The US National Institutes of Health has seen grant applications held up by extra layers of scrutiny.Credit: Mark Wilson/Newsmakers/Getty
Hundreds of grant applications to the US National Institutes of Health (NIH) are being held up at any given time by unprecedented scrutiny after peer review. Some have been flagged by an algorithm for using terms, such as ‘gender’ and ‘climate change’, that have been deemed not to conform with the priorities of the administration of US President Donald Trump.
These new layers of review have delayed delivery of funds to labs and research institutions — and have even resulted in the outright rejection of some applications that had been approved by outside and agency scientists. Before 2025, it was unheard of for grants that had received such approvals to be rejected, says one of six NIH officials who spoke to Nature on condition of anonymity. The extra scrutiny ramped up in early 2026 and is conducted by NIH leaders and officials at the NIH’s parent agency, the Department of Health and Human Services (HHS).
The extra scrutiny has particularly affected researchers waiting for previously approved projects to be renewed. Most applications clear the supplemental review in two weeks, but 10% of the grant-renewal applications that have entered this phase in the 2026 fiscal year have become stuck for more than seven weeks, and some have been held up indefinitely, according to data and internal e-mails that Nature has obtained (see ‘Holding pattern’).
Source: US National Institutes of Health data obtained by an agency programme officer.
Similar reviews could be coming to other federal agencies. On 29 May, the White House proposed rules that would give political appointees unheard-of control over scientific grant-making.
“We’re the test case,” says an NIH programme officer. “The new rules would essentially codify the administration’s ability to restrict funding on anything because they don’t like the topics or the words” used in the applications.
The Trump administration has criticized the federal grant-making process as opaque and unaccountable. Grants have been used “to promote a ‘woke’ policy agenda that did not reflect the values of the vast majority of the American public”, says the 29 May proposal to overhaul grant-making.
In response to Nature‘s queries about the grant-screening process, an NIH spokesperson said that “there are no ‘banned word’ lists at NIH, and funding decisions are not based on specific words or phrases.” Internal documents suggest that the agency does not ban these words, but projects that use them are much more likely to be scrutinized by programme officers who use the tool and by NIH and HHS officials reviewing the project.
Automated check
Weeks after Trump took office in January 2025, his administration began terminating thousands of grants on disfavoured topics, such as diversity, equity and inclusion, climate change, misinformation, vaccine hesitancy and what it called ‘gender ideology’. Billionaire Elon Musk’s Department of Government Efficiency (DOGE) directed these terminations, Nature previously reported.
As DOGE’s influence in government waned, NIH leaders formalized a screening process by creating an algorithm that checks applications for a list of disfavoured terms (see Supplementary information). As of February, the list included 235 terms, which serve as an “indicator that the grant likely needs, at a minimum, a revision to align its terminology with NIH priorities”, according to one internal document.
NIH grant cuts disproportionately hit minority and female scientists
The tool checks every application’s title, abstract and other summary sections for terms including ‘racism’, ‘fossil fuel’, ‘queer’ and ‘sexual minorit’ (referring to all forms of the phrase ‘sexual minority’).
This algorithm is applied after applications have passed two rounds of peer review, and after programme officials and the relevant institute director have deemed them worthy of funding and in alignment with institute priorities (see ‘How an NIH grant is processed’).
If the application contains one of the 235 disfavoured terms, it is classified as “not clean”. Programme officers are then instructed to renegotiate the language or scope of the proposed project with the applicant, or to scrap it altogether. Officers must also write a ‘decision memo’ that describes the actions taken.
Grant limbo
After this step, applications enter a phase called Status 19. Historically, the NIH has used this status to give members of the US Congress advance notice of the projects being funded in their districts.
Now, this designation also gives NIH leaders and the “HHS counselor” an opportunity to review all applications, according to internal documents that Nature reviewed.
Feedback from both NIH leaders and HHS officials sometimes questions whether the applications should be funded, according to the documents. “Does [the institute] really want to fund this study? Seems likely to end up in a Congressional waste report,” reads one comment, referring to analyses that highlight examples of allegedly improper federal spending.
US science after a year of Trump: what has been lost and what remains
In such cases, the application is sent back to the handling NIH programme officer to remedy the issue, often requiring that they renegotiate the scope of the project with the applicant before resubmitting the grant for subsequent rounds of Status 19 review. Sometimes, e-mails reprimand the people who processed the applications: “There were a lot of problematic [National Institute on Aging] grants in the queue this week”, says one e-mail from NIH leaders in March. “Many of them had significant problems that should have been caught by your staff.”
Applications are often sent back because they include disfavoured topics or have foreign collaborators on projects that do not show a “clear benefit” to the United States. Applications are also bounced back to programme officers for alleged “lobbying activities”, a reference to projects that aim to inform policy; or for containing “vague and non-specific language”, such as the phrase ‘social determinants of health’, which is deemed not to include “measurable concepts”, according to e-mails that Nature obtained.
