After already killing more than 1700 active grants related to politically charged topics, the National Institutes of Health (NIH) has flagged another roughly 3200 grants for review and possible termination, Science has learned. Some newly flagged grants had already been cleared by previous reviews or have no obvious link to controversial issues.
The new agency review comes as a federal judge is poised to consider the legality of such terminations in a 16 June hearing. Plaintiffs in that lawsuit say NIH may be violating its obligations to turn over relevant documents as part of the case’s discovery process. The agency claims it has no official guidance documents on how to select grants to kill—or that they are privileged and don’t need to be revealed.
But Science has obtained several internal NIH directives on that matter, revealing inconsistent methods and sometimes conflicting definitions of review terms.
Researchers and state attorneys general challenging the grant terminations that have taken place since President Donald Trump assumed office in January have asked NIH to reveal its process to the court, as well as to disclose communications that show whether such decisions were made by the agency or the so-called Department of Government Efficiency (DOGE) initially run by billionaire Elon Musk. In communications with these plaintiffs, NIH denied that such records exist or said they were privileged information.
But each of NIH’s 27 institutes and centers (ICs) were explicitly ordered to come up with their own versions of guidelines delineating which projects should and should not be funded under the agency’s new “priorities,” according to an internal email. It’s not clear whether every IC did so, but Science has seen such guidelines from five different ICs and they vary widely. They include, for instance, different definitions of what constitutes diversity, equity, and inclusion (DEI)—a topic the agency says it no longer funds. Some ICs’ guidelines include lists of keywords such as “gender,” “justice,” “marginalized,” or “climate” that should trigger extra attention by NIH program officers (POs)—scientifically trained employees who manage grants.
As previously reported by Science and others, the POs have been directed to put any flagged grants into one of three categories—acceptable, unacceptable, or possibly modifiable—before sending them back for review by NIH or DOGE officials who determine whether to cancel them. An NIH PO who asked for anonymity out of fear of retaliation said NIH staff are befuddled by the disparate rules. “What is fundable in one institute would not be in another,” the PO said. “It’s crazymaking.”
At the core of the court hearing on Monday—a partial consolidation of two lawsuits by various parties—is the question of whether the earlier terminations were “arbitrary and capricious,” meaning they did not involve reasoned decision-making. Scott Delaney, a Harvard University epidemiologist and former attorney who maintains a database of terminated NIH and National Science Foundation grants, says the guidance documents in some ways support the government’s case that their decisions were not arbitrary.
But the wide variation reflects the challenge that NIH employees have faced in making sense of directives from the administration, he says. “What you see in the end is no cohesive policy and a lot of elements that could lead a court to say it’s arbitrary and capricious and vague and therefore unlawful.”
In response to a 9 June court filing in which the plaintiffs complained that NIH had not produced information on the cancellation process they had requested, attorneys representing NIH countered that any such documents were “deliberative,” meaning they did not reflect the government’s final decisions and did not need to be disclosed. Additionally, they said, the “plaintiffs have not identified clear, specific evidence that any such guidance exists.”
But in a 5 May email to POs, Jon Lorsch, NIH’s acting deputy director for extramural research, wrote that “all ICs need to have processes in place for screening all awards.” He said several grants on unapproved topics had recently been funded, contrary to NIH policy. “The consequences of approving an award that should not have been approved could be very serious for NIH and your IC, so please ensure that you have the needed processes in place,” he wrote in the memo, first reported on by Nature.
Each set of institute guidelines viewed by Science contains somewhat different instructions on how to screen grants. For example, at the National Institute of Child Health and Human Development, a brief February document lists just five topics to look for—research on gender identity, increasing workforce diversity, climate justice or climate change, foreign institutions, and enhancing diverse perspectives. At the National Heart, Lung, and Blood Institute, seven pages of guidance last updated on 22 May list the same topics but also include vaccine hesitancy and COVID-19. The document says a grant containing “diversity” does not need to be canceled if the word refers to animals, cell types, and other basic research terms. And it includes a list of 18 terms to search for, such as “marginalized,” “ethnic,” and “foreign.”
The National Institute of Mental Health’s seven-page document has a list of 20 terms but they are slightly different, and include, for example, “vulnerable.” This guidance also says to look for “anything that may appear controversial.” The National Institute of General Medical Sciences’s document is by far the longest and most detailed at 14 pages, containing more grant categories than others. And a National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) document has unique wording on studies of specific populations. It states that the grants cannot study only a specific group defined by race, color, religion, sex, or national origin—suggesting, the PO says, that a study of pregnant people would not be funded.
“This is some of the first evidence we have about how DEI is being interpreted and defined within certain ICs,” Delaney says, noting that NIDDK’s guidelines are particularly unusual. “Without such studies, the vast majority of science is in white folks.”
Many NIH observers have assumed DOGE is telling the agency specific grants that must be reviewed, and recent internal emails provide further evidence of that. In a 10 June email seen by Science, Lorsch asked POs to “please turn in your categorizations for the first two batches of grants DOGE has asked us to review.” Another 9 June email from Michelle Bulls, director of NIH’s Office of Policy for Extramural Research Administration, said DOGE had flagged yet another set of grants as requiring additional review by the ICs.
The role of DOGE is key, say the lawsuits’ plaintiffs, because it suggests NIH is canceling grants because of external influence and not NIH priorities. DOGE’s role had previously been confirmed in another lawsuit in depositions by NIH officials, as first reported by ProPublica.
Science has obtained three spreadsheets detailing the 3200 newly flagged grants. Many are on topics targeted by the Trump administration—transgender health, racially diverse populations, and COVID-19, to name a few. The lists also include numerous training grants and centers that aim to recruit racial minorities into trials on Alzheimer’s and other diseases.
But some grants do not appear to fall into any of the categories previously banned by NIH. A number deal with women’s and children’s health, or issues such as alcohol abuse or violence that disproportionately affect impoverished communities. One project on the list, for instance, seeks the causes of respiratory wheezing in economically disadvantaged preschool children. Another studies how wildfire smoke affects children with asthma. A few are even more mystifying, such as a project studying how social factors affect captive research primates’ life spans.
The PO who spoke with Science confirmed that the inclusion of many grants on the new lists can’t be explained by NIH’s screening procedures up to this point. “Program staff have to guess what led them to end up on this list and try to write a justification in response,” they said. “And then they may get sent back to us again, as many on these lists were already screened and approved. No one knows what to do.” NIH has not responded to a request for comment.
