In April, National Institutes of Health (NIH) Director Jay Bhattacharya described a proposal for a new effort to "develop national disease registries, including a new one for autism." Days later, Department of Health and Human Services (HHS) officials walked back claims that the intention was to develop an autism registry, and instead focused on plans to create a real-world data (RWD) platform that would help researchers bypass fragmented and difficult-to-obtain existing data. This was followed in May by an announcement that NIH would be partnering with the Centers for Medicare & Medicaid Services (CMS) to create a data platform using information from Medicare and Medicaid, and a statement released by Secretary Kennedy. confirmed this partnership aims to "uncover the root causes of autism and other chronic diseases". Despite descriptions of a rapid timeline, there are no additional details of study protocols or timelines to inform assessment of how this work will be completed.
What is a real-world data platform and why are patient registries important?
Real-world data (RWD) platforms collect, store, integrate, and analyze health-related data from sources beyond traditional clinical trials, including electronic health records, claims and billing records, product and disease registries, and patient-generated data from apps and wearables. Such platforms can be transformative for research by offering comprehensive datasets that reflect patient behaviors, experiences, and outcomes. By harnessing these data, researchers can gain insights into health care access, utilization, safety, and effectiveness patterns across diverse populations, environments, and clinical settings. This facilitates understanding of complex risk and disease dynamics, opening the door for precision medicine to tailor interventions to individual genetic, environmental, and lifestyle factors. This approach has the potential to improve treatment outcomes, reduce adverse effects of care, and make health care more cost-effective. There has been growing momentum toward the development of RWD platforms for more than two decades, with the landmark launch of the FDA’s Sentinel Initiative in 2008 followed by several other national research initiatives with funding from NIH, PCORI, and others.
As contributors to RWD, patient registries have long been valued for research because they offer a comprehensive and longitudinal view of populations sharing key characteristics or conditions, allowing researchers to study disease patterns, identify risk factors, and monitor treatment effectiveness. They provide a way to gather data from larger, more diverse populations than clinical trials, which is crucial for understanding rare diseases or specific patient subgroups. Powerful examples can be found in state and national cancer registries and the diverse range of rare disease registries participating in the Rare Disease Clinical Research Network (RDCRN), many of which are housed in patient advocacy organizations.
Why does the autism initiative give us pause?
Real-world data platforms can be powerful engines of discovery. Still, they are not without potential risks, including diminished family and community trust, threats to data privacy, and perpetuating narratives of eugenics. Taken together, collecting information about historically minoritized groups without strict safeguards in place can jeopardize individual, family, and community safety.
Government and health care mistrust is often more commonly felt by populations that have experienced historical and present-day harm, including people with disabilities. Disability protections are relatively new, and for many families of children with special health care needs, there are clear and personal memories of autistic children being institutionalized, segregated from their peers, and mistreated. There are real and recent cases where autistic people have been inappropriately discriminated against in health care, employment, education, housing, and public access to resources. This has all happened in a time when disclosure of a diagnosis is left to the discretion of the family, so the idea of involuntary disclosure and the maintenance of a list of those with an autism diagnosis dredge up valid fears.
Early signals from the administration have raised significant concerns that need to be fully addressed with urgency. In April, when announcing the autism data initiative, Secretary Kennedy pledged to have answers about the cause of autism by September, a promise that is unrealistic due to the complexity of the neurodevelopmental condition and the timeframe typically needed for research endeavors of this nature. Couple this with his recent move to reconstitute the Advisory Committee for Immunization Practices (ACIP) with members of questionable credentials and his repeated claims about a link between vaccines and autism—a conclusion soundly rejected by the scientific communitya conclusion soundly rejected by the scientific community—and it surfaces concerns that the research initiative could be intended only to look for the proof needed to confirm an a priori assumption, a practice that would be anathema to the scientific method. Additionally, Secretary Kennedy’s use of language to mischaracterize and pathologize the condition, dehumanize members of the autism community, and treat the autism experience as monolithic is alarming. He has referred to prevalence of the condition as an “epidemic”, disregarding evidence that the increases are due to improved identification as opposed to a true rise in rates, he has committed to research for a “cure”, he has claimed that autism is an “individual tragedy””” that “destroys families,” and he was described those on the spectrum as “kids who will never pay taxes, they’ll never hold a job, they’ll never play baseball, they’ll never write a poem, they’ll never go out on a date. Many of them will never use a toilet unassisted.” Inaccuracies and misleading claims aside, this language positions autistic traits and behaviors as inherently negative, disordered, and pathological, rather than as variations within the spectrum of human experience. People with autistic traits are different, not deficient.
History has shown us that the methods and legitimacy of science can be appropriated by ideology and used to inflict harm—as in the case of eugenics, the pseudoscientific theory that the human race can be “perfected” (i.e., White, cis-gendered, heterosexual, and non-disabled) through strategic genetics and heredity. Both historical and present-day eugenics rely on erroneous science to legitimize bias and discrimination. Eugenics was behind Buck v. Bell, the Supreme Court’s 1927 decision to uphold involuntary sterilization of people the state deemed “feebleminded,” resulting in the forced sterilization of at least 60,000 people through 30 states’ laws by the 1970s. It is similarly behind the health care field’s use of race-based medicine, and disregard for specific populations in the initial response to the COVID-19 pandemic. It is critical to center the legacy of eugenics when developing and using real-world data registries to avoid harmful policies and practices rooted in bias, discrimination, and the devaluation of certain lives.
When eugenics is an underlying principle in the medical model of disability that focuses on “fixing” a disability or condition, society pathologizes differences that create disparities in outcomes rather than improving systems of care.
What does the autism community need to see?
Conducting scientific research as outlined by the NIH can be done, but it requires a careful and ethical approach that prioritizes shared governance, population and data protections, and rigorous scientific methodology. To date, the administration has not demonstrated a clear commitment to these priorities for the RWD platform focused on autism, and, as described above, there is ample evidence that the intentions for this research initiative may not be in good faith. For this work to deliver on its promises, at minimum, the following must be addressed fully and transparently:
- How will HHS and NIH build trust with the autism community through shared governance? A RWD platform for research on autism must be designed and governed in partnership with families using trust-based structures and processes that account for the priorities and concerns of the community. This includes addressing past harms and concerns around surveillance or bias and ensuring that families and autistic individuals have a voice in decisions about what data are collected, how they are interpreted, and how findings are ultimately used. Direct partnership with families could calm concerns by ensuring the RWD platform’s purpose is supportive and destigmatizing, proper controls are in place to prevent use of the platform as a means of surveillance, robust ethical oversight is enforced and the platform is compliant with the HIPAA Privacy Rule, and legal protections are implemented to prevent discrimination or adverse consequences—such as employer access to information or use by insurers to deny coverage. By emphasizing these elements through shared governance, agencies and researchers can build trust and empower families, advancing scientific knowledge while upholding ethical standards and promoting social justice.
- How will data linkages be performed such that privacy, security, and confidentiality are assured? The NIH has emphasized the use of de-identified data, and there are ways to do this. However, the NIA LINKAGE program, which is the primary way NIH links to CMS data now, uses probabilistic matching on personal identifiers for studies that have participant consent (or waivers). If NIH plans to deviate from this publicly available protocol and use de-identified data, what methodology will be used, and how will risk of re-identification be mitigated, particularly when data come from sources such as commercial databases and wearable devices? If NIH will be linking on personal identifiers, how will these processes engage families, obtain consent, and protect privacy? Failure to do any of these diminishes a person’s dignity and autonomy over personal information and how it is used.
- How will study design be assessed and bias mitigated? Real-world data can be powerful, but there are significant threats to internal and external validity. By their very nature, these data are not collected for the purpose of research, and they are susceptible to confounding variables, selection bias, and information bias. Indeed the challenges in using big data for autism research, inclusive of Medicaid and Medicare claims specifically, have been characterized. Minimum requirements for study designs, experience and outcomes measures, and statistical methodologies—as well as the specific process for peer review—should be transparently described, particularly given recent actions by the administration to override peer review decisions and terminate grants based on opaque determinations of study foci as “unscientific”.
History is watching.
We have made progress through increased awareness and diagnoses that have led to the positive effects of treatment and support for people with autism. Yet, despite this progress, the idea of “labeling” a child with autism deters many families from seeking services, especially those who have experienced harm or mistrust. Moreover, even with diagnoses, families face barriers to accessing services. Adding another layer of deterrent may mean that another generation of children goes undiagnosed and unsupported. Without building deep trust and inclusion in autism research and co-creating solutions to adequate service provision and treatment, the potential ramifications will impact the efforts of both medical and education professionals to support families accessing services for diagnosis and treatment. Instead of creating wider gaps between health care access and the community, we have the opportunity to leverage research and data to improve health for all, build trust with families and individuals most impacted, and do so in a way that honors the humanity of disabled people and prioritizes their quality of life. There are ways to do this work well. We have strong examples from PCORI, RDCRN, and other initiatives. A primary ingredient to the success of these programs has been the deep engagement of patients, families, and communities affected by the conditions under study. We must embody "nothing about us without us."
When families are part of the design and oversight process, and given legitimate power to contribute to decision-making, when trust is built and humanity is respected, amazing things can be done. This country has a track record, not only of exclusion, but of the weaponization of science against those powerless to stop it. We are not operating on a baseline of trust where the logical response is to be surprised that people would question the intentions of researchers. On the contrary, the logical response is to honor our collective history, do the work, and show how the scientific community will get this right.
Key Takeaways
- The NIH has proposed developing a real-world data (RWD) platform to uncover the root causes of autism.
- RWD platforms enable potentially transformative research by offering comprehensive insights into patient risks, experiences, and outcomes—opening the door for precision medicine.
- The lack of clear details outlining plans for governance, data collection and management, and engagement with the autism community raise significant concerns.
- Mistrust in health care is particularly intense among historically marginalized groups, including people with autism, due to prior mistreatment by the scientific community.
- For the initiative to be successful, the following must be addressed: (1) shared governance and stewardship of the platform that includes members of the autism community; (2) ethical protections for patient consent, data privacy, and confidentiality; and (3) application of rigorous study designs, methods, and measures for studying root causes of autism.
- Successful models of RWD platforms exist, and all of them include robust strategies for these components. The platform for autism should be no different. In the absence of this, progress in autism awareness, support, and service-seeking will be undermined.