The US National Institutes of Health (NIH) has proposed overhauling how it decides which research projects to fund, and how it reports the outcome of those deliberations. For the past five decades, the agency reported to researchers the precise numerical scores assigned to their project proposals by peer reviewers. The new plan would instead see the agency report whether projects were ranked in one of three broad categories — a top, middle and bottom — based on their merit.

Inside the new political screening that’s stalling NIH grants

The scheme, which was announced on Friday, is part of a shift away from relying on scientific peer review in favour of greater discretion to award funds based on agency priorities. The new categories encourage the NIH’s 27 institutes and centres to balance competing factors when ranking proposals — including peer review, but also “health priorities” and “the wider research portfolio”, the announcement says.

Proponents of the change argue that the current scoring system creates a false sense of precision, given that a particular proposal’s score is typically an average of the opinions of just three peer reviewers. But given the increased influence of political appointees in federal grant decisions under the administration of US President Donald Trump, some scientists say that removing the transparency of exactly how a proposal was scored could open the door to further political interference.

Jenna Norton, a programme official at the NIH who organized a protest letter against changes at the agency under Trump, says that if researchers never know how high their score is, they are less likely to raise an alarm if their project is rejected, perhaps in favour of a project that more closely aligns with the administration’s political priorities.

“This change is quite substantial,” says Michael Lauer — who for a decade ran the NIH’s extramural research programme, which funds scientists across the United States. However, he adds, “if it were not for the concerns about political interference, there’s a lot here that makes sense”.

A new system

To evaluate the scientific merit of grant applications, the NIH, the largest public funder of biomedical research, usually convenes specialized review panels of independent scientists from across the United States. First, a minimum of three expert reviewers from a panel are assigned a proposal to give a preliminary score. After the scores are averaged, only those proposals in the top 50th percentile advance for a discussion among the full panel’s reviewers, who provide a final ‘impact score’.

NIH limits funding for research on the health effects of public policy

These scores and the panel’s written critiques are then shared with the researchers who submitted the proposals, their institutions and agency programme staff members. Each of the NIH’s institutes has an advisory council, composed of a different panel of outside scientists, that then reviews applications they receive. These councils make recommendations on which projects to fund, and the institute directors have the final say.

Under the new plan, the NIH would stop sharing impact scores with researchers, university officials and the agency’s own programme staff members. Instead, applications would be grouped into three distinct categories: ‘most competitive’ for the top 25th percentile of scores, ‘competitive’ for those between the 26th and 50th percentiles and ‘not discussed’ for the bottom half.

Lauer says that “there is strong scholarship showing that peer review is inherently biased against novelty”, so this proposal could nudge the agency to fund more innovative projects by reducing bias tied to a precise score and instead treating applications that score in the top quartile similarly. Norton says that encouraging more innovative research is important, but that it would be best accomplished through ‘targeted solicitations’, where the agency calls for research proposals on specific scientific topics or gaps. Under the Trump administration, the NIH has issued a small fraction of the targeted solicitations it did before 2025.

NIH pivots away from agency-directed science

Still, Norton and other researchers with whom Nature spoke agree that the current peer-review system has its flaws. For example, although the system successfully discerns science worth funding from science not yet worth funding (scores below the 50th percentile), scientific papers that result from NIH grants receiving the very highest scores are cited nearly the same amount as those in the 20th percentile. That’s according to an analysis by Ferric Fang¹, a molecular bacteriologist at the University of Washington in Seattle, and his colleagues.

“When it comes down to the top quarter of grants, it becomes much more arbitrary, the criteria you use to rank them, because all these grants have a lot of merit,” Fang says. He and his colleagues have advocated for a lottery-style system that awards funds randomly among applications that score above a certain threshold, a scheme that is becoming more popular, particularly in Europe.