NIH employees warn of ‘institutionalized destruction’ one year after Bethesda Declaration

One year after nearly 500 National Institutes of Health (NIH) employees signed a letter warning that political interference was threatening the agency’s mission, the authors of a new report say the situation has only gotten worse. 

“The chaos of 2025 has been replaced with coordinated, systematic, institutionalized destruction in 2026,” the Bethesda Declaration: One Year Later states. The 37-page report was released on June 9, 2026 by dozens of current and former NIH employees and hosted by the grassroots advocacy group 27 UNIHTED, founded by Anna Culbertson and others after she was placed on administrative leave in February of 2025 and then later fired in May of the same year. 

“Director Bhattacharya is committed to transparency, open inquiry, and constructive debate and remains open to continuing direct conversations with the authors of the Bethesda Declaration to discuss their concerns firsthand,” said HHS Press Secretary Emily Hilliard in an email to R&D World. 

‘A living document’ 

The original Bethesda Declaration called for the restoration of canceled grants and raised alarm about what signatories described as the politicization of federally funded research. Director Jay Bhattacharya met with some of the named signers in July 2025, but the authors say their concerns went largely unaddressed. 

Culbertson described the new report as a resource designed to help advocates and lawmakers talk about the damage being done to NIH. 

“I think the hope is that this is a resource that can be used by many people who talk about the changes that have happened at NIH, like how detrimental they are: a living document,” she said. 

Culbertson worked as a program specialist in NIH’s peer review system before she was dismissed and she helped to build 27 UNIHTED in the wake of the April 2025 reduction in force to share information and legal resources among affected employees.

The new paradigm: ‘censorship’ of grants

While the mass termination of grants in early 2025 drew national press coverage and congressional scrutiny, the new report says the administration has shifted to what it describes as censorship of NIH grants, with grantees “compelled to comply with censorship” from HHS. 

Jenna Norton, an NIH program officer speaking in a personal capacity, described how the process has shifted from blunt termination to systematic censorship through computational screening.

Beginning in late 2024 and accelerating through 2025, she explained, program staff began receiving spreadsheets flagging grants whose language was potentially “misaligned with agency priorities,” based on an algorithm scanning for banned terms. Staff were then required to go to grantees and ask them to renegotiate the language of their grants in order to receive funding. 

“The full list [of banned words] is kept very close,” Norton said. Program officers only received the specific words flagged within individual grants, not the master list. 

Racial and ethnic identifiers like “African American” and “Hispanic American” were flagged for a period last year but have since been removed from the list, Norton said. Terms like “social determinants of health” and “policy” have been added, she said. 

Now, a team within the NIH’s Office of Extramural Research, and sometimes the HHS, conducts additional review of renegotiated grants, sometimes asking for further changes.  

“Sometimes they request things that are absurd, like changing the outcome three years into a study,” Norton said. The NIH did not respond to a request for comment about changing the outcomes of studies. 

Science reported that HHS staff now weigh in on the scientific substance of NIH grants and have requested substantive changes to ongoing studies, including recommending that a depression study add genetic influences in its fourth year. An analysis by former NIH official Jeremy Berg found that more than 700 multi-year NIH grants had their titles changed in 2025 to avoid terms such as “equity” and “disparity,” STAT reported. Bhattacharya has denied that a banned-words list exists, and an HHS spokesperson has said claims that NIH issued a banned-words list are “unequivocally false.”

Bhattacharya has stated publicly that research designed to inform policymakers is now out of scope for the agency, though this guidance has not been put in writing, Norton said. “It is an absurd argument, because to the extent that we are charged with public health, most public health is implemented through policy.”

Bhattacharya does not appear to have addressed the policy-research claim directly on the record. In announcing the agency’s new research priorities in 2025, he ordered a review of current and planned NIH research for alignment with those priorities, Science reported. He has also said that DEI-related grants reinstated by court order will not be renewed, STAT reported. HHS did not respond to a request for comment on whether policy-focused research is out of scope for the agency.

Norton said the deliberate vagueness of the process appears designed to generate overcompliance. “When you create an environment of fear, where people believe they will experience retaliation if you don’t do what they want, and then don’t tell them where the line is, they’re going to guess, and then take five steps back from the line to be safe,” she said. 

For researchers navigating this environment, Norton had some advice. “My advice is to write their application as they would have always written it,” she said, adding that self-censorship actually undermines researchers in peer review. 

Reviewers are still evaluating scientific coherence, and grants stripped of necessary terminology often fail to explain their own aims. And since the banned word list is constantly evolving, pre-emptive self-censorship will likely be wrong by the time a renegotiation actually happens.

“If the scientific community collectively stops using all of the words that we think might possibly be screened, then we actually lose our ability to know what is getting screened,” Norton added.

There is an appeals process for terminated grants, but Norton said she did not see a single successful appeal in her portfolio. 

Constant changes to foreign collaboration

Rui Carlos Pereira de Sá, a program director at NIH and an author on the new report speaking in a personal capacity, said the rapid-fire policy changes around foreign collaboration have been damaging research. Sá manages a portfolio of over 60 grants, four of which were terminated last year for having foreign components. 

Two of the terminated grants involved research on maternal health conditions with low incidence rates in the U.S., so data must be collected abroad, he explained. 

“What could be done in about a year in Kenya and Uganda for different diseases would take five-plus years” in the U.S., given lower incidence rates, Sá explained.

One of the studies reached approximately 80% of its target data acquisition before the collaboration had to be shut down. Another was less than halfway through and, with two and a half years of investment already spent, is unlikely to fully deliver on its scientific promise. 

“That project is probably going to be unable to fully deliver on its promise, just because of external forces that they do not control,” Sá said.

The policy around foreign collaboration today requires that each foreign component of a grant be carved out into a separate award, which Sá described as “triple the work” and, in some cases, a significant waste of taxpayer money for grants where the foreign collaboration is only a fraction of the budget. 

A new funding mechanism called PF5, intended to allow foreign partnerships to resume through U.S.-based applicants, has just entered its first review cycle. Sá said it remains burdensome and discouraging for applicants. 

None of these mechanisms were developed with input from NIH program staff. “Typically, these things have a work group, take years to come alive,” Sá said. Staff were informed of this draft policy after it had already been finalized. 

“We had no clear guidance. And because they didn’t get input, some of their initial policies were not really enforceable or applicable, so they had to come back and change them and then change them a third time,” Sá said.