Sahil Lavingia, who wrote the code, told it to cancel, or in his words “munch,” anything that wasn’t “directly supporting patient care.” Unfortunately, neither Lavingia nor the model had the knowledge required to make such determinations.

“I think that mistakes were made,” said Lavingia, who worked at DOGE for nearly two months, in an interview with ProPublica. “I’m sure mistakes were made. Mistakes are always made.”

It turns out, a lot of mistakes were made as DOGE and the VA rushed to implement President Donald Trump’s February executive order mandating all of the VA’s contracts be reviewed within 30 days.

ProPublica obtained the code and prompts — the instructions given to the AI model — used to review the contracts and interviewed Lavingia and experts in both AI and government procurement. We are publishing an analysis of those prompts to help the public understand how this technology is being deployed in the federal government.

The experts found numerous and troubling flaws: the code relied on older, general-purpose models not suited for the task; the model hallucinated contract amounts, deciding around 1,100 of the agreements were each worth $34 million when they were sometimes worth thousands; and the AI did not analyze the entire text of contracts. Most experts said that, in addition to the technical issues, using off-the-shelf AI models for the task — with little context on how the VA works — should have been a nonstarter.

Lavingia, a software engineer enlisted by DOGE, acknowledged there were flaws in what he created and blamed, in part, a lack of time and proper tools. He also stressed that he knew his list of what he called “MUNCHABLE” contracts would be vetted by others before a final decision was made.

Portions of the prompt are pasted below along with commentary from experts we interviewed. Lavingia published a complete version of it on his personal GitHub account.

Problems with how the model was constructed can be detected from the very opening lines of code, where the DOGE employee instructs the model how to behave:

This part of the prompt, known as a system prompt, is intended to shape the overall behavior of the large language model, or LLM, the technology behind AI bots like ChatGPT. In this case, it was used before both steps of the process: first, before Lavingia used it to obtain information like contract amounts; then, before determining if a contract should be canceled.

Including information not related to the task at hand can confuse AI. At this point, it’s only being asked to gather information from the text of the contract. Everything related to “munchable status,” “soft-services” or “DEI” is irrelevant. Experts told ProPublica that trying to fix issues by adding more instructions can actually have the opposite effect — especially if they’re irrelevant.

The models were only shown the first 10,000 characters from each document, or approximately 2,500 words. Experts were confused by this, noting that OpenAI models support inputs over 50 times that size. Lavingia said that he had to use an older AI model that the VA had already signed a contract for.

This portion of the prompt instructs the AI to extract the contract number and other key details of a contract, such as the “total contract value.”

This was error-prone and not necessary, as accurate contract information can already be found in publicly available databases like USASpending. In some cases, this led to the AI system being given an outdated version of a contract, which led to it reporting a misleadingly large contract amount. In other cases, the model mistakenly pulled an irrelevant number from the page instead of the contract value.

“They are looking for information where it’s easy to get, rather than where it’s correct,” said Waldo Jaquith, a former Obama appointee who oversaw IT contracting at the Treasury Department. “This is the lazy approach to gathering the information that they want. It’s faster, but it’s less accurate.”

Lavingia acknowledged that this approach led to errors but said that those errors were later corrected by VA staff.

Once the program extracted this information, it ran a second pass to determine if the contract was “munchable.”

Again, only the first 10,000 characters were shown to the model. As a result, the munchable determination was based purely on the first few pages of the contract document.

The above prompt section is the first set of instructions telling the AI how to flag contracts. The prompt provides little explanation of what it’s looking for, failing to define what qualifies as “core medical/benefits” and lacking information about what a “necessary consultant” is.

For the types of models the DOGE analysis used, including all the necessary information to make an accurate determination is critical.

Cary Coglianese, a University of Pennsylvania professor who studies the governmental use of artificial intelligence, said that knowing which jobs could be done in-house “calls for a very sophisticated understanding of medical care, of institutional management, of availability of human resources” that the model does not have.

The prompt above tries to implement a fundamental policy of the Trump administration: killing all DEI programs. But the prompt fails to include a definition of what DEI is, leaving the model to decide.

Despite the instruction to cancel DEI-related contracts, very few were flagged for this reason. Procurement experts noted that it’s very unlikely for information like this to be found in the first few pages of a contract.

These two lines — which experts say were poorly defined — carried the most weight in the DOGE analysis. The response from the AI frequently cited these reasons as the justification for munchability. Nearly every justification included a form of the phrase “direct patient care,” and in a third of cases the model flagged contracts because it stated the services could be handled in-house.

The poorly defined requirements led to several contracts for VA office internet services being flagged for cancellation. In one justification, the model had this to say:

The contract provides data services for internet connectivity, which is an IT infrastructure service that is multiple layers removed from direct clinical patient care and could likely be performed in-house, making it classified as munchable.

Despite these instructions, AI flagged many audit- and compliance-related contracts as “munchable,” labeling them as “soft services.”

In one case, the model even acknowledged the importance of compliance while flagging a contract for cancellation, stating: “Although essential to ensuring accurate medical records and billing, these services are an administrative support function (a ‘soft service’) rather than direct patient care.”

Shobita Parthasarathy, professor of public policy and director of the Science, Technology, and Public Policy Program at University of Michigan, told ProPublica that this piece of the prompt was notable in that it instructs the model to “distinguish” between the two types of services without instructing the model what to save and what to kill.

The emphasis on “direct patient care” is reflected in how often the AI cited it in its recommendations, even when the model did not have any information about a contract. In one instance where it labeled every field “not found,” it still decided the contract was munchable. It gave this reason:

Without evidence that it involves essential medical procedures or direct clinical support, and assuming the contract is for administrative or related support services, it meets the criteria for being classified as munchable.

In reality, this contract was for the preventative maintenance of important safety devices known as ceiling lifts at VA medical centers, including three sites in Maryland. The contract itself stated:

Ceiling Lifts are used by employees to reposition patients during their care. They are critical safety devices for employees and patients, and must be maintained and inspected appropriately.

This portion of the prompt attempts to define “soft services.” It uses many highly specific examples but also throws in vague categories without definitions like “non-performing/non-essential contracts.”

Experts said that in order for a model to properly determine this, it would need to be given information about the essential activities and what’s required to support them.

This section of the prompt was the result of analysis by Lavingia and other DOGE staff, Lavingia explained. “This is probably from a session where I ran a prior version of the script that most likely a DOGE person was like, ‘It’s not being aggressive enough.’ I don’t know why it starts with a 2. I guess I disagreed with one of them, and so we only put 2, 3 and 4 here.”

Notably, our review found that the only clarifications related to past errors were related to scenarios where the model wasn’t flagging enough contracts for cancellation.

This section of the prompt provides the most detail about what constitutes “direct patient care.” While it does cover many aspects of care, it still leaves a lot of ambiguity and forces the model to make its own judgements about what constitutes “proven efficacy” and “critical” medical equipment.

In addition to the limited information given on what constitutes direct patient care, there is no information about how to determine if a price is “reasonable,” especially since the LLM only sees the first few pages of the document. The models lack knowledge about what’s normal for government contracts.

“I just do not understand how it would be possible. This is hard for a human to figure out,” Jaquith said about whether AI could accurately determine if a contract was reasonably priced. “I don’t see any way that an LLM could know this without a lot of really specialized training.”

This section explicitly lists which tasks could be “easily insourced” by VA staff, and more than 500 different contracts were flagged as “munchable” for this reason.

“A larger issue with all of this is there seems to be an assumption here that contracts are almost inherently wasteful,” Coglianese said when shown this section of the prompt. “Other services, like the kinds that are here, are cheaper to contract for. In fact, these are exactly the sorts of things that we would not want to treat as ‘munchable.’” He went on to explain that insourcing some of these tasks could also “siphon human sources away from direct primary patient care.”

In an interview, Lavingia acknowledged some of these jobs might be better handled externally. “We don’t want to cut the ones that would make the VA less efficient or cause us to hire a bunch of people in-house,” Lavingia explained. “Which currently they can’t do because there’s a hiring freeze.”

The VA is standing behind its use of AI to examine contracts, calling it “a commonsense precedent.” And documents obtained by ProPublica suggest the VA is looking at additional ways AI can be deployed. A March email from a top VA official to DOGE stated:

Today, VA receives over 2 million disability claims per year, and the average time for a decision is 130 days. We believe that key technical improvements (including AI and other automation), combined with Veteran-first process/culture changes pushed from our Secretary’s office could dramatically improve this. A small existing pilot in this space has resulted in 3% of recent claims being processed in less than 30 days. Our mission is to figure out how to grow from 3% to 30% and then upwards such that only the most complex claims take more than a few days.

The engineer, working for the Department of Government Efficiency, quickly built an artificial intelligence tool to identify which services from private companies were not essential. He labeled those contracts “MUNCHABLE.”

The code, using outdated and inexpensive AI models, produced results with glaring mistakes. For instance, it hallucinated the size of contracts, frequently misreading them and inflating their value. It concluded more than a thousand were each worth $34 million, when in fact some were for as little as $35,000.

The DOGE AI tool flagged more than 2,000 contracts for “munching.” It’s unclear how many have been or are on track to be canceled — the Trump administration’s decisions on VA contracts have largely been a black box. The VA uses contractors for many reasons, including to support hospitals, research and other services aimed at caring for ailing veterans.

VA officials have said they’ve killed nearly 600 contracts overall. Congressional Democrats have been pressing VA leaders for specific details of what’s been canceled without success.

We identified at least two dozen on the DOGE list that have been canceled so far. Among the canceled contracts was one to maintain a gene sequencing device used to develop better cancer treatments. Another was for blood sample analysis in support of a VA research project. Another was to provide additional tools to measure and improve the care nurses provide.

ProPublica obtained the code and the contracts it flagged from a source and shared them with a half dozen AI and procurement experts. All said the script was flawed. Many criticized the concept of using AI to guide budgetary cuts at the VA, with one calling it “deeply problematic.”

Cary Coglianese, professor of law and of political science at the University of Pennsylvania who studies the governmental use and regulation of artificial intelligence, said he was troubled by the use of these general-purpose large language models, or LLMs. “I don’t think off-the-shelf LLMs have a great deal of reliability for something as complex and involved as this,” he said.

Sahil Lavingia, the programmer enlisted by DOGE, which was then run by Elon Musk, acknowledged flaws in the code.

“I think that mistakes were made,” said Lavingia, who worked at DOGE for nearly two months. “I’m sure mistakes were made. Mistakes are always made. I would never recommend someone run my code and do what it says. It’s like that ‘Office’ episode where Steve Carell drives into the lake because Google Maps says drive into the lake. Do not drive into the lake.”

Though Lavingia has talked about his time at DOGE previously, this is the first time his work has been examined in detail and the first time he’s publicly explained his process, down to specific lines of code.

Lavingia has nearly 15 years of experience as a software engineer and entrepreneur but no formal training in AI. He briefly worked at Pinterest before starting Gumroad, a small e-commerce company that nearly collapsed in 2015. “I laid off 75% of my company — including many of my best friends. It really sucked,” he said. Lavingia kept the company afloat by “replacing every manual process with an automated one,” according to a post on his personal blog.

Lavingia did not have much time to immerse himself in how the VA handles veterans’ care between starting on March 17 and writing the tool on the following day. Yet his experience with his own company aligned with the direction of the Trump administration, which has embraced the use of AI across government to streamline operations and save money.

Lavingia said the quick timeline of Trump’s February executive order, which gave agencies 30 days to complete a review of contracts and grants, was too short to do the job manually. “That’s not possible — you have 90,000 contracts,” he said. “Unless you write some code. But even then it’s not really possible.”

Under a time crunch, Lavingia said he finished the first version of his contract-munching tool on his second day on the job — using AI to help write the code for him. He told ProPublica he then spent his first week downloading VA contracts to his laptop and analyzing them.

VA press secretary Pete Kasperowicz lauded DOGE’s work on vetting contracts in a statement to ProPublica. “As far as we know, this sort of review has never been done before, but we are happy to set this commonsense precedent,” he said.

The VA is reviewing all of its 76,000 contracts to ensure each of them benefits veterans and is a good use of taxpayer money, he said. Decisions to cancel or reduce the size of contracts are made after multiple reviews by VA employees, including agency contracting experts and senior staff, he wrote.

Kasperowicz said that the VA will not cancel contracts for work that provides services to veterans or that the agency cannot do itself without a contingency plan in place. He added that contracts that are “wasteful, duplicative or involve services VA has the ability to perform itself” will typically be terminated.

Trump officials have said they are working toward a “goal” of cutting around 80,000 people from the VA’s workforce of nearly 500,000. Most employees work in one of the VA’s 170 hospitals and nearly 1,200 clinics.

The VA has said it would avoid cutting contracts that directly impact care out of fear that it would cause harm to veterans. ProPublica recently reported that relatively small cuts at the agency have already been jeopardizing veterans’ care.

The VA has not explained how it plans to simultaneously move services in-house, as Lavingia’s code suggested was the plan, while also slashing staff.

Many inside the VA told ProPublica the process for reviewing contracts was so opaque they couldn’t even see who made the ultimate decisions to kill specific contracts. Once the “munching” script had selected a list of contracts, Lavingia said he would pass it off to others who would decide what to cancel and what to keep. No contracts, he said, were terminated “without human review.”

“I just delivered the [list of contracts] to the VA employees,” he said. “I basically put munchable at the top and then the others below.”

VA staffers told ProPublica that when DOGE identified contracts to be canceled early this year — before Lavingia was brought on — employees sometimes were given little time to justify retaining the service. One recalled being given just a few hours. The staffers asked not to be named because they feared losing their jobs for talking to reporters.

According to one internal email that predated Lavingia’s AI analysis, staff members had to respond in 255 characters or fewer — just shy of the 280 character limit on Musk’s X social media platform.

Once he started on DOGE’s contract analysis, Lavingia said he was confronted with technological limitations. At least some of the errors produced by his code can be traced to using older versions of OpenAI models available through the VA — models not capable of solving complex tasks, according to the experts consulted by ProPublica.

Moreover, the tool’s underlying instructions were deeply flawed. Records show Lavingia programmed the AI system to make intricate judgments based on the first few pages of each contract — about the first 2,500 words — which contain only sparse summary information.

“AI is absolutely the wrong tool for this,” said Waldo Jaquith, a former Obama appointee who oversaw IT contracting at the Treasury Department. “AI gives convincing looking answers that are frequently wrong. There needs to be humans whose job it is to do this work.”

Lavingia’s prompts did not include context about how the VA operates, what contracts are essential or which ones are required by federal law. This led AI to determine a core piece of the agency’s own contract procurement system was “munchable.”

At the core of Lavingia’s prompt is the direction to spare contracts involved in “direct patient care.”

Such an approach, experts said, doesn’t grapple with the reality that the work done by doctors and nurses to care for veterans in hospitals is only possible with significant support around them.

Lavingia’s system also used AI to extract details like the contract number and “total contract value.” This led to avoidable errors, where AI returned the wrong dollar value when multiple were found in a contract. Experts said the correct information was readily available from public databases.

Lavingia acknowledged that errors resulted from this approach but said those errors were later corrected by VA staff.

In late March, Lavingia published a version of the “munchable” script on his GitHub account to invite others to use and improve it, he told ProPublica. “It would have been cool if the entire federal government used this script and anyone in the public could see that this is how the VA is thinking about cutting contracts.”

According to a post on his blog, this was done with the approval of Musk before he left DOGE. “When he asked the room about improving DOGE’s public perception, I asked if I could open-source the code I’d been writing,” Lavingia said. “He said yes — it aligned with DOGE’s goal of maximum transparency.”

That openness may have eventually led to Lavingia’s dismissal. Lavingia confirmed he was terminated from DOGE after giving an interview to Fast Company magazine about his work with the department. A VA spokesperson declined to comment on Lavingia’s dismissal.

VA officials have declined to say whether they will continue to use the “munchable” tool moving forward. But the administration may deploy AI to help the agency replace employees. Documents previously obtained by ProPublica show DOGE officials proposed in March consolidating the benefits claims department by relying more on AI.

And the government’s contractors are paying attention. After Lavingia posted his code, he said he heard from people trying to understand how to keep the money flowing.

“I got a couple DMs from VA contractors who had questions when they saw this code,” he said. “They were trying to make sure that their contracts don’t get cut. Or learn why they got cut.

“At the end of the day, humans are the ones terminating the contracts, but it is helpful for them to see how DOGE or Trump or the agency heads are thinking about what contracts they are going to munch. Transparency is a good thing.”

“There are real-life dangerous impacts for veterans,” said Rep. Chris Deluzio of Pennsylvania, citing the news organization’s work.

This week, ProPublica reported on dozens of emails sent from staff at VA hospitals and clinics across the country to headquarters warning how cuts could, and in some cases are, degrading the agency’s ability to provide for the roughly 9 million veterans who rely on it.

Hiring freezes and other edicts from the White House have left medical providers scrambling and short-staffed amid an ever-shifting series of policy moves, including the cancellation of contracts with companies that maintain cancer registries, the emails said. Staffers at VA centers in Pennsylvania warned the cuts were causing “severe and immediate impacts,” including to “life-saving cancer trials.”

“Enrollment in clinical trials is stopping,” one wrote, “meaning veterans lose access to therapies.” Staffers at the hospital warned more than 1,000 veterans would lose access to treatment for diseases ranging from metastatic head and neck cancers, to kidney disease, to traumatic brain injuries.

On Thursday, the House members, several of whom are veterans, demanded VA leadership provide more details on how cuts are affecting such work, in which service members often receive treatment they would not otherwise have access to.

“We all want to cut waste, fraud and abuse, but what we see today is when you cancel a contract, it means the end of a clinical trial that’s going to save someone’s life,” Rep. Maggie Goodlander of New Hampshire said.

Notably, Deluzio, an Iraq War veteran whose Pittsburgh-area district includes a VA facility, and other lawmakers said they had learned about the impact for the first time from ProPublica’s reporting. On Thursday, they accused agency Secretary Doug Collins of stonewalling their efforts to find out what positions have been laid off, what contracts have been canceled and what future cuts will look like.

“We want the country to understand that this administration is hiding what they are doing, not just from us and the Congress, but from veterans and the American people,” Deluzio said.

“And the worst part is, we don’t know if anyone has died,” he added.

President Donald Trump has long said his administration will prioritize veterans and not compromise their care.

The disruptions at the VA have come even as the department has laid off just a few thousand staffers — a small fraction of the employees it said it ultimately plans to remove. Collins has said the agency is developing plans with Elon Musk’s Department of Government Efficiency to cut at least 70,000 employees — a number that he has underscored is a “goal.” “Could be more, could be less,” he told lawmakers this week.

On Thursday, in a post on X, Collins pushed back on criticism, calling ProPublica’s reporting “misleading” and saying it was based on “some outdated reports from the internal system VA uses to quickly identify and fix issues across the department.”

In a statement, VA press secretary Pete Kasperowicz said that Collins was working to fix a “broken bureaucracy” that has long had problems with patient safety and access to care, among other issues. “Unfortunately, many in the media, government union bosses and some in Congress are fighting to keep in place the broken status quo,” he said. “Our message to Veterans is simple: Despite major opposition from those who don’t want to change a thing at VA, we will reform the department to make it work better for Veterans, families, caregivers and survivors.”

Kasperowicz previously told the news organization that the issues in Pennsylvania have been resolved, though locals there with knowledge of the issues said that’s not the case and that the impact is ongoing. Kasperowicz also said in regard to the contracts to maintain the cancer registries that there had been “no effect on patients.” He added that the VA is moving to create a national contract to administer them.

According to some providers, even the temporary disruptions have hurt the care of veterans. One clinical trial to treat veterans for opioid addiction was hobbled by temporary layoffs. “We couldn’t give veterans a tool that could save their lives,” said Ellie Gordon, the CEO of the startup Behavior, which is testing biosensors to alert veterans to the risk of relapse.

Collins touted the cuts in a sometimes-contentious hearing on Tuesday before the U.S Senate Committee on Veterans’ Affairs.

“We’re going to maintain VA’s mission-essential jobs like doctors, nurses and claims processors, while phasing out non-mission essential roles like interior designers and DEI officers,” he said in an opening statement. The funds saved will be rerouted into direct health care and benefits for veterans, he added.

Some Republicans at the hearing defended the administration’s proposed cuts. “The VA has become a bloated bureaucracy,” said Sen. Tommy Tuberville, who represents Alabama. “I think most of us will agree with that.”

But Sen. Richard Blumenthal, D-Conn., pushed back on Collins’ statements, saying that laying off such a large portion of the staff will inevitably involve letting go of health care workers, like nurses and doctors. “You cannot slash and trash the VA without eliminating those essential positions which provide access and availability of health care,” he said. “It simply cannot be done.”

Others at the hearing took Collins to task for a lack of transparency. Sen. Angus King, I-Maine, admonished the secretary for refusing to provide a list of the 538 canceled contracts since his appointment. Collins said he would provide the information, but only after it’s finalized.

“We’re looking at every step we can, but also, I’m not going to play it out in a public arena,” he said.

The email said more than 1,000 veterans would lose access to treatment for diseases ranging from metastatic head and neck cancers, to kidney disease, to traumatic brain injuries.

“Enrollment in clinical trials is stopping,” the email warned, “meaning veterans lose access to therapies.”

The administration reversed some of its decisions, allowing some trials to continue for now. Still, other research, including the trials for treating head and neck cancer, has been stalled.

President Donald Trump has long promised to prioritize veterans.

“We love our veterans,” he said in February. “We are going to take good care of them.”

After the Department of Veterans Affairs began shedding employees and contracts, Trump’s pick to run the agency, Secretary Doug Collins, pledged, “Veterans are going to notice a change for the better.”

But dozens of internal emails obtained by ProPublica reveal a far different reality. Doctors and others at VA hospitals and clinics across the country have been sending often desperate messages to headquarters detailing how cuts will harm veterans’ care. The VA provides health care to roughly 9 million veterans.

In March, VA officials across the country warned that a critical resource — databases for tracking cancer — would no longer be kept up to date. As officials in the Pacific Northwest explained, the Department of Government Efficiency was moving to kill its contract with the outside company that maintained and ran its cancer registry, where information on the treatment of patients is collected and analyzed. DOGE had marked it for “immediate termination.”

The VA in Detroit raised a similar alarm in an email, warning of the “inability to track oncology treatment and recurrences.” The emails obtained by ProPublica detail a wide variety of disruptions. In Colorado, for instance, layoffs to social workers were causing homeless veterans waiting for temporary housing to go without help.

The warnings, sent as part of a longstanding system at the VA to alert higher-ups of problems, paint a portrait of chaotic retrenchment at an agency that just three years ago was mandated by Congress through the PACT Act to expand care and benefits for veterans facing cancer and other issues after exposure to Agent Orange, burn pits or other toxins.

Doctors and other health care providers across the VA have been left scrambling and short-staffed amid an ever-shifting series of cuts, hiring freezes and other edicts from the White House.

The upheaval laid bare in the emails is particularly striking because the cuts so far would be dwarfed by the dramatic downsizing in staff and shift in priorities the administration has said is coming.

The VA has cut just a few thousand staffers this year. But the administration has said it plans to eliminate at least 70,000 through layoffs and voluntary buyouts within the coming months. The agency, which is the largest integrated health care system in the U.S., currentlyhas nearly 500,000 employees, most of whom work in one of the VA’s 170 hospitals and nearly 1,200 clinics.

Despite an expanded role mandated by Congress through the PACT Act, administration officials have said their goal is to trim the agency to the size it was before the legislation passed.

“The Biden Administration understood what it meant to pay for the cost of war; it seems the Trump Administration does not,” said Rep. Mark Takano, a California Democrat and chief author of the PACT Act.

Documents obtained by ProPublica show DOGE officials working at the VA in March prepared an outline to “transform” the agency that focused on ways to consolidate operations and introduce artificial intelligence tools to handle benefits claims. One DOGE document proposed closing 17 hospitals — and perhaps a dozen more.

VA press secretary Pete Kasperowicz told ProPublica that there would be no hospital closures. “Just because a VA employee wrote something down, doesn’t make it VA policy,” he said in a written statement. But he did say that use of AI will be a big part of what he called VA’s “reform” efforts.

Kasperowicz dismissed the idea that the emails obtained by ProPublica show chaos.

“The only thing these reports show is that VA has a robust and well-functioning system to flag potential issues and quickly fix them so we can provide the best possible care to Veterans,” he wrote.

DOGE did not respond to requests for comment.

The White House released a budget proposal last week that calls for a 4% increase in the VA’s budget. That total includes more money for medical care, though a portion of that would be used to pay for veterans to seek care outside the VA medical system.

More answers to the VA’s larger plans may come today, when Collins is scheduled to testify before the Senate Veterans Committee, his first hearing on Capitol Hill since coming into office.

David Shulkin, who headed the VA in Trump’s first term, said the administration is too focused on cuts rather than communicating a strategy for improving care for vets.

“I think it’s very, very hard to be successful with the approach that they’re taking,” Shulkin told ProPublica.

One way local VA officials have tried to limit the damage has been by sending warnings — formally known as an issue brief — to higher-ups. And sometimes it works.

After officials in Los Angeles warned that “all chemotherapy” would stop unless Washington backed off killing a service contract, the VA reversed its decision.

And, amid growing scrutiny, the administration also made some researchers in Pennsylvania and elsewhere exempt from cuts. The laid-off social workers who helped homeless vets in Colorado were also brought back after about a month away from their jobs. Kasperowicz said that four social workers were affected but “their caseload was temporarily redistributed to other members of the homeless team.”

The warnings from officials across the country underscore how the comparatively modest cuts so far are already affecting the work of the VA’s medical system, with the study and treatment of cancer cited in multiple warnings to agency leadership.

“We have absolutely felt the impact of the chaos all around us. We’re already losing people,“ said one senior researcher, who spoke to ProPublica anonymously for fear of retaliation.

Referring to studies, he added: “We’re going to be losing things that can’t restart.”

And while Kasperowicz told ProPublica that the issues in Pennsylvania have been resolved, locals there said that’s not the case and that the impact is ongoing.

In Pittsburgh, two trials to treat veterans with advanced head and neck cancer, which officials in March had warned were at risk because of hiring freezes, have still not started, according to Alanna Caffas, who heads a Pittsburgh nonprofit, the Veterans Health Foundation, that partners with the VA on research.

“It’s insane,” Caffas said. “These veterans should be able to get access to research treatments, but they can’t.”

A third trial there, to help veterans with opioid addiction, wasn’t halted. Instead, it was hobbled by layoffs of key team members, according to Caffas and another person involved in the research.

Regarding the issues with cancer registries, Kasperowicz said there had been “no effect on patients.” He added that the VA is moving to create a national contract to administer those registries.

Rosie Torres, founder of Burn Pits 360, the veterans advocacy group that also pushed hard for the legislation, called the emails showing impeded cancer treatment a “crisis in the making” and “gutwrenching.”

That the decisions are being made without input from the communities of vets they affect is worse, she added.

“If they are killing contracts that may affect the delivery of care, then we have a right to know,” she said.

Last week, as the second Trump administration marked its first 100 days in office, Collins celebrated what he described as its achievements.

In a recorded address, he said that under his stewardship the VA processed record numbers of benefit claims, ended “divisive” spending on diversity initiatives and redirected millions of agency dollars from “non-mission-critical” programs back toward services to benefit veterans.

“We will not stop working to put veterans first,” he wrote in an accompanying op-ed.

Others say Collins has done no such thing. Instead of focusing on veterans, said one VA oncologist, “we’re spending an enormous amount of time preparing for a staffing catastrophe.”

“Veterans’ lives are on the line,” the doctor said. “Let us go back to work and take care of them.”

The closure of the Veterans Benefits Administration’s Office of Equity Assurance effectively hobbles internal efforts at the VA to investigate and eliminate long-standing racial inequities the department itself has acknowledged.

The office was eliminated as part of the Trump administration’s purge of programs broadly aimed at addressing diversity, equity or inclusion, according to emails obtained by ProPublica. But several VA sources said that the office was not exclusively focused on race, and that it takes on cases for a range of veterans to ensure no one is denied proper benefits — including for reasons of age, religion, gender identity, sexual orientation and geographic location.

Rep. Mark Takano, a California Democrat, criticized the Trump administration’s action as “excessive” and “reckless.”

“The closure of the OEA will undoubtedly have disastrous effects on the care we offer veterans,” Takano, former chair and now ranking member of the House Committee on Veterans’ Affairs, said in a statement to ProPublica. “This office was making it easier for minority veterans to access care and benefits. Its closure will directly impact the care and benefits received by minority veterans.”

Richard Brookshire, co-founder of the Black Veterans Project, a nonprofit focused on rectifying discrimination faced by Black veterans, echoed Takano’s concerns.

“It’s a first step toward gutting the second-largest agency in our federal government,” he said. “The consequences will be dire, wide-reaching and deadly.”

VA spokesperson Peter Kasperowicz declined to say whether the agency would continue to study racial disparities. But he emphasized in a statement that newly installed VA Secretary Douglas Collins “treats all veterans and beneficiaries fairly and equally, so the Office of Equity Assurance is no longer needed.”

He added: “The money saved by closing the office will be redirected to improve health care, benefits and services for Veterans, all of whom we treat fairly and equally. VA will always fulfill its duty to provide veterans, families, caregivers and survivors the health care and benefits they have earned. That is a promise.”

The VA grew significantly under the Biden administration, with tens of thousands of employees added to beef up capacity in conjunction with the passage of the PACT Act. The 2022 law expanded health care and benefits for an estimated 3.5 million veterans exposed to toxic substances from burn pits and other chemicals.

The Biden administration also created the OEA and several other initiatives to help analyze and rectify discrimination in the delivery of health care, benefits and other services. Those moves were seen as a direct response to long-standing complaints by minority veterans.

The closure of the OEA is just one of several disruptive staff cuts at the VA in recent weeks. Around 2,400 VA employees have lost their positions since the Trump administration began slashing the federal workforce, with significantly more firings to come.

The department currently employs around 470,000 workers. Kasperowicz said that the administration plans would shed nearly 15% of its workforce, dropping the total to roughly 398,000.

Workers assigned to the OEA were informed on Feb. 14 via email that their positions were terminated immediately and that the office was being “liquidated.” The notices were sent to nearly all of the office’s employees, effectively dissolving the unit, sources familiar with the firings told ProPublica.

The administration reversed at least some of the terminations later that month, according to correspondence obtained by ProPublica. Workers who were attached to the OEA have now been placed on administrative leave pending a possible reassignment within the VA or another federal department, according to sources familiar with the department who spoke on the condition of anonymity for fear of retaliation.

“But even if they are reassigned, it won’t be to the OEA,” said one official familiar with the moves. “It’s definitely gone.”

Black veterans and their advocates have long complained of a divide in how claims are handled. An ongoing suit filed by the National Veterans Council for Legal Redress against the government gave the suspicions new life.

The group’s case was bolstered by data unearthed through Freedom of Information Act requests from the Black Veterans Project, which found that the VA was far more likely to reject applications for service-related disabilities by Black veterans than their white counterparts.

Attorneys for the federal government asked the judge overseeing the suit to dismiss the claims, stating that only the VA secretary has jurisdiction to decide disputes over award benefits and that the court lacks the jurisdiction to hear them.

A 2023 U.S. Government Accountability Office report also concluded that there were disparities. It found that the department approved compensation applications for service-related disabilities like hearing loss, impaired limb movement and post-traumatic stress by Black veterans at lower rates than veterans of other races.

The report found that between 2010 and 2020 the approval rate for benefit applications made by White male veterans was 3% to 22% higher than Black male veterans for the selected medical conditions.

The OEA’s demise is just one part of an ongoing rollback of the racial equity programs the Trump administration has called “radical and wasteful.”

Since Trump entered office, a webpage detailing the VA’s work to address equity and diversity appears to have been scrubbed from its website. In January, the administration fired the heads of internal advisory groups formed to address the concerns and needs of minority and female veterans.

One of those groups, the Center for Minority Veterans, worked in conjunction with the OEA to address racial disparities in disability compensation. Between 2013 and 2018, the advisory group raised its concerns over the Black veterans’ lower rate of claims approval to VA leadership five times, according to the National Veterans Council for Legal Redress suit.

Mariela Roca, a former Republican congressional hopeful, took over as director of the advisory group last week. It’s unclear what specific strategies the group will pursue to advance the needs and concerns of minority veterans under new leadership. Roca did not respond to a request for comment.