RFK Jr.'s autism registry is looking for collaborators

$50 million in grants are being offered to build a "real-world data platform"

(Photo by David Pupăză on Unsplash)

When Robert F. Kennedy, Jr. became Secretary of Health of Human Services (HHS), he pledged to identify the “cause” of autism by September 2025. His plan for doing so raised eyebrows not just for its aggressive timeline but also for the preferred approach, a national registry of people diagnosed with autism spectrum disorders.

The announcement prompted important questions like: What exactly is the American government going to do with this list? Answers were scarce, but just days later NIH director Jay Bhattacharya announced that the institutes had already begun gathered medical records to build the registry.

While Kennedy and Bhattacharya were excited to get started, many advocates and scientists were outraged. One parent of autistic children referred to the proposal as “a slippery slope to eugenics.” In this vein, the proposed registry has already drawn comparisons to organized murders of disabled people under the Nazi regime which required medical practitioners to report people with “incurable” conditions to the government.

In the face of this pushback and just four days after Bhattacharya’s announcement, an HHS official contacted the press to bluntly say, “We are not creating an autism registry.” Instead, the government would work on a “real-world data platform.” Yet, the difference between a “registry” and a “real-world data platform” remained unstated.

By late May, the NIH released a funding announcement for the new Autism Data Science Initiative (ADSI). ADSI is dispensing a total of $50 million in grant money to “create new integrated data resources” and analyze them to “link data on genetic and nongenetic factors.”

The precise information that would be included in the database is up to the researchers who submit proposals. The NIH provided examples of information that could be stored in these databases: clinical diagnoses, behavior, neurophysiology, genetics, etc.

Proposals were due by June 27, and awardees will be notified by September 1. The turnaround here is tight by NIH standards, especially considering that the NIH in 2025 is way behind its normal pace of awarding grants. The speed at which the ADSI is moving is also reflected in its review process which is just one round of scientific review rather than the normal two.

Additionally, the NIH reserves the right to select proposals for funding, a rare provision in NIH funding opportunities. This gaping loophole leaves room for political interference in the selection process and ultimately undermines trust in the whole initiative. For example, Kennedy could direct the NIH to select his favorite proposals or disburse grants to his cronies.

In crafting the funding announcement, the NIH provides hints as to the types of proposals that they are looking for. Required skills (shown below) include building datasets “from disparate sources,” creating new datasets (including geo-coding data for some reason), and using AI to analyze these data.

(Photo: Source: NIH | Red markup added for emphasis.)

Honestly, it’s a stupid idea. And, it’s dangerous. The amount of influence NIH staff (including Bhattacharya as director) has over the process in the review phase and after acceptance is seriously concerning. Because of the funding mechanism used by the ADSI, NIH staff can use the generated datasets however they want. This includes commandeering the data without permission from the academic investigator who generated it.

To the Trump regime, such political interference is “gold standard science.” Yet, the approach has already undermined trust in the ADSI within the autism researcher community. Why should someone go through the (hopefully good-faith) effort to compile this information if the NIH could politicize the data and use it for whatever it chooses?

And then there’s the problem of bad-faith researchers who are eager for the NIH money and hustled to put together a half-assed proposal. Privacy, scientific integrity, and ethical engagement with the autistic community could all be sacrificed for the sake of grant money.

Considering all this, we should all be highly skeptical of the ADSI. The initiative is poorly planned, scientifically lackadaisical, and likely harmful to autistic people. Fortunately, there are things that you can do to protect yourself.

Don’t consent to ADSI studies. As a scientist, it pains me to have to write that, but it is clear that the ADSI is a political project and not a scientific project. Giving your consent to participate in an ADSI study is giving your consent to not just the research team but also NIH staff to mine your personal data.

The generation of new datasets will still be subjected to informed consent. If you are approached to take part in an autism study, you do not have to agree to participate. It is fair to ask the researcher whether the study is funded by the ADSI. If so, then think carefully about the risks of your involvement.

There is currently no indication that these ethical and political shortcomings will affect existing and/or new studies of autism that are funded by the NIH outside of the ADSI. It should be OK to participate in these non-ADSI studies, but the decision is ultimately yours.

Lobby your state governments to protect existing data on people diagnosed with autism. Initially, a major arm of the proposed registry was to assemble medical records of autistic Americans by compiling state-level data that the federal government already has access to.

To prevent this, the state of Illinois (via executive order from Governor JB Pritzker) has limited the collection and dissemination of health information related to autism. The order requires informed consent from the autistic individual for their data to be shared with federal authorities. More states should follow Illinois’ lead, and you can encourage them to do so by reaching out to your governor’s office.1 

This action is particularly important for residents of Delaware, Indiana, North Dakota, New Jersey, Rhode Island, Utah and West Virginia. These states each have mandatory autism registries where doctors are required to report patients with autism diagnoses. State governments can still restrict access to these information (as in Illinois), but they are likely going to need grassroots advocacy to make these changes.

(You can also lobby your elected representatives to dismantle these registries. New Hampshire did so legislatively just last year, in a rather prescient decision.)

Ultimately, the ADSI is a dangerous “real-world data platform.” Given the timeline presented by the funding announcement, its construction will rely on external collaborators (largely in academia) and will begin as soon as this fall. While Kennedy will fail to meet his goal of identifying the causes of autism by September 2025, he will have (by then) assembled a team of collaborators to build out the registry that he so craves.

There are serious risks to participating in these ADSI studies. But as a reminder, the choice to participate in the generation of new data for this database is yours, and you can take steps to protect your existing health data by contacting your state government. The ADSI can be ineffectual if enough people reject it as the political ploy it is.

---

(c) 2025, Queer Science Lab

https://queersciencelab.substack.com/p/rfk-jrs-autism-registry-is-looking