We have entered an age of austerity that, given the fecklessness of the political classes, won't end any time soon. Against that backdrop, the global science community must decide how to succeed with less money.

Major changes in the way science is funded and conducted are needed to ensure the most pressing scientific problems are addressed by the best scientists. The patients who rely on the biomedical innovation ecosystem deserve nothing less, and the taxpayers who foot the bill should be demanding it.

Basic science is a classic example of a public good that should be funded by government. Indeed, it doesn't take much scratching to find publicly funded research underlying virtually every biomedical innovation of the modern era.

Yet publicly funded research is but a part of a complex innovation ecosystem. It produces maximum value to society only when it dovetails with private sector institutions capable of exploiting the scientific opportunities it creates.

With very few exceptions, medical breakthroughs that began with publicly funded discoveries would not and could not have occurred without the private sector making enormous investments, taking financial risks and applying its own scientific inventiveness as well as commercial innovation.

An example of this link is HCV. In 1997, researchers funded by NIH published in Science that they had figured out how to clone the virus. That in turn allowed research teams from multiple institutes, including NIH's National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), to figure out how to culture HCV in 2005. This basic science then made it possible for drug companies to pour billions of dollars into developing targeted treatments that actually cure some patients.

In this manner publicly funded basic research has played a huge role in creating the biotechnology industry.

Seeking to catch up, China, Korea, India and other countries aiming to participate in the 21st century economy have been investing more in basic science research and science training with the aim of building biopharma industries of their own.

But now, a number of countries are in austerity mode, and science budgets are increasing moderately, if at all (see "R&D Trajectories," A3).

In the U.S., budget sequestration could lead NIH - by far the world's biggest source of public research money - to fund 2,300 fewer new grants in FY13, slashing a quarter of last year's total. The burden would fall most heavily on first-time grant applicants, NIH Director Francis Collins told a Senate committee in March.

Even if Congress spares NIH from cuts, there is no chance the next budget will increase enough to reverse the erosion in science purchasing power experienced over the last four years.

No other country funds bioscience for as long or as generously as the U.S., and no other country has as much flexibility in terms of career options for scientists. But NIH's approach to training and research funding has become sclerotic and hasn't kept up with changes in the real world.

Markers of sclerosis, not just at NIH but also within academia, include the fact that a smaller and smaller proportion of young researchers get funding to do independent work, and the fact that a separate pool of money had to be set aside to fund potentially paradigm-shifting science.

Neither has academia kept up with the changing world. Indeed, the university research model is a house of cards based on the idea that public funds will increase into infinity.

China, Korea, India, Brazil and other countries investing to develop basic biomedical science engines have their own problems, including in some cases lack of a research tradition and lack of a developed industry that can translate and develop novel discoveries.

But these countries have an opportunity to do it differently, learning from what does and does not work at NIH.

As other countries improve their scientific base, resetting the system in the U.S. will be crucial if biomedical research and a thriving biopharmaceutical industry are to be drivers of national competitiveness.

But more money will not be the answer in the age of austerity.

Put simply, the $30 billion American taxpayers give NIH annually, the more than $10 billion EU citizens invest in health-related R&D, and the billions donated by philanthropies, $2 billion from the Wellcome Trust and Howard Hughes Medical Institute alone, must be enough.

It is imperative that the global research enterprise find ways to create lasting value for society with the resources available. This 20th Back to School Commentary argues there is no reason it can't. While the will of entrenched interests may be lacking, it is not hard to identify fat in the system. Moreover, the numbers show excessive money is being spent on science that will not yield transformational value, while drip feeding innovation the public good requires.

As has been the case in industry since the meltdown began, righting the ship may mean the absolute size of the life sciences establishment has to shrink.

But just as Back to School previously has argued that rebasing is good for industry, it also will be healthy for the scientific endeavor. If and when the biomedical research enterprise restructures so that it can live - and thrive - with current funding levels, it will be in a much better position to justify future funding increases.

The situation

Any discussion of saving publicly funded basic research must start by acknowledging the system does consistently produce important discoveries that lead to broad improvements in overall health and quality of life.

It is nonetheless clear there is fat in the system that diverts resources from the conduct of research; that not enough transformational ideas get funded; that not enough young researchers are permitted to do their own work; and that there is too little collaboration between countries, institutions and individuals.

While no one who understands and values the scientific enterprise would sensibly argue in favor of a top-down system that prescribes what work scientists should do, it is still true that the current approach is not strategic enough. And even with the recent explosion of precompetitive consortia and corporate-academic collaborations, there is still too little interaction between those who perform basic research and those who translate basic discoveries into new treatments.

Though not all of these problems are exclusive to NIH, many are most easily illustrated by looking at NIH because of its size and the amount of information it discloses publicly.

Ironically, some of these problems were caused by a doubling of the NIH budget from 1998 to 2003.

The doubling was Miracle Grow for university biomedical research programs, which responded with an explosion of infrastructure and activity as if the gusher of money would continue indefinitely.

According to a paper published by researchers from the Association of American Medical Colleges in the New England Journal of Medicine in September 2007, in the eight years prior to the doubling (1990-97), medical colleges spent an aggregate annual average of $400 million on construction and renovation of biomedical facilities.

In a 2002 survey, AAMC schools said they expected to spend an annual average of $1.9 billion on construction and renovation over 2003-07. While an estimate, those amounts were mostly already committed in multi-year construction contracts.

Universities and institutes also increased the number of faculty and trainees in response to the doubling of NIH's budget.

The predictable result of the increase in academic research capacity was a surge in grant applications to NIH that began near the end of the doubling period and has continued to the present. NIH reviewed 18,807 new grants in FY98. By FY03 that figure was 28,355, and last year it was 43,639 (see "Fear of Failure").

However, NIH funding increased by 2% annually between FY03 and FY11, while the cost of doing research as measured by the Biomedical Research and Development Price Index (BRDPI) increased 4% a year.

The BRDPI is calculated each year by the Bureau of Economic Analysis in the U.S. Department of Commerce; however, it is autocorrelated with research institutions driving up payroll and overhead in anticipation of future hikes in the NIH budget.

But the NIH budget also has lagged the Consumer Price Index, which has increased 3% annually from 2003-11. By this benchmark, NIH's budget has actually decreased 1% annually in real terms, from $24.1 billion in FY03 to $22.4 billion in FY11 using 1998 dollars (see "Facing Reality," A5).

Meanwhile, universities and institutes have been relying heavily on NIH to support their bloated infrastructures.

Data from the University of California office of the president show that over two decades, the number of senior administrators in the UC system has grown four times faster than the number of teachers. In 1993, there were 2.5 faculty members to one senior manager. The ratio is now 1:1.

In FY10, on average only 73% of NIH research grants went to project-related costs, while 27% went to overhead, including facilities and administration, at the institution where the research was conducted.

But averages never tell the whole story. Academics whose institutions receive NIH grants and drug developers who have contracted with academics for research told BioCentury that, at some of the larger universities, the overhead charges can routinely be 50% to more than 100% of the project-related costs. A $250,000 research project could come with more than $250,000 in overhead charges.

In many cases, NIH grant monies are used to pay for 100% of faculty salaries. And multiple grants that are awarded to a single PI would each have overhead built in, presumably covering the same laboratory facilities, even if salary is pro-rated for each project.

The increasing numbers of applications, coupled with NIH's reduced purchasing power, have had disturbing effects on academic researchers.

In FY02, NIH reviewed 24,403 new research project grant applications and approved 6,505 of them, for a 26.7% success rate. The average annual funding per grant was $304,951.

By FY11, the number of new applications had increased to about 43,639 and there were 6,629 awards, producing a 15.2% success rate. The average annual funding per grant was $406,980, or $325,491 in 2002 dollars.

The falling success rates amplified a worsening trend toward limiting access to NIH money. The average age at which a scientist receives his or her first independent grant had already begun to climb before the doubling. In 1970, the average age for a first R01 grant was 35. By 1980 it had crept up to 37. It rose to 39 in 1990, hit 42 in 2000 and has been steady for the last dozen years.

This would be an indicator that something was wrong even without the imperatives of austerity.

The age distribution of NIH-supported PIs tells the same bleak story.

"In 1980, less than 1% of PIs were over age 65, and now PIs over age 65 constitute nearly 7% of the total," noted Sally Rockey, NIH's deputy director for extramural research, in a recent blog posting. "In parallel, in 1980, close to 18% of all PIs were age 36 and under. That number has fallen to about 3% in recent years."

The protracted adolescence of biomedical scientists is imposing enormous unseen opportunity costs. Younger scientists are much more likely than their mentors to be familiar with a wide range of technologies and ideas outside their area of specialization. A system that keeps scientists occupied on someone else's research until after their 42nd birthday creates evolutionary pressure for docility and against innovation.

It also keeps a high proportion of the workforce in a kind of postadolescent holding pond, rather than sending them into the world during the very stage of their careers when they should be moving up the ladder.

The numbers of people getting Ph.D.s in biomedical science in the U.S. has more than doubled in the last 20 years. There are more young women and men doing science than ever, they just aren't doing their own science.

As Georgia State University economist Paula Stephan noted in How Economics Shapes Science, U.S. science operates on a pyramid scheme. The best way for PIs to ensure future success and the production of research that will enhance their reputations is to recruit as many talented postdocs and graduate students as possible to work in their labs.

Stephan is no mere pundit, having served on boards for NIH, the National Research Council and the National Science Foundation (NSF).

In exchange for this dedication, postdocs supposedly are receiving valuable training that will lead them to tenure-track academic positions and one day to become PIs themselves. But their mentors have no incentive to inform postdocs there cannot possibly be enough tenure-track positions to absorb them.

The result is that after dedicating many years to graduate school and postdoc training, many people who earn Ph.D.s in biomedical science never work in the profession.

This represents an immense waste of talent and resources - and money.

The knock-on effects don't stop there.

Middle-aged scientists who have spent their entire careers working for other scientists are desperate to get their first grant, so the incentive is to play it safe by submitting an application that is interesting enough to appeal to the majority of the reviewers, but not so cutting-edge they decide it is unlikely to succeed.

It is an open secret that the most effective way to improve one's chances of receiving a grant is to have already performed all or most of the work described in the application. This kind of gamesmanship was routine in the Soviet Union, where there were strong incentives to hit five-year plan targets and few rewards for unexpected accomplishments. The USSR wasn't a font of innovation.

We know from the biotech industry, and from Silicon Valley, that failure is almost a prerequisite for success. But there is little tolerance for and certainly no reward for failure in the current NIH grant-making structure. Safe science isn't likely to be breakthrough science.

This is not to deny the incredible biomedical breakthroughs that have come out of the U.S. system. Rather, it is to say that the kinds of inefficiencies that were tolerable in an era of plenty now will be dead weight in an era of austerity.

While the NIH peer review system is often cited as the gold standard by researchers from other countries, U.S. researchers frequently note it is a conservative old boys' network. Less frequently, they say the peer-review network has too many third-rate reviewers who give high scores to mediocre proposals.

Either way, the fact that NIH has had to create four grant programs to fund so-called "high risk-high reward" science provides evidence that the peer review system favors safe, predictable research over innovative, paradigm-testing projects.

NIH admits as much on its own website: "NIH has traditionally supported research projects, not individual investigators. However, complementary means might be necessary to identify scientists with ideas that have the potential for high impact, but that may be too novel, span too diverse a range of disciplines, or be at a stage too early to fare well in the traditional peer review process."

The website goes on: "To address this, the NIH Common Fund created three companion awards, the NIH Director's Pioneer, New Innovator, and Transformative Research Award, to encourage creative, outside-the-box thinkers to pursue exciting and innovative ideas about biomedical research. A fourth award in the High-Risk Research Program, the NIH Director's Early Independence Award (EIA), was created in fiscal year 2011 to support exceptional early career scientists who possess the intellect, scientific creativity, drive, and maturity to flourish independently immediately following their graduate training, eliminating the need for traditional post-doctoral training."

In FY11, these four programs awarded 89 new grants totaling an underwhelming $147 million.

The ever-intensifying competition for research grants creates a second set of anti-innovation evolutionary pressures. Virtually everywhere, not just at NIH, academic scientists spend at least as much time applying for funding as doing research.

And there is far too little collaboration between scientists, institutions and countries on the most compelling questions that require a diversity of scientific disciplines and approaches, as well as the economies that are gained from pooled resources and reduction in duplication of efforts.

The path forward

Fiscal austerity in the U.S. and Europe is widely viewed within the science establishment as the cause of the crisis. But it actually should be the impetus for fundamental changes that will make the global science enterprise stronger and more productive.

Even NIH's Collins has concluded that just putting in more money isn't the answer. There are those who disagree with his solution - dedicating a new center to translational research - but at least he's thinking about how to get a bigger and faster return from the nation's investment in biomedical research.

In any case, hope is not a strategy. And muddling through until the economy improves and governments become more generous is not a viable option.

Some of the urgently needed changes are obvious, like cutting funding for pseudo-science.

NIH's Center for Complementary and Alternative Medicine (NCAM) is an obvious example. It was established because Sen. Tom Harkin (D-Iowa) and a couple of other members of Congress firmly believe in the power of bee pollen and cow colostrum.

Over $1.2 billion has been spent on NCAM over the last decade, and it is set to receive $120 million next year. NCAM has funded a depressing catalog of hypotheses, such as the ideas that smelling lavender or lemon will promote wound healing and coffee enemas will cure pancreatic cancer. It also has given financial support and intellectual stature to training programs for homeopathy.

But while eliminating pseudo-science would be intellectually satisfying, freeing up $120 million a year from the NIH budget is little more than looking under the sofa cushions for loose change.

Key opportunities for emerging science establishments throughout the world, and challenges for the U.S. and EU, include making publicly funded science as nimble as the private sector it feeds into; making more science directly relevant to medical product developers; and supporting and enabling new generations of scientists.

The overarching objectives for those who control the public purse should be to prioritize basic over clinical research; prioritize individual investigators over top-down big science; and create a bigger marketplace of ideas.

In the U.S., for example, huge amounts of money could be made available for basic research - and for translational research - by reducing spending on clinical research. In FY12, NIH estimates it will spend about $3.1 billion, or 10% of the budget, on clinical trials.

This means there should be limits on the kinds of clinical research government conducts and funds. These limits are based on practicality rather than principle: clinical research is phenomenally expensive and, with a few exceptions, government doesn't do it very well.

Clinical research shouldn't be contemplated unless there is a clear and compelling need for the results, and a plan for rapidly translating them into patient benefit.

If and when government funds clinical research, trials should be done only where industry clearly has no incentive to do the work, and they should be designed to produce results that could be transformative.

The flood of activity unleashed by the Orphan Drug laws throughout the world proves that there is very little truth in the argument only government will invest in the search for treatments for rare diseases.

On the other hand, industry is not going to fund clinical trials comparing the safety and efficacy of medical procedures. And industry is unlikely to fund the kinds of trials the National Cancer Institute has run to optimize the sequencing and dosing regimens of already approved drugs.

Whenever possible, government funding for clinical research should be contingent on validation by co-investment by industry and/or patient philanthropies.

Nor should government be in the business of funding clinical R&D with the goal of getting new drugs approved - it isn't good at it, and there is little chance that the consensus processes required by government could be consistently successful.

Importantly, getting government out of the clinical trial business means industry must accept its obligation to conduct and/or fund the downstream work that is necessary to translate basic research discoveries into treatments.

Countries that lack a well developed biopharmaceutical industry could be an exception. There, it may be appropriate for governments to fund more translational or clinical work.

Singapore, for example, has pharmaceutical manufacturing but lacks drug discovery and development, venture backing and the entrepreneurial spirit that would drive researchers to start companies.

Thus, according to Francis Yeoh, "the funding mechanism scales have been tipped to incentivize translational science."

Yeoh stepped down in July as head of the Singapore National Research Foundation.

Public science funders also should prioritize individual investigators or collaborations of investigators over top-down big science. It doesn't make sense to think of most life sciences research problems as "moon shots" or Manhattan Projects that can be solved primarily with large infusions of money, a great deal of organizational and logistics expertise, and some applied engineering. The human genome project was a successful exception, but it is not the rule.

At the same time, public science funders need to find ways to create a bigger marketplace of ideas, by funding more breakthrough science, funding multidisciplinary research and funding talented young scientists with good ideas without shackling them to their mentors indefinitely.

More fundamental discoveries that will lead to solutions for critical unmet medical needs will come from small projects, sometimes from previously unknown researchers working in institutions lacking prestige.

Accomplishing these objectives will require specific actions, described in more detail in this essay.

Within basic science, there should be more strategic ways of setting priorities based on societal need. This should include formal mechanisms for translational scientists and drug developers to tell basic researchers what specific questions are holding them back, and there should be funds dedicated to research that will answer those questions.

At the same time, the peer review system needs to be revamped and to make room for mechanisms that support bold, innovative ideas.

Meanwhile, a much greater proportion of public funds should be used for research, not for administrative overhead and facilities. When tools and facilities are needed, public money should be spent whenever possible on large-scale resources that many can use.

In terms of human capital, training has to be revamped to reduce money that now is spent to create a surfeit of scientists, and to allow the best of them to conduct independent research earlier in their careers.

Finally, the adaptation to austerity also should result in more collaboration, not only on the relatively small scale between and among investigators and institutions, but also on a much larger scale between governments.

While it is harder to identify how this will happen, Back to School argues austerity itself will force the issue.

Priority setting

Within basic science, there should be more strategic ways of setting priorities based on societal need. This should include formal mechanisms for translational scientists and drug developers to tell basic researchers what specific questions are holding them back, and there should be funds dedicated to research that will answer those questions.

It is essential to preserve the independence of academic investigators, and attempts to centrally impose topics and questions would be disastrous. At the same time, the justification for government investment in life sciences research is that it supports medical progress.

However, academic researchers spend little time, if any, with the scientists who are working to turn their discoveries into new therapies.

Publicly funded scientists should spend a lot more time with the people and institutions that translate their work into products, and to explicitly orient more of their work to solving puzzles that are preventing the translation of scientific discoveries into therapies.

This doesn't mean abandoning basic science for translation. It means that translational work raises basic science questions that need to be answered.

"Society has to think about how it wants to spend money efficiently to solve its most significant challenges. Where we are faced with an aging population in this part of the world, in other parts of the world we're tasked with addressing significant chronic and infectious diseases. The question is how do we use the resources we have to solve the world's questions," said Paul Stoffels, worldwide chairman of pharmaceuticals at Johnson & Johnson.

"Progressing science faster together and then developing specific products will be good for industry, because then everyone will be at a higher level together and then they can compete on how to get products to market," he added.

According to Joshua Boger, former CEO of Vertex Pharmaceuticals Inc. and chair of the board of fellows at Harvard Medical School, a challenge for basic research should be to answer the question: "Who is waiting for the results of my work?"

Science funders and academic scientists, Boger said, "are not giving a lot of thought to organizing how we are investing in research. They are not asking if we are orienting basic research objectives around questions that the folks who are trying to translate basic research into societal benefit are asking."

The rigor and relevance of basic biomedical research would be enhanced substantially if the system were more tuned to reward investigator-initiated research that answered the question, "who is waiting for the results?"

According to Yeoh, during the current funding cycle in Singapore, some grants are available only if there is an industry partner. Also, some large grants ($10 million) are earmarked specifically for interdisciplinary teams of scientists.

Singapore has also established "national innovation challenges" where additional research is needed in a particular area, and the NRF will direct some percentage of funding toward that larger goal.

As a benchmark, in the U.S., about 13% of research project grants in 2011 were awarded in response to requests for applications (RFAs) issued by institutes within NIH for proposals in targeted areas. These grants accounted for 24% of total research project grant funding.

The problem is that most priority setting is carried out by members of the extramural research community, and doesn't include stakeholders with hands-on knowledge of how to build the solutions.

The National Institute of Allergy and Infectious Diseases (NIAID) is part way there: priorities are set by an 18-member council comprising 12 researchers and six laypeople representing people living with disease.

While patient involvement is a good start, the process would be strengthened by incorporating more points of view, especially those of translational and clinical scientists, and those working on drug discovery and development in industry.

Among its four strategic objectives, the National Institute of Mental Health (NIMH) says it aims to "strengthen the public health impact of NIMH-supported research" - an admirable goal.

However, the description of how it proposes to do so is bland to the point of being meaningless: "Through research, evaluation, and collaboration, we will further develop the capacity of the Institute to help close the gap between the development of new, research-tested interventions and their widespread use by those most in need."

NIMH does note its success will depend on collaborating with all stakeholders in the field of mental health. It lists these stakeholders as payers, service providers, patients, families, advocacy groups and professional organizations.

Scientists and companies that actually develop new therapies are not on the list. But those are precisely the people and organizations who are the first to know if a basic research finding is panning out in practice, or if it is raising new questions that need to be answered before there is a path forward.

"In neuroscience, we're working on how to go to the next stage," said Stoffels. "We should collaborate more between the entire community from basic science to companies to identify initial steps - what do you need to know - biomarkers, clinical trial designs."

Stoffels told BioCentury: "I went to NIMH, and we had a discussion about how to do this. That was 18 months ago. A lot of discussions are going on. What we need is for both government and industry to say this collaboration needs to happen, so that the science would advance faster."

Jim Greenwood, president and CEO of BIO, agreed that the missing link between academic scientists, NIH and industry is a way to communicate questions that arise in applied research back to scientists working on basic research - especially when studies fail for unknown reasons.

Investigator-initiated research that allows scientists to independently pursue novel thinking is "a very important thing for academic researchers to be involved in and for the federal government to fund - but it is not the only thing necessary to get the job done," he said.

"Hypothetical science questions can be developed as a direct result of failures in the clinic and can be reviewed by NIH, which will know which academic researchers are following their noses in this area and might be interested in working on them," he said.

Providing funds for research aimed at answering those questions is a way to "make sure tax-funded research has the greatest likelihood of resulting in medical products," Greenwood said.

Although it is focused on a different type of scientific question, the National Center for Advancing Translational Sciences (NCATS) could provide a model for how NIH and industry can work together to identify priorities without devolving into a top-down, centrally directed approach.

"We are working with NIH on NCATS and helping define what industry thinks would be most useful," Greenwood said.

For example, in May, NCATS announced its first initiative would be to partner with pharmas to repurpose compounds. Greenwood said BIO thinks it would be more useful for NCATS to work on broad studies that would help modernize clinical trials and clinical development, rather than conducting studies on a particular molecule.

Reforming peer review

The peer review system needs to be revamped and to make room for mechanisms that support bold, innovative ideas.

Expert review is intended to ensure that the best scientific proposals from diverse research fields are identified and funded. Large organizations conduct peer review of project proposals and then fund those with the highest scores for 3-5 years, after which scientists may apply for a renewal grant. NIH and the European Research Council (ERC) predominantly use this model (see "Drawing the Payline," A8 & "High Risk, High Gain," A9).

But peer review only works with the participation of engaged high-caliber reviewers, and to be successful the system must be structured to ensure innovation is properly recognized as a criterion of excellence.

This is an issue that researchers in emerging economies are acutely aware of.

"The problem is that if the whole system is producing science at a particular level, this is the system that will be judging the proposals. You are in a kind of trap where average people are judging average projects, and it's difficult to get rid of that [mindset] and drive excellent projects," said Kleber Franchini, director of the Brazilian Biosciences National Laboratory.

In Korea, the government is building a new Institute for Basic Science that plans to have 3,000 researchers and staff based at 50 newly established research centers by 2017. According to Ulf Nehrbass, CEO of Institut Pasteur Korea, the goal is to recruit "excellent people meant to have the freedom to explore within the constraints of peer review."

But in countries with a tradition of excellence in biomedical science, conservatism in peer review can stifle innovation. Conservatism is especially pernicious when only a small percentage of top applications are eligible for funding and innovative ideas can so easily miss the cut.

This is reflected in oft-heard criticisms of NIH: that peer reviewers demand too much preliminary data, that the study sections responsible for reviewing proposals do not have the expertise to evaluate multidisciplinary projects, that young researchers cannot compete, and that the practice of allowing applicants to revise and resubmit proposals results in what amounts to experimental design by committee.

NIH has tacitly acknowledged its peer review system does not reward innovation by creating the four new grant programs with the explicit mandate to fund what it terms "high risk-high reward" research (see "Risk-Reward at NIH").

However, the FY12 budget allocates just $196.3M for these awards - a drop in the bucket compared to the $16.5 billion total allocated to research project grants.

Next year NIH will provide at most $15 million for Director's Transformative Research Awards and will award up to seven Pioneer Awards. It expects to grant at least 33 New Innovator Awards.

Back to School argues these numbers should be bigger. For example, expanding the four programs to 5% of the total research project grant budget would move $825 million toward research that promises innovation, and still is only a dollop of money compared to the potential societal reward.

But these carve-outs are not all that is needed. They must be accompanied by fundamental changes to the application and peer review system - even if such changes may lead to reduced funding for safer projects that would be more likely to provide solid, if not groundbreaking, results.

Back to School argues that research proposals should be cut to five from 12 pages, and there should be no requirement for preliminary data.

In 2009, the maximum length of a research proposal was cut from 25 to 12 pages. "It became apparent that the only way to keep reviewers from focusing on the experimental details was not to have them in the application," said Keith Yamamoto, vice chancellor for research at the University of California, San Francisco and co-chair of an external advisory committee convened in 2007-08 to advise the NIH on peer review reform.

He favors a seven-page limit. But five is already the limit for the NIH Director's Pioneer Awards.

Back to School would go on to decree that preliminary data should not be required for new R01 applications. It is reasonable to assume that reviewers will still look for preliminary data to support the feasibility of grant applications, but NIH can begin to change this mentality by explicitly instructing reviewers that these are not required, as it does for its subset of Director's awards.

Better still would be to disallow preliminary data for Director's awards. Truly innovative ideas that would not be funded by any other mechanism by definition should not be validated with preliminary results. Permitting "optional" preliminary data simply provides an excuse for the peer review process to play it safe and flies in the face of the spirit of the award.

Again, ERC validates the idea. It has adopted a funding scheme that aims to promote "frontier research" and does not require any preliminary data - regardless of the career stage of the investigator.

Without a requirement for preliminary data to support feasibility, more projects selected for funding may fail, and first-time renewal rates would decline. However, these tradeoffs are worth making if they result in funding more innovative research projects.

Eliminating the requirement for preliminary data also would enable skilled young investigators whose projects fail to more quickly move on to new avenues of research.

Removing the preliminary data crutch also would force NIH study sections to re-evaluate how they judge the feasibility of novel experimental approaches and lines of research.

Fundamental changes also must be made to the makeup of study sections to adapt to the increasingly multidisciplinary nature of applications and to further encourage participation by top investigators.

NIH currently recruits technical specialists into individual study sections as ad hoc reviewers. For example, if the tumor microenvironment section were reviewing an application that uses a new microscopy technique, a specialist in that technology could be added to the study section. But instead of reviewing only the application involving that technology, the specialist, who may not have extensive knowledge of the tumor microenvironment, would participate in the discussion and scoring of all of the applications presented at the meeting.

This unnecessarily bloats the size of study sections and is an unnecessary burden for specialists, who may be more willing to volunteer their time if their expertise were more efficiently utilized.

For select multidisciplinary applications, including for SBIR and Director's Transformative Grants, NIH piloted a two-stage process to address this concern. In these pilots technical merit was first assessed by specialist reviewers who provide a written critique, but not a score. This was then followed by an in-person meeting of topical experts to discuss the overall impact and significance of the science and score the applications.

Back to School would ask why the pilot has not been expanded to traditional R01 applications in select study sections as a first step toward expanding it throughout the NIH peer review process.

This approach would ensure that the most appropriate experts in rapidly developing technology contribute to the review process, even if they don't have expertise in the particular field of study being discussed.

It also would free study sections to spend more time discussing the innovation and significance of grant applications in a given area of biology, which is where their expertise primarily lies and where their time is better spent.

While the peer-review system provides the best system for the large-scale evaluation of scientific proposals, Back to School also supports complementing the system with a second model that selects individual researchers with high potential, and then gives them sufficient funding to follow their noses for 5-7 years.

Long-term support allows a scientist's work to diverge into unexpected avenues. It is also a way for countries that do not have a long-standing tradition of excellence in biomedical science to begin to build one.

This system has been highly successful at institutions such as Howard Hughes, the Wellcome Trust and the National Institute of Biological Sciences (NIBS) in Beijing (see "People Over Projects," A11).

Such a model is not scalable to the size of NIH. However, NIH should do more to identify exceptional investigators and give them the trust and funding to pursue independent lines of innovative research.

Along this line, NIH has experimented with awards that emphasize the qualifications of the investigator more than a traditional R01 would, but it needs to set aside a larger portion of its budget for this purpose and explicitly fund the investigator, not the project.

The Director's Pioneer Award is on the right track. It was established in 2004 to "support individual scientists of exceptional creativity who propose pioneering approaches to major challenges in biomedical and behavioral research."

However, the program is grossly underfunded, having issued a total of $10.4 million in 13 grants in FY11. NIH expects to hand out "at least 7" awards in 2013.

The agency's Method to Extend Research in Time (MERIT) award also could be restructured to provide support to individual investigators to pursue breakthroughs. Currently, the five-year award serves as a prize that continues to support research that is in little danger of losing funding.

MERIT is given by individual NIH institutes to top experienced investigators in lieu of an R01 that is up for renewal. The prize is intended to allow a PI to pursue more innovative or risky work by providing temporary relief from competitive grant renewal. In 2007, MERIT awards accounted for about 3% of R01 grants.

However, last year the National Heart, Lung, and Blood Institute (NHLBI) stopped participating in the program because "much of the work funded by the MERIT program, while highly meritorious, has not been especially 'high impact.'"

The program should be restructured as an additional source of funds available to excellent investigators who would apply on the strength of their past research accomplishments. These investigators would then be free to pursue innovative lines of research of their choosing.

Howard Hughes provides a model with the selection criteria it uses to judge the quality of candidates up for award renewal. According to CSO Jack Dixon, one is to apply the "subtraction test," which asks: "If you were not there, would the field be worse off?"

Capping overhead

A much greater proportion of public funds should be used for research, not for administrative overhead and facilities. When tools and facilities are needed, public money should be spent whenever possible on large-scale resources that many can use.

In the age of austerity, it is no longer acceptable - if it ever was - for universities and research institutes to skim an average of 27% off the top of grants funded with public money.

Nobody would donate money to a charity he or she knew was spending 27% of donations on overhead. And if the public knew that was the case with research funds, that wouldn't fly either.

In the current U.S. system, overhead charges are determined from a base that consists of a project's direct costs. Direct costs include salaries for investigators, technicians, postdocs and other staff working on the project; reagents and other supplies; and equipment. Direct costs also may include tuition and stipends for graduate students working on a project.

Indirect costs, which are intended to cover the operating and administrative costs of labs, are most typically calculated as a percentage of direct costs.

This creates perverse incentives for the universities, including cases in which universities actually are not paying the investigator, because pushing salaries into a grant's direct costs also bumps up the overhead charge.

These effects are illustrated both by anecdotes and by data showing that university capital spending on laboratory space tripled in response to the doubling of NIH's budget in anticipation that the cost of their operation would be covered by public money.

Peter Ho, a founder and president of China's BeiGene Ltd., told BioCentury that when he was first starting his career in big pharma, there was no question research buildings and laboratories in industry were newer and better designed than facilities in even fine institutions.

Now, he says, when he visits top-tier institutions, their labs are not just as good as industry's - they're better.

"Long-term, universities are really not doing themselves a favor by charging exorbitant overhead," he said. "If you look at some of these buildings, do you really need to spend that money building temples to research? Maybe you can spend that money in other ways. Maybe more faculty get funding and do more research."

How much should overhead be cut?

ERC is making it work with less. Overhead in its grants is calculated as a flat 20% of eligible direct costs.

Horizon 2020, the research and innovation program in the EU for 2014-20, which will succeed Framework Programme 7 (FP7), also has a proposed 20% flat rate for overhead.

The Innovative Medicines Initiative, the public-private partnership of the European Commission and the European Federation of Pharmaceutical Industries and Assocations (EFPIA), offers two options. Organizations that are able to calculate actual indirect per-project costs can receive full reimbursement for that amount. Organizations that are not able to do so - which is most of them - get 20%.

If Congress were to cap the amount of overhead that can be included in an NIH research grant at 20% of direct costs, or 17% of the grant's total, that would save $1.7 billion based on the FY12 allocation of $16.5 billion and the current average overhead, which is 27% of total grant funds.

Obviously, expensive technologies that are necessary to conduct basic research must be paid for, and Back to School would argue this can be a particularly good use of public money when it is used to fund large-scale resources that many can use.

For example, Brandeis University has a nuclear magnetic resonance facility that may be used by any NIH-funded researcher, especially those in New England.

Another example is the Broad Institute of MIT and Harvard's Chemical Biology Program, a high throughput screening facility that can be used by the public research community.

In New York, 11 institutes have banded together to build the New York Genome Center, a large-scale genome sequencing facility that will provide a dedicated resource for DNA sequencing and bioinformatics and is intended to facilitate collaborations between institutes.

Finally, Back to School also agrees with arguments put forth in a 2010 editorial by Bruce Alberts, editor-in-chief of Science, that grantee institutions actually need to have skin in the game.

According to Alberts, NIH should require institutions to pay at least half of a PI's salary.

"A new NIH policy must make it unambiguously clear that expansion through laboratory building construction requires a substantial, nonreimbursable, long-term commitment of resources, including 'hard-money' faculty support, by any institution that wants to increase its facilities and research staff," he wrote.

Leveraging human capital

Training has to be revamped to reduce money that now is spent to create a surfeit of scientists, and to allow the best of them to conduct independent research earlier in their careers.

An inevitable consequence of rebasing is that university research programs will shrink. But Back to School argues that the science establishment, which is producing many more postdocs than can find jobs in their fields, is due for a right-sizing.

NIH is by far the largest funder of science training in the world, and thus provides the object example.

For starters, the U.S. is demonstrably cranking out many more Ph.D.s than the system can support.

It is not in the public interest for NIH to fund training for a growing and significant fraction of individuals who ultimately have no career path in the biomedical enterprise. Nor is it in the public interest to continue to use dear dollars on a training system that prevents talented young researchers from making independent contributions.

Between 2000 and 2010, the number of Ph.D. graduates in the biological or biomedical sciences in the U.S. increased by 38% from 5,853 to 8,052, according to survey data from the National Science Foundation (NSF). Yet the proportion of life sciences Ph.D.s employed in occupations that are closely related to their field dropped to 59% in 2008 from 70% in 1997.

In addition, Ph.D.s can go on to two or three postdocs. These are intended to be temporary positions that provide additional training, but they more closely resemble indentured servitude lasting until or unless independent research positions can be obtained. Only then can scientists apply for and receive a major independent grant, which Back to School already has noted does not happen until an average age of 42 in the U.S.

According to NIH's Biomedical Research Workforce Working Group, the NSF survey showed a "significant number" of postdocs remain in these positions 5-8 years - hardly temporary.

Back in 2005, the National Academy of Sciences argued no postdoc should receive funding for more than five years. This limit would benefit the ecosystem by pushing scientists into independent exploration of more novel ideas sooner. And it would force individuals with no academic career prospects to find work where they can be more productive.

On this latter point, NIH should insist that the graduate programs it pays for educate students about careers outside of academia.

"Less than half of the people trained in Ph.D. programs in biomedicine end up in tenure track positions, and many more of them expect that will be the outcome," Collins told BioCentury. He said training programs should give students and postdocs the opportunity to sample different career options, and added universities should no longer refer to these opportunities as "alternative careers."

"It is not necessarily the case that being a faculty member at a university is the highest, boldest of all options," Collins said.

Some science funders have launched initiatives to help accelerate the transition from mentored postdoc to independent investigator, and to allow some Ph.D.s to go straight from graduate school to independent research without a postdoc. But these initiatives are too small to change the overall demographics.

In FY2012, NIH's allocation for training totals $777.8 million. The agency thus has the power to dramatically reshape the training of scientists from around the globe by changing the way it allocates funding and by demanding that the training programs it pays for prepare students for multiple career paths.

It does have two such programs funded through the NIH Common Fund.

The Pathway to Independence award, which was established in 2007, provides two years of mentored funding to postdocs followed by three years of independent funding contingent on securing a faculty position. In 2011 there were 180 awards funded - a 22% success rate for applications.

However, postdocs are eligible to apply for the award through their fifth year of support, which leads to an unnecessary extension of their mentored training and unfairly pits them against the postdocs with little experience whom the award is intended to nurture.

The Biomedical Research Workforce Working Group has recommended cutting back eligibility from fifth-year to third-year postdocs, along with a doubling of the number of awards.

A second NIH program, launched in 2011, is the Director's Early Independence award. It allows exceptional Ph.D. and M.D. candidates to forgo a postdoc and immediately begin independent research. The award provides five years of funding and is available within the first year after a candidate receives his or her terminal degree.

This program is modeled after prestigious and competitive fellows programs funded by the Whitehead Institute for Biomedical Research and UCSF.

NIH expects to award 10 early independence grants in 2013. Back to School finds it hard to believe there aren't more than 10 graduates ready to be given the chance to conduct independent research. If the number of awards were quadrupled to 40, it still would amount to only about 0.5% of the current number of graduates each year.

Initiatives that help speed the transition of Ph.D.s and postdocs to independence are necessary, but they do not address the reality that it is more difficult for new independent investigators to receive grant funding in a system that requires them to compete with established researchers.

Howard Hughes' Dixon argues it is particularly important to fund new independent investigators. "This is when you are really at your best, and you don't want to burn these young people out - you want to give them the opportunity," he said.

Howard Hughes is one of several institutes inside and outside the U.S. that have established grant programs specifically targeted to the career stage of the investigator.

ERC has separate grant programs for independent scientists who are 2-7, 7-12 or any number of years past completion of their Ph.D. These programs are called Starting Grants, Consolidator Grants and Advanced Grants.

"In a competitive granting scheme, you have to compare like with like," said Anna Tramontano, a member of the ERC Scientific Council.

One reason, she said, is that one would expect scientists at different career stages to be capable of different kinds of research. For example, a more experienced researcher may be better able to manage a project requiring a large group. In contrast, newly minted investigators don't usually have a group and would be likely to propose different kinds of projects.

NIH's approach has been to identify and label grant applications from new or early stage investigators, discuss and score them separately, and then require that individual institutes lower the cutoff hurdle for funding them. New investigators are defined as those who have never before received an R01 grant. Early stage investigators are a subset of new investigators who graduated up to 10 years ago.

For example, in FY12 NIAID funded all R01 grants that scored in the 10th percentile for established investigators, but extended that cutoff to the 14th percentile for new investigators.

Critics would argue this policy supports young investigators at the expense of higher quality science. However, there are legitimate reasons why young investigators may obtain lower grant scores that have little to do with their science, including grantsmanship by seasoned PIs, lack of resources to obtain preliminary data, and a less developed history of accomplishments.

Scaling up collaboration

The adaptation to austerity also should result in more collaboration, not only on the relatively small scale between and among investigators and institutions, but also on a much larger scale between governments.

In 2011, Back to School mined this theme of "Innovation & Collaboration." It argued the biopharma industry was "busy restructuring for the future," and that better engines of value creation would result "only if truly fresh thinking is allowed to replace old expectations and habits."

Back to School argued this will require "an already heavily partnered industry to be even more broad-minded about how collaborations must be at the center of value creation" (see BioCentury, Sept. 5, 2011).

The age of austerity will impose the same imperative on the publicly funded research enterprise.

There are no obvious mechanisms for forcing global research collaborations between institutions or between governments. But the prospect of starvation will be a mighty motivator for finding ways to pool resources and expertise on problems that could benefit from the diversity of scientific disciplines and approaches that are a precondition for solving today's most compelling scientific questions.

For example, progress in many areas will be contingent on skill in managing big data - skills that life sciences academics and industry lack.

The Weizmann Institute of Science has done particularly well, although that is in large part due to the institute's small size and the fact that scientists at the Weizmann spend so much time together.

The institute comprises five faculties: biology; biochemistry; chemistry; physics; and mathematics and computer science.

"We have 250 active PIs, and the science buildings are clustered together, so we are all physically very close together," said Irit Sagi, the Maurizio Pontecorvo professorial chair in the department of structural biology. She noted many of the researchers live on campus.

"All this informal interaction leads to fruitful collaborations and unexpected collaborations. It's not something like the institute directs these collaborations; it's very natural. Scientists love their work, and when we get together, after five minutes we talk science," she told BioCentury (see "A Matter of Character," A15).

While that hot house environment would be difficult to replicate, and probably is impossible to replicate on a large scale, the Weizmann still provides lessons for other science funders about what kinds of collaborations can work well.

"Usually a good collaboration is when two or three collaborators come from absolutely different and distinct disciplines, and each is able to bring a specific component to the project," Sagi said.

That thinking applies to industry-academia collaborations as well. While these have been rapidly increasing in number in recent years, Back to School argues for more.

"I would hate to see NIH take money away from basic research because NIH is not good at applied and translational research. That is the role of industry or industry-academic partnerships," said Douglas Williams, EVP of R&D at Biogen Idec Inc.

Biogen Idec has recently announced its participation in a handful of research consortia and collaborations. The company is providing funding to support academic investigators in basic research, but expects its own role will be to add value downstream via applied experiments.

In two of these - an amyotrophic lateral sclerosis (ALS) consortium with the Hudson Alpha Institute for Biotechnology; Duke University; the University of Massachusetts Medical School and Columbia University Medical Center; and the human interactome project with Harvard - the academic institutions will own resulting IP and are expected to publish the results (see BioCentury, July 2 & SciBX: Science-Business eXchange, July 26).

International collaborations are another area where greater efforts are needed.

For example, One Mind for Research aims to coordinate global multidisciplinary expertise in the neuroscience space. Launched last year, One Mind is a program sponsored by the International Mental Health Research Organization focused on precompetitive neuroscience research, education and awareness-building programs. The program involves a global coalition of neuroscientists, advocates, policy makers and pharmaceutical companies.

J&J's Janssen Pharmaceuticals Inc. unit has committed $2 million in cash and $1 million in matching donations to the program.

The program's first initiative is focused on creating a "knowledge integration network" for research on traumatic brain injury. It includes a partnership with the International Neuroinformatics Coordinating Facility to create a system for multiple-source data sharing from research efforts around the world.

The initiative's many collaborators also include data, informatics and analytics organizations such as the Neuroscience Information Framework, Recombinant Data Corp. and GNS Healthcare Inc.; research organizations such as UCSF, the University of Michigan, SAGE Bionetworks and NIH; health information-sharing website PatientsLikeMe Inc.; and therapeutic and diagnostic companies such as Eli Lilly and Co. and General Electric Co.'s GE Healthcare unit.

Another international initiative is led by the Structural Genomics Consortium, which is supported by the Canadian government, the Wellcome Trust, and six pharmaceutical companies. SGC works with partners in the U.S., the U.K., France, and Canada to determine crystal structures and develop chemical probes against human proteins of biomedical importance.

In July, an SGC-GlaxoSmithKline plc team published the first potent and selective histone demethylase inhibitor (see SciBX: Science Business eXchange, Aug. 9).

International collaboration also can help develop and strengthen expertise in countries that are trying to build a research base.

In Brazil, the Sao Paulo Research Foundation (FAPESP) has issued calls for proposals that provide funding to joint teams of researchers from local institutions working with international collaborators. FAPESP Scientific Director Carlos Henrique de Brito Cruz told BioCentury the initiative began in 2006 to fund the international exchange of researchers and now includes the co-funding of joint research projects.

FAPESP has relationships with organizations including Research Councils UK, the French National Research Agency (ANR), and the German Research Foundation (DFG).

System reset

Adopting Back to School's proposals will not fix every problem that threatens the continued productivity of the global research enterprise. Nor does Back to School have a corner on the market for solutions to the problems discussed in this essay. Other ideas should be put on the table.

No matter what the source, Back to School argues the best proposals would act synergistically to reduce waste in the system; provide greater rewards to the most innovative projects and investigators; expand the pool of ideas; and more closely tie the efforts of basic research to the desired outcome of medical breakthroughs.

For example, reforming training and peer review in parallel should together help ensure the individuals and projects with the greatest potential to transform the treatment of disease receive the funding they need.

Reforming peer review also would increase the likelihood that multidisciplinary projects could receive funding and thus also would promote the kind of collaboration that can lead to quantum leaps in knowledge.

Making meaningful communication between product developers and basic researchers a prerequisite for funding would speed the solution to logjams where therapeutics are urgently needed in areas of unmet need - particularly where diseases are poorly understood - and make it much easier to assess and measure the productivity of basic research investments.

Reducing the funding for overhead that goes to individual institutions while investing in more large scale resources that can be shared ought to not only reduce wasteful costs, but also promote collaboration.

In turn, promoting larger-scale collaboration at the institute and even government level should not only help to move the science forward faster, but also should help reduce waste in the form of redundant experiments - and especially redundant failures.

Entrenched interests will fight these reforms with politically charged warnings about the erosion of global competitiveness and the loss of high quality jobs. But global austerity will not be denied. A leaner system of training and research, one that identifies and nurtures the very best, is the only way to weather the belt tightening that is already upon us.

It also will put the research community in the best position to ask for and receive more funding when the coffers open again.

Finally, as all this transpires, the social value of research will only be realized if it is translated from bench to bedside. This will be industry's job, and it will have to live up to this obligation.

The 20th Back to School Commentary is a collaborative work led this year by BioCentury Editor Susan Schaeffer and co-writers Washington Editor Steve Usdin and Senior Writer Chris Cain. Data were developed by Research Director Walter Yang and News Editor Meredith Durkin. The package was edited by Chairman & Editor-in-Chief Karen Bernstein and President & CEO David Flores.

COMPANIES AND INSTITUTIONS MENTIONED:

Association of American Medical Colleges (AAMC), Washington, D.C.

BeiGene Ltd., Beijing, China

Biogen Idec Inc. (NASDAQ:BIIB), Weston, Mass.

Biotechnology Industry Organization (BIO), Washington, D.C.

Brandeis University, Waltham, Mass.

Brazilian Biosciences National Laboratory, Sao Paulo, Brazil

Broad Institute of MIT and Harvard, Cambridge, Mass.

Columbia University, New York, N.Y.

Duke University, Durham, N.C.

Eli Lilly and Co. (NYSE:LLY), Indianapolis, Ind.

European Commission (EC), Brussels, Belgium

European Federation of Pharmaceutical Industries and Associations (EFPIA), Brussels, Belgium

European Research Council (ERC), Brussels, Belgium

French National Research Agency (ANR), Paris, France

General Electric Co. (NYSE:GE), Fairfield, Conn.

Georgia State University, Atlanta, Ga.

German Research Foundation (DFG), Bonn, Germany

GlaxoSmithKline plc (LSE:GSK; NYSE:GSK), London, U.K.

GNS Healthcare Inc., Cambridge, Mass.

Harvard Medical School, Boston, Mass.

Howard Hughes Medical Institute, Chevy Chase, Md.

Hudson Alpha Institute for Biotechnology, Huntsville, Ala.

Innovative Medicines Initiative (IMI), Brussels, Belgium

Institute of Medicine (IOM), Washington, D.C.

Institut Pasteur Korea, Seoul, South Korea

International Mental Health Research Organization, Rutherford, Calif.

International Neuroinformatics Coordinating Facility, Stockholm, Sweden

Johnson & Johnson (NYSE:JNJ), New Brunswick, N.J.

National Academy of Sciences (NAS), Washington, D.C.

National Cancer Institute (NCI), Bethesda, Md.

National Center for Advancing Translational Sciences (NCATS), Bethesda, Md.

National Heart, Lung, and Blood Institute (NHLBI), Bethesda, Md.

National Institute on Alcohol Abuse and Alcoholism (NIAAA), Bethesda, Md.

National Institute of Allergy and Infectious Diseases (NIAID), Bethesda, Md.

National Institute of Biological Sciences (NIBS), Beijing, China

National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), Bethesda, Md.

National Institute on Drug Abuse (NIDA), Bethesda, Md.

National Institute of Mental Health (NIMH), Bethesda, Md.

National Institutes of Health (NIH), Bethesda, Md.

National Research Council (NRC), Washington, D.C.

National Science Foundation (NSF), Arlington, Va.

New York Genome Center, New York, N.Y.

One Mind for Research, Rutherford, Calif.

PatientsLikeMe Inc., Cambridge, Mass.

Recombinant Data Corp., Newton, Mass.

Research Councils UK, Swindon, U.K.

SAGE Bionetworks, Seattle, Wash.

Sao Paulo Research Foundation (FAPESP), Sao Paolo, Brazil

Singapore National Research Foundation, Singapore

Structural Genomics Consortium, Oxford, U.K.

University of California, San Francisco (UCSF), San Francisco, Calif.

University of Massachusetts Medical School, Worcester, Mass.

University of Michigan, Ann Arbor, Mich.

Vertex Pharmaceuticals Inc. (NASDAQ:VRTX), Cambridge, Mass.

The Weizmann Institute of Science, Rehovot, Israel

Wellcome Trust, London, U.K.

Whitehead Institute for Biomedical Research, Cambridge, Mass.