Accusations that the leading U.S. funders of biomedical research "ignore truly innovative thinkers" and "encourage conformity if not mediocrity" are seldom heard in the polite precincts of top science journals. Yet they are front and center in a paper published Wednesday in the journal Nature, which concludes that fewer than half of America's most influential and productive biomedical scientists now receive funding from the National Institutes of Health.

Critics have long argued that NIH, which spends some $30 billion a year on biomedical research at universities and medical centers worldwide, funds conventional, incremental science rather than swing-for-the-fences studies more likely to produce breakthroughs. But the new analysis goes further: It marshals data to show that U.S. biomedical researchers who make the most influential discoveries are not getting NIH support.

"I was astonished by the findings," said Jack Dixon, vice president and chief scientific officer of the nonprofit Howard Hughes Medical Institute (HHMI), who was not involved in the study. "It's just amazing that most of NIH's $30 billion is going to scientists who haven't had the greatest impact."

NIH Director Francis Collins said the agency is "always trying to improve" how it evaluates grant proposals and has recently launched several programs to explicitly support groundbreaking science, which he concedes can get lost in the bureaucracy.

The process of deciding which studies NIH should fund "can be very conservative," said Collins. "Faced with a very innovative, intriguing proposal that could crash and burn versus an incremental one, reviewers can slip in the direction of the sure thing" as they try to ensure that there will be some result, even a less-than-groundbreaking one, for taxpayers' money.


For their analysis, Dr. John Ioannidis, director of the Prevention Research Center at Stanford School of Medicine in Palo Alto, California, and graduate student Joshua Nicholson of Virginia Tech in Blacksburg identified the 700 biomedical papers published since 2001 that had been cited at least 1,000 times. (Citations, in which one paper refers to another, are considered to be the best measure of a study's impact on a particular field.)

Of those papers' primary authors, only 40 percent of those not serving on NIH panels are currently funded by NIH. Conversely, a mere 0.8 percent of the scientists in these "study sections," which evaluate some 80,000 grant proposals a year, are primary authors of 1,000-citation papers.

"Not only do the most highly cited authors not get funded," the Nature authors write. "Worse, those who influence the funding process are not among those who drive the scientific literature."

Although members of study sections fall short in scientific impact, they do well in the funding sweepstakes. They are significantly more likely to receive NIH money than are scientists who do not volunteer for the panels. Yet their work tends to be similar to other NIH-funded studies, the Nature study found, not envelope-pushing research.

"Even without conscious bias," said Ioannidis, "members of study sections are likely to prefer avenues of research they have gone down themselves rather than bold innovation. These are people who are good at networking and at promoting their own research agenda. Exceptional creative ideas may have a hard time."

Top scientists are familiar with NIH's penchant for the safe and incremental. Years ago, biologist Mario Capecchi of the University of Utah applied for NIH funding for a genetics study with three parts. The study section liked two of them but said the third would not work.

Capecchi got the grant and put all the money into the part the reviewers discouraged. "If nothing happened, I'd be sweeping floors now," he said. Instead, he discovered how to disable specific genes in animals and shared the 2007 Nobel Prize for medicine for it.

NIH's Collins pointed out possible flaws in the new analysis, including that citations might not be the best way to identify the most influential studies. In fields with only a few scientists, He said, even blockbuster papers might fall short of 1,000 citations. And as the Nature authors concede, the reason NIH supports so few high-impact scientists is not necessarily that it shies from creative, risk-taking studies but that some of these scientists might have stopped doing research.

NIH has not barred the door to original thinkers; scientists it has funded have won 135 Nobel prizes. "Innovative" is one of five criteria that panels are told to consider when they weigh grant proposals, and in recent years NIH has established new grants explicitly for what it calls "pioneering -- and possibly transforming -- research."

After reading the newspaper before publication, Collins asked a colleague to identify the 10 most influential papers published in the world's top two science journals since 2007. "One hundred percent were funded by NIH," he said.


Nevertheless, top scientists argue that NIH could be bolder. The year before his Nobel, Capecchi said, NIH denied his grant proposal. He instead received funding from HHMI, which spends some $800 million a year to support biomedical research.

"Hughes has a different model," Capecchi said. "They encourage us to take risks" -- that is, to try studies that might fail but might change the world.

HHMI's Dixon, who has served on NIH study sections, goes further. Unlike NIH, HHMI "embraces risk-taking," he said. "We ask, 'Where would your field be if you weren't there?' "

That results in "more highly cited papers as well as more duds," said Pierre Azoulay of the Massachusetts Institute of Technology, who compared NIH- and HHMI-funded studies. "In terms of creativity, Hughes investigators are doing more. The way the NIH system is set up, proposals that do the worst are ones that cause passionate disagreement, which tend to be those that are very creative and out there."

Capecchi heard years later from the NIH panel that discouraged him from his genetic moon shot: "We are glad that you didn't follow our advice."