New friction surfaces over replicating research

Axios

24-03-2025

The Trump administration wants to spend more federal dollars replicating medical research. A key question will be which studies get repeated and, with limited resources, at what expense.
Why it matters: Many findings can't be replicated — a problem scientists say needs to be addressed. But it could also consume increasingly scarce resources as the administration cuts spending and freezes federal grants.
And some warn repeating accepted studies into how diseases originate or drugs work could undermine science for political gain.
"We should ask questions, ensure reproducibility, and grow our evidence base with replication," David Higgins, a practicing pediatrician and health services researcher at the University of Colorado Anschutz Medical Campus, said in an email.
But that "requires considering many factors," he adds.
Catch up quick: Health and Human Services Secretary Robert F. Kennedy Jr. and National Institutes of Health director nominee Jay Bhattacharya say they want to make replication a pillar of what the institutes do, pointing to fraud in the research community.
"The gold standard means real scientific research with replication of studies, which very rarely happens now at NIH," Kennedy said during a Senate confirmation hearing in January.
"We should be giving at least 20% of the NIH budgets to replication," he added, citing a landmark paper on Alzheimer's disease that was later found to contain doctored images, calling many subsequent studies into question.
In one early sign of the administration's priorities, the Centers of Disease Control and Prevention is reportedly planning a broad study into connections between vaccines and autism, despite substantial evidence disproving any link.
NIH last year launched a program that invited researchers to nominate their own studies for replication — and promised up to $50,000 plus overhead costs to contract with an outside organization to repeat the work, according to Science. Interest was "modest," the outlet reported.
What they're saying: "CDC will leave no stone unturned in its mission to figure out what exactly is happening" with the increase in autism cases in the U.S., HHS spokesperson Andrew Nixon told Axios in an email. "The American people expect high quality research and transparency and that is what CDC will deliver."
The big picture: Replication is "expensive, time-consuming and draws resources from other work but if you're interested in improving science, the scientific record and knowledge, it needs to be applied equally and universally," says Ivan Oransky, who teaches journalism at New York University and is a co-founder of Retraction Watch, which tracks withdrawals of scientific papers,
"In a world of endless resources, you should replicate every study," Oransky says.
But resources are limited and the cost of doing science has "vastly outpaced inflation," he says.
The White House already is trying to cut billions of dollars in NIH grants for research overhead at universities and medical research centers. The Trump administration has said those savings could be reallocated directly to research. But scientists and university administrators say those indirect costs are crucial for the infrastructure that enables research.
Zoom in: Reports that the CDC will conduct a new large-scale study to look into already unsupported claims of a link between autism and vaccines are raising concern about political influence. Kennedy has for years repeated the debunked theory.
"We have already done that many times over. It wastes valuable resources to revisit the same question instead of using them to address critical health challenges," Higgins said.
"More than 20 major studies involving over 10 million children across multiple countries, populations, and decades have found that there is no link between vaccines and autism," according to a new review of studies by Higgins and others.
Bhattacharya said during his Senate confirmation hearing that he doesn't "generally believe that there is a link" and doesn't want "to disprove a negative" but added that another study might help to convince people who are vaccine hesitant.
But Higgins says "reexamining settled questions that have already been repeated, replicated, and tested many times is not healthy skepticism; it's cynicism and science denial."
Research, like other investments, can be set on a spectrum of risk, says Brian Nosek, executive director of the Center for Open Science, a nonprofit that supports replication studies, and a psychology professor at the University of Virginia.
On the low-risk end is replicating studies as closely as one can to verify and validate their results. Nosek says "an investment on the order of single digits of the percent of the budget" at NIH could be helpful.
In the middle is more incremental science that comprises the bulk of research — and that has drawn the ire of politicians who characterize it as wasteful. "Incrementalism is used pejoratively and I think that is insanity for how science actually makes progress," Nosek says.
On the opposite end of the spectrum is high-risk — and potentially high-reward — research that "is open-ended and sometimes looks frivolous and impossible" but in some cases may be ultimately groundbreaking, says Stuart Buck, executive director of the Good Science Project.
Between the lines: The deeper issue underlying debates about replication and where science funding should be directed is that scientists have an incentive to build on existing studies, because it's likelier to allow them to publish often, attract more funding and advance their careers.
"Big bureaucracies in science tend to fund consensus opinion ... and to not be interested in replication and bias against groundbreaking ideas," Buck says.
Bhattacharya has said "a tentativeness to focus on the big ideas" and replicability are among the problems at NIH he'll address if confirmed, and that "no matter what the budget is, I want to reform it in that direction."