Earlier this year, the National Institutes of Health (NIH) made an unusual offer to many of its 37,500 principal investigators: If you have a laboratory study you think could have a major impact on health—such as a mouse experiment testing a possible heart disease drug—we may pay for a contract lab to repeat the work to make sure it’s solid.
Only a few people applied to the pilot phase of NIH’s initiative, which is finalizing its picks for the first handful of studies this month. But its leader at NIH says it has enough participants to study the feasibility of the program, which has the support of Congress and President-elect Donald Trump’s nominee to head NIH, Stanford University health economist Jay Bhattacharya. He recently told The Wall Street Journal that replication studies should be “a centerpiece of what the NIH does.”
The initiative comes with a big caveat: The agency has no plans to make the resulting data public. That “limits the appeal and value,” says Tim Errington of the Center for Open Science (COS), a nonprofit that supports replication studies. Still, he says, the pilot “is a step in the right direction.”
For years, concerns have mounted that many basic biomedical experiments don’t hold up when another lab attempts them, casting doubt on plans to translate the work into a treatment. Cases of apparent scientific fraud, such as work underlying Alzheimer’s disease drugs that Science investigated, have added to worries about the integrity of these preclinical studies.
A string of recent efforts has sprung up to test reproducibility. One high-profile COS initiative tested key experiments from 23 high-impact papers in cancer biology. It found after 8 years that only a fraction could be fully replicated. Other countries have launched their own programs—the main Dutch science agency has invited researchers to repeat landmark studies, with mixed success, and a Brazilian nonprofit and Germany’s science ministry both have replication efforts underway.
NIH is joining in with a nudge from Congress, which urged the agency “to establish a program to fund replication experiments on significant lines of research” as well as monitor for scientific fraud, according to a report accompanying a funding bill this year. (The U.S. House of Representatives suggested $50 million for the replication and fraud efforts and the Senate $10 million, but the final bill did not specify an amount.)
Instead of selecting the work to be replicated, NIH invited researchers to nominate their own NIH-funded studies—even unpublished ones. Program managers looked for those that have “high potential impact on public health.” (Clinical studies and primate experiments would take too much time and money, but NIH has indicated other animal studies could qualify.)
NIH is then giving the investigator up to $50,000, plus overhead costs, to help a contract research organization (CRO) arrange the exact same techniques, protocols, and reagents. The two participating contract labs—one focusing on technologies, the other experiments—have separately received $2 million in total to complete the work by January 2026.
Two requests for proposals in August and November drew a “modest” response, says Douglas Sheeley, who helps manage NIH’s Common Fund, which is running the initiative. NIH can’t release proposal numbers, but at most, 32 people total signed up for its three webinars on the proposal process. Still, Sheeley expects to fund “around a half-dozen different projects.”
Unlike some other replication projects, NIH has decided not to publicize the results. One concern is protecting intellectual property. But NIH also wants to build “a trusted relationship between the NIH and the researchers,” Sheeley says. “Right now, we’re focused on just making sure that we do the pilot well and learn as much as we can from it,” including the cost and challenges of the replications.
The CROs will submit “aggregated, anonymized data” to NIH, according to an FAQ page. The original investigator and contract lab are also free to publish the full results. But, “We’re not going to force that,” Sheeley says. (The curious will, however, be able to see the funded projects in NIH’s grants database, NIH Reporter, in February 2025.)
One scientist involved with the cancer biology replications is skeptical. Stem cell biologist Sean Morrison of the University of Texas Southwestern Medical Center, who edited some of the papers resulting from that project, notes the contract labs didn’t have the resources to do pilot experiments or repeat studies to work out kinks. And such labs often lack “the expertise of academic laboratories, especially when it comes to advanced or specialized techniques,” he says. In some cases, “this led to uninterpretable results.” The murky outcome of some replication attempts may have left scientists behind the original studies feeling their reputation was unfairly tarnished, he adds.
In NIH’s better funded effort, investigators will work more closely with the CROs than they did in the cancer biology project, where communication was often minimal. Still, Morrison says, that project sounds a note of caution for future replication efforts: “It’s not easy to test the reproducibility of results at scale.”
More: https://www.science.org/content/article/nih-launches-initiative-double-check-biomedical-studies
