top of page
Search

Five Priorities the Federal Autism Strategic Plan Should Not Overlook

Caregiver education, adaptive support, equitable access, responsible AI, and rigorous evaluation must be part of the national autism strategy


The autism service system is confronting a scale problem.

In the Centers for Disease Control and Prevention's most recent Autism and Developmental Disabilities Monitoring Network report, approximately 1 in 31 8-year-old children across 16 participating U.S. communities were identified with autism in 2022. This estimate does not represent every child in the United States. The CDC states that the increase in identified prevalence underscores increased demand for equitable diagnostic, treatment, and support services.


At the same time, the 2021-2023 IACC Strategic Plan documented continuing disparities in access to assessment, diagnosis, interventions, and other services. It also identified barriers involving geography, income, language, broadband, transportation, and workforce capacity. Workforce expansion remains essential, but the system cannot solve every educational and support need by relying exclusively on clinician-delivered services.


That does not mean replacing clinicians with technology or directing families toward less support than they need. It means building a more responsive continuum in which professional care, caregiver education, peer support, community resources, and responsibly designed technology each have an appropriate role.

In response to the 2026-2028 Interagency Autism Coordinating Committee Strategic Plan Working Draft, I prepared public comments focused on five areas that deserve clear attention in the final plan. The IACC is a congressionally mandated federal advisory committee that coordinates autism-related activities across the U.S. Department of Health and Human Services and advises the HHS Secretary on autism-related issues.


1. Fund adaptive, parent-mediated, and stepped-support models

Autistic people and their families do not all need the same services, delivered at the same intensity, in the same format, at every point in their lives.

The 2021-2023 IACC Strategic Plan cited the Lancet Commission on the future of care and clinical research in autism and described a personalized stepped-care approach in which support changes as a person's strengths and needs evolve. The next federal strategy should carry that principle forward and fund the research required to implement it responsibly.

Stepped support should not mean automatically starting every family with the least expensive option. It should mean matching support to the individual's needs, preferences, circumstances, and response, then adjusting that support as those factors change.


Research comparing self-directed and therapist-assisted parent-mediated telehealth programs illustrates why this distinction matters. In a 2016 pilot randomized trial, both groups improved on several parent and child measures, while the therapist-assisted group showed greater gains in parent intervention fidelity and positive perceptions of the child. In a 2024 randomized study, therapist assistance predicted program engagement and produced a significant parent-learning effect in the full randomized sample; the self-directed group showed a learning effect when the analysis was limited to participants who completed the program. A separate 2021 proof-of-concept trial examined the feasibility of a stepped-care telehealth program, but it was not a definitive effectiveness trial. These findings support further study of who benefits from each level of assistance. They do not establish that self-directed education is sufficient for every family.


The appropriate research question is therefore not whether digital education can replace clinicians. It is how to determine:

  • Who benefits from self-directed caregiver education

  • Who benefits from peer or group support

  • Who requires clinician-assisted education or coaching

  • Who requires more intensive individualized services

  • When support should be increased, modified, or reduced

  • How family preference and autistic people's goals should influence those decisions


The Strategic Plan should support comparative-effectiveness studies, adaptive-intervention trials, and real-world implementation research that answer these questions.

2. Recognize caregiver education as a distinct and complementary support

Behavioral challenges and everyday support needs do not occur only during scheduled appointments.

Family caregivers may need practical education while waiting for services, between professional appointments, during transitions, and after a formal episode of care has ended. Well-designed caregiver education may help families understand behavior, prepare for difficult situations, communicate more effectively, and use strategies more consistently across daily environments. Those potential benefits should be tested rather than assumed.


But caregiver education must be defined accurately. It is not automatically therapy, diagnosis, treatment, crisis intervention, or clinical decision-making. It should not be used to shift professional responsibilities onto families or justify withholding medically necessary care.


Federal research should examine caregiver education as its own category within a broader support system. Meaningful measures could include:

  • Usability and sustained engagement

  • Caregiver self-efficacy and confidence

  • Generalization across daily routines and environments

  • Caregiver burden and satisfaction

  • Safety and appropriate escalation to professional care

  • Outcomes identified as valuable by autistic people and their families

  • Equitable participation across languages, regions, and socioeconomic groups


Research should also investigate why families disengage from self-directed programs. A resource cannot improve access if families cannot understand it, do not trust it, cannot use it consistently, or find that it does not reflect their circumstances.


3. Build linguistic, economic, disability, and geographic accessibility into the research

Accessibility cannot be added after a program has already been designed and evaluated for a narrow population.

Scalable caregiver supports should be developed and studied in plain language and multiple languages. Translation alone is not enough. Programs require cultural adaptation, accessible design, testing with intended users, and attention to differences in family structure, resources, values, and service systems.

Research must include rural communities, families with limited financial resources, households with limited broadband or digital literacy, people with disabilities, and individuals who communicate or learn in different ways. Autistic people and family caregivers should participate meaningfully in the design, governance, implementation, and evaluation of the programs intended to support them.

Economic accessibility also requires sustainable funding. Policymakers should examine reimbursement, grant, sponsor, and public-private funding models that make evidence-supported resources available without assuming that developers, providers, or community organizations can provide continued access without payment.

Equity requires more than creating a resource. It requires a realistic pathway for the people who need it to reach it, understand it, use it, and continue accessing it.


4. Establish responsible governance for AI-assisted caregiver tools

Public agencies are already establishing expectations for responsible AI use. The National Institute of Standards and Technology's AI Risk Management Framework addresses governance, transparency, privacy, fairness, and ongoing risk management. California's Department of Developmental Services has also issued guidance for ethical and responsible generative AI use within regional centers, emphasizing human judgment, privacy, transparency, accountability, and protection from bias. Autism-related caregiver tools should be held to similarly clear expectations for how they are designed, evaluated, described, and monitored.


For autism-related caregiver education, responsible implementation should include:

  • Clear disclosure that the user is interacting with an AI-assisted system

  • Plain-language explanations of what the system can and cannot do

  • Human control over decisions

  • Clear boundaries between education and clinical treatment

  • Data minimization and meaningful privacy and security safeguards

  • Bias, accessibility, and safety testing

  • Defined crisis limitations and directions to appropriate professional or emergency resources

  • Ongoing monitoring and a process for correcting unsafe or inaccurate content

  • Independent evaluation rather than reliance on marketing claims

AI-assisted education should never be represented as therapy, diagnosis, treatment, or a substitute for professional judgment. It also should not be treated as inherently effective because it is new, scalable, or technologically sophisticated.

Responsible innovation requires both opportunity and restraint.


5. Create pathways for independent evaluation and real-world pilots

Emerging caregiver-support technologies need a credible route from promising concept to responsibly evaluated resource.

That requires more than testimonials or internal usage statistics. Federal agencies and research partners should support carefully defined pilots with predetermined measures, transparent reporting, subgroup analysis, safety monitoring, and clear distinctions among feasibility, acceptability, educational benefit, clinical outcomes, service utilization, and cost.


A pilot may show that families can use a tool and find it acceptable. That does not automatically establish that the tool produces clinical improvement, reduces costs, or replaces another service. Those are separate questions requiring appropriate study designs.


The federal strategy should encourage collaboration among autistic people, caregivers, researchers, clinicians, public agencies, health plans, community organizations, and responsible technology developers. It should also establish standards that allow public partners to distinguish credible educational tools from products making unsupported clinical claims.


Why I am raising this issue

My perspective comes from nearly 30 years in behavioral health, including direct clinical practice and leadership within an autism service organization. It also comes from developing CareBridge Copilot, a paid subscription-based, AI-assisted educational platform for family caregivers navigating autism and other developmental disabilities.


According to current company product information, CareBridge provides structured guidance, scripts, printable plans, and checklists in English and Spanish. It is an educational caregiver-support tool, not therapy, supervision, diagnosis, or medical care, and it does not replace a BCBA, therapist, pediatrician, or school team. The company states that the platform does not store protected health information. These are product-design and policy statements, not evidence of clinical effectiveness.


CareBridge is one example within a much broader category of emerging caregiver-support technologies. It should not be endorsed, funded, or adopted based solely on how it is described by its founder. Like other commercial platforms, it should be evaluated against clear standards for usability, accessibility, safety, ethics, educational value, and real-world implementation.


That is precisely why the federal strategy matters.


The goal should not be to place an AI tool between every family and a clinician. The goal should be to build a system in which families and autistic people can access the right type of support at the right time, professional care remains available when it is needed, and new technologies are evaluated with the same seriousness as the problems they claim to address.


A national strategy must address both access and evidence

The United States needs more qualified clinicians, stronger service systems, better transition supports, and sustained investment across the lifespan. It also needs responsible ways to extend education and support beyond the limited hours available through formal services.


Those goals are not in conflict.


A strong federal autism strategy can protect access to necessary professional care while also advancing caregiver education, adaptive models, linguistic accessibility, responsible technology, and rigorous evaluation. Doing both is how we move from isolated programs toward a more responsive and equitable support system.

According to a public notice from IACC Chair Sylvia Fogel, comments intended for consideration at the next meeting should be submitted by August 20, 2026, at 5:00 p.m. Eastern Time. The notice also states that IACC accepts comments on a rolling basis. Comments may be emailed to IACCPublicInquiries@mail.nih.gov, the address published by the National Institute of Mental Health for IACC inquiries.


About the author

Yrenka Lolli-Sunderlin, M.S., BCBA, LBA, CLC, is the founder and CEO of Sunderlin Behavioral Interventions, Inc. and the creator of CareBridge Copilot. She has nearly 30 years of experience in behavioral health and has worked across clinical practice, organizational leadership, caregiver education, and responsible AI-assisted support.

Selected sources

 
 
 

Comments

bottom of page