at UCSF would cut the researchers’ share of the federal research dollar by more than one third, while that share would decrease, by considerably more than half, research funded by non-federal sources. The university could then spend the money “saved” for other purposes, while research itself would wither away.

5. A Public Voice, Online Newsletter, National Academies Names Research Universities Committee. Web 1 July 2010, Association of Public and Land-grant Universities.

Why ignore those icebergs? (I)

Instead, let’s first ask where they come from. We begin with institutional addiction to expansion

BiomedWatch began by comparing the US biomedical research enterprise to the RMS Titanic 100 years ago, heedlessly steaming into dark night, where icebergs loom (1). Why do so many scientists—in academic corridors and labs, in research societies, NIH, and Congress—persist in ignoring icebergs?

One simple, obvious answer: jostled daily by the same old icebergs, we avoid thinking about them because the underlying problems of our biomedical juggernaut are too overwhelmingly big, too complicated and numerous, or just too sadly familiar. Every-day icebergs would be scary if we hadn’t already gotten used to them. Examples:

  • The near-impossibility of publishing in a journal glitzy enough for the first author to get a job interview.
  • The growing fraction of soft money in faculty researchers’ salaries.
  • Shrinking academic budgets in public institutions, and many private ones.
  • Arbitrary peer review, increasingly onerous for both grantees and reviewers.
  • Apparently inexorable graying of biomedical researchers (2).
  • Competing (for funds, papers, trainees) steals time from thinking and experiments.
  • Conflicting goals: training young scientists vs. getting exciting results.

The same icebergs plagued us in the late 20th century, but we managed somehow to ignore them. Now we are trying hard to ignore an even scarier iceberg: no one wants to think about Washington’s terrifying political gridlock, which has produced flat-line NIH budgets for the past nine years and (if it continues through January) will trigger “sequestration” of federal funds, removing more billions from the NIH budget (3). Instead, many of us revile evil gnomes for instigating the icebergs. Do such gnomes populate teeming covens of our colleagues, peer reviewers, journal editors, deans, NIH administrators, or politicians? Not likely—evil gnomes are too few, even in Congress.

The obvious alternative is to figure out what is happening, and why. To do so, we need a conceptual framework to help us pose questions and devise strategies to shape the future. Here I lay out a brief, schematic version of one such framework. For fuller description and more facts, consult the framework’s most thoughtful architects: Paula Stephan, an economist (4), and Michael Teitelbaum, a demographer (5).

The framework, as we shall see, suggests that many forces driving our present troubles originated in choices that were rational at the time. The key choice-makers include stakeholders (no gnomes) throughout the biomedical research community, among scientists and in universities, medical schools, research institutes, NIH and other federal agencies, and Congress. Each set of stakeholders interacts with the others, via positive and negative feedback loops. Perhaps useful at one time, the feedbacks can become genuinely dangerous later. Today BiomedWatch will focus mainly on feedbacks that underlie rampant expansion of the academic research enterprise (see diagram). The following blog will connect this expansionism to training of young scientists.

Historically, biomedical research in the US began to expand soon after World War II, in keeping with: (i) transformation of many colleges and universities from venues for training privileged youths to take their fathers’ place in society into “multiversities” capable of creating new knowledge (6); (ii) rapid increases in federal research funding, via the NIH; and (iii) Medicare, which by paying for care of indigent patients produced an unintended consequence: new funds for faculty salaries in medical schools. Medical schools and universities quickly learned to expand their research faculties, lab facilities, and commitment to research and patient care. NIH funding for extramural biomedical research continued to increase, by 8-9% per year from 1970 to 1998 (7)—a rate faster than GNP, inflation, or the two combined. Yearly increases from 1999 to 2003, the period when Congress doubled the NIH budget, rose even higher.

By itself, increased NIH funding would certainly have supported a substantial size increase in the US biomedical enterprise, but two additional developments further accelerated that increase. One was a gradual shift, beginning in the 1960s, toward paying larger fractions of research faculty salaries from grant funds (“soft money”), rather than from university coffers (“hard money”) (8). Reducing academia’s contribution to such salaries allowed faculty members to teach less and thus to devote more time to the lab. Simultaneously, it relieved institutions from part of their obligation to pay faculty researchers’ increasing salaries. Soft-money salaries gradually began to prevail in medical schools and research institutes, where teaching obligations were lighter (even in the 1960s) than in universities. The shift toward soft-money lifted a heavy financial foot from a potential brake on institutional expansionism. No longer obliged to invest scarce funds in researchers’ salaries, institutions could use those funds to construct new research labs and pay start-up costs for beginning labs. In comparison, I imagine, institutions see salary support for a tenured scientist as a riskier and more expensive long-term investment.

In the latter half of the 20th century, the federal government was more directly responsible for a second accelerant to expansion of academic biomedical research, in the form of rules governing indirect cost payments on research grants. Federal grant costs are direct (incurred by performing the work proposed, including salaries and necessary equipment or supplies) or indirect (incurred by the institution in providing adequate conditions for performing the work). Indirect cost payments typically account for approximately one third of NIH’s total investment in any individual research grant. Rules determined by the Office of Management and Budget (OMB, in the federal executive branch) specify that indirect costs be negotiated by each institution as a percentage (often 50%, or more) of direct costs incurred by the investigator. Institutions can use such indirect costs to pay for heat, light, security, and research administration, plus interest on money borrowed for lab construction.

OMB’s indirect cost rules foster addictive institutional expansionism in two ways (9): (i) by paying interest on lab construction loans, the government encourages institutions to borrow and over-build research capacity even when their funds are stretched; (ii) soft money salaries, charged to an NIH research grant as a direct cost, bring the institution the extra bonus of a larger indirect cost payment on that grant. Thus returns on previous years’ investments begets further investment. By rewarding research centers for failing to commit their own funds (to paying off debts or to faculty salaries), the indirect cost rules encourage expansion and discourage constraints on bricks-and-mortar investment and hiring more researchers. It should be no wonder that academic institutions increased biomedical research space by 62% in 1998-2009 (10), or that medical schools and institutes pay an ever-decreasing proportion of research faculty salaries.

Back in the 20th century, rampant expansion produced extraordinary discoveries and clinical advances that benefitted scientists, universities, physicians, and patients while creating jobs and fueling economic growth (11). In addition to these intended consequences, others were not anticipated. In fact, expansion also created a research behemoth with new properties, including tightly-wound adaptations, such as soft-money salaries and indirect cost rules, that fostered dangerous addiction to ever-continuing growth. One unintended result, among others: a growing chasm between faculty that teach and those that do research. The latter occupy large labs and work hard to keep them going, but as grant funds dwindle they become increasingly aware that the institution values them primarily as earners of indirect costs, rather than as integral members of a common polity with shared broad institutional goals.

Separate BiomedWatch posts will focus on an additional set of adaptations to expansionism, the feedback loops that connect expansion to using research trainees as the source of cheap labor in labs—and vice versa.


1. As described in BiomedWatch, Why blog? Why this blog?

2. The average NIH grantee is now 51 years old—that is, 13 years older than the average in 1980. N Ruiz-Bravo, Conversation with NIH: The Health of the Scientific Workforce, presentation 4 December 2007. Web 29 November 2011.

3. The Budget Control Act of 2011 required caps on discretionary programs, beginning in 2013, which will amount to more than $1 trillion over the ten years from 2012 to 2021. If Congress does not agree on a way to do this, in early 2013 across-the-board mandatory cuts in discretionary funding will “sequester” about $109 billion from the next year’s budget (and from subsequent budgets, until the total budget reduction has been accomplished). The Federation of American Societies for Experimental Biology recently estimated (FASEB analysis demonstrates devastating impact of sequestration on medical research, FASEB Washington Update) that such sequestration would in its first year reduce extramural NIH funding by 11.1%.

4. PE Stephan, How Economics Shapes Science, Harvard University Press, Cambridge, MA (2011).

5. M Teitelbaum, Structural disequilibria in biomedical research, Science 321, 644 (2008). In addition, I recommend a careful reading of the tour de force summary of feedback loops, recently published as appendix D (pages 72-80) of the Tilghman-Rockey report on the Biomedical Workforce. Although not billed specifically as Teitelbaum’s work, he was a member of the panel responsible for this appendix and (I guess) its principal author.

6. C Kerr, The Uses of the University , Cambridge, Massachusetts, Harvard University Press (1963).

7. D Korn, et al., The NIH Budget in the “Postdoubling” Era, Science 296, 1401 (2002).

8. Early on, most of a faculty researcher’s salary was paid by the academic institution (i.e., hard money) or (in medical schools) by fees from patients. Then grants paid small portions of such salaries—called “soft” because money from the parent institution was considered more predictable and reliable—but such soft-money contributions to faculty researcher salaries gradually increased over the years, especially after 1980. Now a part of the salary of almost every researcher is charged to a research grant (or often, to several such grants).

9. B Alberts, Overbuilding Research Capacity, Science 329, 1257 (2010).

10. National Science Foundation: National Center for Science and Engineering Statistics, Research Facilities, Data Tables 3 and 8. Web 28 November 2011.

11. RA Atkinson, SJ Ezell, LV Giddings, LA Stewart, SM Andes, Leadership in Decline: Assessing US international competitiveness in biomedical research, Web 20 May 2012[].