Last week, the Senate Appropriations Committee released a report on the country’s federal labor, health, human services, and education agencies in which staffers buried a small note of concern.
It has to do with a case of duplication.
Under director Jay Bhattacharya, the National Institutes of Health (NIH) appears to be repurposing a project that Congress shot down last year, the note says. That project, which NIH abandoned amid pressure from Congress, would have amassed huge amounts of personal data from Americans with little oversight. NIH had planned, through contractors, to vacuum up information from phone providers, wearable device companies, health insurers, medical records keepers, and government agencies in order to create a “real-world data platform” for monitoring Alzheimer’s disease.
Though ambitious, the project was poorly managed, according to a 2024 report by the Government Accountability Office (GAO), a legislative branch watchdog agency. It was at risk of failing to manage appropriately hundreds of millions of dollars in spending, per the report, and it raised a host of privacy concerns around the protection of government-collected data.
In part because of the findings of that GAO report, Congress expressed “serious reservations” about the NIH effort, according to the Appropriations Committee document released last week. NIH would eventually shut down the project.
In last week’s little-noticed note, which was buried 150 pages into a document laying out the health agencies’ funding for the coming fiscal year, Senate Appropriations Committee staffers expressed a fresh concern: NIH is reviving the same program under a different name, despite Congress’ earlier objections. This time, it’s to study autism, the report said.
I set out over the past week to understand what this project is, who is working on it, the scope of the data that it may collect, and how it came to be that NIH revived a study of this sweeping scale — and one with documented, congressional concerns — and contorted it to conduct a mass study of autism.
What emerges is a bizarre picture, but one that’s characteristic of Trump II. The government is set to engage in a massive data collection effort despite past, documented privacy concerns about a similar effort. As part of it, the agency now plans to engage a contractor that the federal government has turned to time and again — one founded by a friendly billionaire with authoritarian impulses. It’s in service, this time, of an initiative that likely would not be taking place but for the conspiracy theorizing of a top political appointee.
At the same time, the note from Senate Appropriations — included in a report authored by subcommittee chair Rep. Shelley Moore Capito (R-WV) — means that two different Senate committees under control by each party have opposed versions of the project. It suggests that the concerns are bipartisan.
The revived, newly autism-focused project appears to be part of Health and Human Services Secretary Robert F. Kennedy Jr.’s highly controversial plan to seek a cure for autism and, in doing so, to create a database that some have worried will amount to an “autism registry.” HHS has pushed back on that characterization, saying it wants to create a “comprehensive” database to better understand the causes of autism. But that initiative has already sparked some debate within MAHA, and further antagonized an already-skeptical autism research community.
In a statement, the Autism Science Foundation, a nonprofit that has partnered with NIH in the past, criticized Kennedy’s plans.
“The lack of clarity around how data will be collected, shared, maintained and tested for accuracy raises red flags,” it said.
When reached for comment by TPM, an HHS spokesperson did not deny that the agency had revived the Alzheimer’s program under a new guise. NIH’s new autism initiative “includes the development of a secure data repository that will enable researchers to analyze large-scale, de-identified data to better understand the causes of conditions like autism and chronic diseases,” the agency said in a statement.
“All NIH-managed databases follow the highest standards of security and privacy, with the protection of personal health information as a top priority,” the statement reads.
Since the new administration took office, NIH has been ground zero for massive cuts. Research has been stopped across the board; funding to universities and research hospitals has been halted. Scores of researchers have been fired. It’s left the agency in a precarious position as it struggles to navigate wild new proposals from Kennedy.
Per an internal NIH slide show obtained by TPM, an updated version of the project is slated to use a platform developed by Palantir, the data behemoth co-founded by Peter Thiel, the influential billionaire. Palantir has benefited extensively from the Trump II alliance with big tech: as DOGE has sought to cut government services across the board, Palantir has stepped in as a private, big data operator, receiving in the process, by one tally, billions of dollars worth of federal contracts, including from agencies like NIH.
A $300 million project dropped
The original version of the project, which congressional staffers now say is being revived in a different form, was aimed at tracking Alzheimer’s in real time, across the country, and was underway by 2023.
The potential benefit there for medicine is obvious. The proposal would have allowed Alzheimer’s researchers to link vast new data sets related to people suffering from the degenerative disease. The National Institutes of Aging (NIA) was set to invest $50 million annually into the project. The project could, potentially, help identify young, healthy people early on as likely to develop the devastating illness. A CNN story in April 2023 heralded it as a means to “transform” research into the disorder.
“Real-world data is what we need to make a lot of decisions about the effectiveness of medications and looking really at a much broader population than most clinical trials can cover,” a top NIA official told Reuters at the time.
But within months, the project began to run into trouble.
The following spring, news reports emerged that NIH had shuttered the project without any announcement. The July 2024 report from the GAO accused it of being on track to mismanage more than $300 million, and of failing to manage data privacy risks. “Data privacy, protection, and consent are key considerations in any real-world data project,” the GAO report said, describing “lessons learned” by the NIA.
Dave Hinchman authored the report for GAO, and spoke to TPM about the 2024 paper. The project appeared to have been designed without any guardrails to keep the price tag for the $312 million project from ballooning, he said.
“That’s a big check that Congress is willing to write for that. And I think that given that size and scope that you can’t treat it like a small, little project,” he said.
The project also failed to prepare for risks like potential cyberattacks or other privacy problems, Hinchman said.
“We just didn’t see any evidence that they had really taken a look at how they were going to ensure the privacy of the participants in the studies,” he said.
A national ‘autism registry’ controversy
So, what brought the project back from the dead?
Enter RFK Jr.
Since his confirmation as HHS secretary, Kennedy has wrested federal medical research towards a longtime priority: autism.
Autism has played a central role in a story that the activist-turned-Cabinet secretary has told to audiences around the country for years. Autism is an epidemic, he says, and a “preventable disease” that leaves hundreds of thousands — maybe millions — of Americans mute, angry, and cut off from society, in his telling. It’s a description that autism advocates find deeply offensive, with a premise fundamentally opposed by current medical research: that autism is something for which there may be a cure. We all know who he has blamed as the culprit: first vaccines, then, more recently, such things as food additives and environmental toxins. “We know that it’s an environmental exposure,” Kennedy said in April.
At the time, Kennedy committed to a “massive testing and research effort” that would determine the cause of autism by September, in part through the “digitalization of health records.”
To Dr. Helen Tager-Flusberg, a prominent autism researcher and professor emerita at Boston University, the entire approach has deep flaws. Current research suggests that autism is rooted in “genetic factors” and early pre-natal development, she told TPM.
“It’s not like asthma, where the causes are genetic but also strongly environmental,” she said. “The brain differences are there probably at birth, and are certainly evident in the first few months of life.”
Weeks after Kennedy announced the autism initiative, NIH director Jay Bhattacharya said that the agency would develop a “real-world data platform” to study autism, using the same terminology that NIH used to describe the earlier Alzheimer’s effort. It would amass a huge amount of data, he suggested, collecting information from medical claims, wearable devices, and more to provide real-time coverage of where autism is appearing and what information correlates with it.
The proposal faced the same concerns that dogged the Alzheimer’s study: data privacy, project management, and use of congressional funding, among others.
Tager-Flusberg added that even if the data is anonymized, it’s very easy to determine people’s identities from simple information like date of birth and current residence. The Trump administration has already shared internal data with law enforcement agencies, she noted; that suggested another risk — that NIH-gathered data could be shared with people “not considered in the least bit trustworthy by the scientific community.” She’s already concerned that “someone might be able to unravel who the participants are in my study.” Participants in ongoing studies are “not coming in for research studies in the same way as they did prior to January 2025,” she noted.
Various news outlets picked up on the mass records collection component of Kennedy’s proposal and linked it to one mention in a presentation Bhattacharya gave that mentioned “national disease registries.” That, in turn, gave rise to discussion of a Kennedy-backed plot to create a national autism registry, prompting some concern that NIH was set to abuse its authority.
As that interpretation circulated, HHS denied that it was creating a national “autism registry.”
Three months later, Congress would raise a fresh concern: that Kennedy’s effort was attempting to repackage the already-shelved Alzheimer’s program.
Big data
In late July, NIH internally circulated a slideshow about the project, in part to address the same concerns that Congress expressed around reviving the project. TPM obtained the slideshow.
In late May, the slideshow notes, as part of Kennedy’s autism-researching initiative, NIH had announced a project called the “Autism Data Science Initiative (ADSI).” While it contains some of the same elements of the real-world data monitoring platform, the slideshow contrasts the two, describing the autism project as being a “research initiative” and separating it from the real-world data platform, describing it newly as a “data infrastructure initiative” used to study both autism and other conditions.
This may be a distinction without a difference. For one, the projects’ goals and methods appear to be largely the same. As Tager-Flusberg, the autism researcher, put it, ADSI seeks to amass data about Americans, thereby creating new data sets. It’s not clear that this will lead to any significant new research, she said.
“Would I put $50 million into the ADSI? Not in the way it’s been framed,” Tager-Flusberg said.
An HHS spokesperson defended the program in a statement as “employing only gold-standard, evidence-based science.”
The agency did not, when asked, distinguish between ADSI and the real-world data platform, saying only that the ADSI is “a collaborative effort designed to gain new knowledge about autism that could be used to further our understanding of the condition and improve the well-being and health outcomes for people on the autism spectrum.”
The document also says that NIH has already agreed with Centers for Medicare and Medicaid Services, currently helmed by Dr. Mehmet Oz, to share data about enrollees with autism.
Relatedly, the slideshow reveals that a platform developed by mass data contractor Palantir would be involved.
Palantir did not return TPM’s request for comment.
Teams assigned to the data platform project will “leverage” a platform called N3C as an “analytical center” to amass data in a central location, it said.
Palantir developed N3C for COVID-19 research starting in August 2020. The database, Palantir says, contains data from millions of patients, allowing researchers to track everything from common side-effects of the virus to long COVID. It’s also been used by NIH for other applications, including cancer research.
Per the slideshow, the idea is to use that same system for the real-world data platform.
Kennedy and Bhattacharya have set very fast timelines for the project. The slideshow projects that the platform will be up and running in four months. (In May, Kennedy had to walk back an earlier promise to find the cause of autism by “September.”)
Whether the project has overcome any of the initial concerns that led Congress to push for an end to its predecessor, the Alzheimer’s study, is unclear. But the level of alarm in the note from the Appropriations Committee, in what is normally a dry document, is telling. The project is “nearly identical” to the Alzheimer’s effort, it said, even though “Congress expressed serious reservations about” that project at that time.
Congress noted that it “strongly urges NIH to pause funding any new awards under the proposal” and to inform relevant congressional committees “on any effort related to” the real-world data platform every 30 days.
