Editor’s note: This essay examines the economic architecture of autism treatment in the United States. It is not an argument that all support is harmful, that parents are wrong for seeking help, or that every practitioner acts in bad faith. It is an argument that America has built a financial and legal system that funds autistic normalization far more readily than autistic accessibility, autonomy, or systemic redesign and that understanding that system is essential to changing it.

America built an economy around making autistic people easier to manage. It called that economy healthcare.

The numbers bear this out. In 2025, analysts estimated the U.S. Applied Behavior Analysis market at nearly eight billion dollars. Private equity firms had acquired more than five hundred autism therapy centers across forty-two states. State Medicaid programs were spending over two billion dollars annually on autism services — a figure that had more than tripled in four years. Federal investigators had identified hundreds of millions of dollars in improper payments. And the industry was still growing.

None of this happened by accident. It happened because the United States made a series of specific legal, financial, and institutional choices about what autism looks like, what it requires, and — most critically — who gets paid to address it.

This essay is about those choices. It is about the economic ecosystem that formed around a particular answer to the question of what autistic people need. And it is about what that answer has produced: a multi-billion-dollar industry oriented around making autistic people more institutionally legible — more compliant, more socially acceptable, less visibly autistic — while investing comparatively little in redesigning the schools, workplaces, healthcare systems, and public infrastructure that autistic people must navigate every day.

America built an industry around training autistic people to survive institutions. It built far less of an industry around redesigning institutions to reduce harm. That is not a funding gap. It is a design choice. And it has consequences.

I. Where This Started: The Origins of the Compliance Model

Applied Behavior Analysis did not begin with autism. It began with a philosophical claim about human beings: that what matters is not what a person experiences internally, but what they can be observed to do. And that if you control the environment, you can control the person.

In 1913, psychologist John B. Watson published a seminal paper arguing that psychology should concern itself not with feelings, intentions, or consciousness, but with observable, measurable behavior. B.F. Skinner built on this foundation in the 1930s and 1940s, developing operant conditioning — the systematic use of reinforcement and punishment to shape voluntary behavior. Skinner demonstrated that behavior could be reliably modified by manipulating its consequences. What he did not examine, with any sustained rigor, was what those consequences felt like from the inside.

In 1968, researchers Donald Baer, Montrose Wolf, and Todd Risley published what became the founding document of Applied Behavior Analysis as a formal discipline. Their paper, “Some Current Dimensions of Applied Behavior Analysis,” appeared in the inaugural issue of the Journal of Applied Behavior Analysis and defined the field’s seven core dimensions. The definition they codified is worth sitting with: ABA is the process of systematically applying interventions based on learning theory to improve “socially significant” behavior. That phrase — socially significant — is doing enormous work. It means behavior that matters to institutions, families, and communities. It does not require that it matter to the person displaying it.

The application of these principles to autistic children was largely the project of one man: Ole Ivar Lovaas, a Norwegian-American psychologist who established the Young Autism Project at UCLA in the early 1960s. Lovaas described his work as “building a person.” He was explicit that his goal was to make autistic children “indistinguishable from their peers” — meaning indistinguishable from non-autistic children in neurotypical classrooms and institutions.

What is rarely discussed with the directness it deserves: at UCLA in the 1960s and 1970s, Lovaas ran two behavioral modification programs simultaneously. One targeted autistic children. The other — known as the Feminine Boy Project — used the identical method of operant conditioning to eliminate behaviors deemed insufficiently masculine in young boys. That program became an early version of what is now recognized as conversion therapy. Both programs shared the same foundational logic: divergence from a behavioral norm is a problem, and the problem can be solved through systematic conditioning. The person undergoing the conditioning is not the authority on whether it is working. The institution is.

This is not a coincidence to be noted and moved past. It is the DNA of the industry. Lovaas did not see autistic children and gender-nonconforming children as requiring different treatments because he understood them as having different needs. He saw them as having the same problem — visible divergence from what institutions expected — and he applied the same solution. The clinical model that would become a multi-billion-dollar autism industry was built on the premise that the goal of treatment is conformity to a norm, not the flourishing of the individual.

Lovaas’s 1987 study became the pivot point that transformed a clinical model into an industry. He reported that 47 percent of children who received 40 hours per week of intensive ABA for two or more years achieved “indistinguishable” functioning in regular classrooms. That word — indistinguishable — became the field’s founding metric of success. Progress was defined as passing for non-autistic. The less visible the autism, the more successful the intervention. That definition has never fully left the field. It has been softened in the language of contemporary practice, but it remains structurally embedded in how outcomes are measured, how insurance reimbursement is triggered, and how progress is reported to families.

II. How the Market Formed: Insurance, Mandates, and the Revenue Engine

ABA did not become a near-eight-billion-dollar industry because autistic people demanded it. It became one because the legal and insurance architecture of the United States made it mandatory — and then made it profitable.

Beginning in 2001, states began requiring private insurance plans to cover ABA therapy for children diagnosed with autism spectrum disorder. By 2019, all fifty states had enacted some form of autism insurance mandate. The Affordable Care Act and a 2014 federal Medicaid directive further expanded coverage, requiring that medically necessary behavioral health services — including ABA — be covered for children enrolled in Medicaid. The effect was to create a guaranteed, nationally scaled revenue stream for a single category of autism intervention.

The financial logic was straightforward. ABA clinics charge tens of thousands of dollars per client per year for services that insurers are legally required to cover. The reimbursement structure rewards volume: more therapy hours, more billable sessions, more revenue. There is no equivalent financial mandate for environmental redesign, sensory-accessible infrastructure, adult employment supports, autistic-led peer services, or communication accommodations. Those things may be valuable, but they are not covered. ABA is covered. The market follows the money.

Insurance coverage for one specific intervention — and not others — does not merely reflect a clinical consensus. It creates one. When families are told their child qualifies for forty hours per week of ABA, covered by insurance, and simultaneously told that sensory-friendly school redesign is not a reimbursable service, the architecture of options shapes the decisions families make. The choice appears individual. The constraints are structural.

The certification board for behavior analysts was established in 1999 — one year before the first state insurance mandates. The timing matters. The mandates did not respond to a credentialed profession. They helped create one. Introduction of state insurance mandates was directly correlated with a sixteen percent increase in board-certified behavior analysts and a rapid expansion of the provider workforce. The field grew because the reimbursement infrastructure made it viable to grow. This is a market formation story, not a scientific consensus story.

III. The Private Equity Chapter

Once insurance mandates made ABA a legally guaranteed revenue stream, private equity moved in.

Between 2017 and 2022, private equity firms completed eighty-five percent of all mergers and acquisitions in the autism healthcare segment — a rate not found in any other segment of healthcare or any other industry. A 2026 study by researchers at Brown University, RAND, and Harvard, published in JAMA Pediatrics, identified 574 autism therapy centers owned by private equity firms as of 2024, spanning forty-two states. Most were acquired in a four-year window. Private equity was particularly concentrated in states with the highest autism prevalence and the most generous insurance mandates — a pattern consistent with financial extraction, not need-driven expansion.

The Center for Economic and Policy Research documented the dynamics of this acquisition wave in their 2023 report. Private equity firms targeted ABA providers specifically because the intervention is covered by state insurance mandates and “offers more opportunities for higher billing and revenue generation.” The regulatory environment was described by industry analysts as “favorable” — meaning high reimbursement, guaranteed payers, fragmented market, and thin oversight. The CEPR report included a case study of Blackstone’s acquisition of the Center for Autism and Related Disorders. After acquiring CARD, Blackstone cut training requirements, shifted from in-person to online training, and halved new-hire orientation. The cost of delivering the service went down. The billable rate did not.

A Wall Street Journal investigation found that autism services companies “have found lucrative opportunities to capitalize on the growing need for such care, sometimes outpacing regulators’ oversight.” Goldman Sachs Alternatives’ 2024 purchase of Center for Social Dynamics was cited by analysts as evidence of continued institutional appetite for scaled ABA assets. By early 2025, Proud Moments ABA — operating across more than seventy clinics in twelve states — was sold from one private equity firm to another. The children in those clinics were not part of the transaction. The revenue they generated was.

What private equity does to any industry it acquires is well documented. It maximizes billable activity, minimizes labor costs, reduces training and oversight, and exits within three to seven years. In autism services, these incentives run directly against clinical quality. High staff turnover — estimated at 35 to 40 percent annually among registered behavior technicians — means the children receiving therapy are regularly losing the practitioners they have built relationships with. Thirteen percent of registered behavior technicians report receiving no initial training before working with clients who display severe problem behavior. Twenty-nine percent report no ongoing training. The system is optimized for billing, not for care.

IV. What the Fraud Tells Us

When a payment structure rewards volume rather than outcomes, fraud becomes structurally predictable. And the fraud in ABA billing has been extensive.

HHS Office of Inspector General audits have now been conducted in multiple states. In Indiana, investigators found at least $56 million in improper Medicaid payments to ABA providers. In Wisconsin, $18.5 million. In Colorado, $77.8 million in clearly improper payments and an additional $207.4 million in potentially improper payments — a combined figure of nearly $285 million from a single state’s two-year audit period. Investigators found providers billing for children who were napping, eating, playing on tablets, or watching movies. Staff without credentials delivering therapy across multiple states. Session notes photocopied between patients or signed before sessions concluded. One child accumulated $677,448 in Medicaid payments from age two to eleven based on a single 2014 referral, with no subsequent reassessment.

Monthly Medicaid billing under key ABA codes increased 792 percent between 2019 and May 2024. The Massachusetts Medicaid Fraud Control Unit has publicly identified ABA therapy as “an area of behavioral health where fraud and abuse is rampant,” noting that high reimbursement rates have “enticed potential fraudsters into the space.” The FBI has opened investigations. State attorneys general have pursued enforcement actions. False Claims Act cases have been filed and settled.

The fraud is not incidental to the system. It is a symptom of it. A reimbursement structure that pays for hours of compliance training — with no requirement that those hours produce measurable improvement in the child’s quality of life — creates the conditions for billing fraud as a rational business strategy. The system cannot distinguish between genuine service and fabricated service because it was not designed to measure what actually matters.

V. What Gets Funded and What Does Not

The central structural argument of this essay is not that ABA is always harmful. It is that the American system has been built to fund one kind of response to autism — behavioral, compliance-oriented, child-focused, professionally delivered — while providing no equivalent mandate, reimbursement category, or financial incentive for other kinds of response.

ABA therapy is mandated insurance coverage in all fifty states. Environmental redesign is not a covered service anywhere. Sensory-accessible school construction is not reimbursable. Flexible workplace accommodations are not a billable clinical intervention. Communication access tools funded at the same scale as ABA do not exist. Autistic-led peer support networks do not generate insurance claims. Adult employment supports are discretionary in a way that childhood behavioral therapy legally is not.

The insurance architecture determines what counts as legitimate treatment. When a family is navigating an autism diagnosis, the reimbursable options are not a neutral menu from which they choose freely. They are a curated list of what the financial system has decided to fund. ABA appears on that list. The sensory-friendly classroom redesign their child might actually need does not.

This matters because what gets funded gets built. The ABA workforce grew by sixteen percent following state insurance mandates because the mandates made it financially viable to build that workforce. No comparable mandate has been created for environmental redesign, universal design in schools, adult autism support services, or autistic-led care. Those things exist in small pockets, funded by grants or goodwill. The compliance-based intervention system is funded by law.

The question this structure raises is not whether behavioral intervention can be helpful. Some of it can, particularly for autistic individuals with complex communication needs or self-injurious behavior. The question is: why has the United States chosen to fund compliance training at massive scale, while failing to fund systemic redesign at any comparable scale? The answer is not clinical. It is economic and political. Compliance training produces institutionally legible children. Institutional redesign does not produce revenue. One of these has a billing code. The other does not.

VI. What “Evidence-Based” Actually Measures

Proponents of ABA frequently invoke its evidence base as justification for the industry’s dominance. Over fifty years of research, they argue, supports the efficacy of ABA. The U.S. Surgeon General and the American Psychological Association have endorsed it. The science is settled.

But what does the evidence base actually measure? This is not a rhetorical question. It is a methodological one — and it has a specific, traceable answer.

ABA’s evidence base was built measuring outcomes that institutions defined as desirable: reduced frequency of flagged behaviors such as stimming, non-compliance, and what practitioners call “challenging behavior”; increased frequency of approved behaviors such as eye contact, verbal response, and social interaction. The original Lovaas study defined success as children becoming indistinguishable from their neurotypical peers. Subsequent research largely followed the same framework — asking whether visible autistic traits decreased, not whether the child’s quality of life improved, whether they felt safe, whether they experienced the intervention as respectful, or whether they would choose it as adults.

There was no robust evidence base built on autistic-defined outcomes. There was no longitudinal research following ABA-exposed children into adulthood to ask whether they thrived — whether they were employed, housed, connected, mentally healthy, and living lives they found meaningful. The evidence base answered the questions the institutions were asking. And those questions were about institutional manageability, not human flourishing.

This distinction matters because the evidence base then became the justification for insurance mandates. The insurance mandates created a market. The market generated revenue that funded provider expansion and more research. That research continued to measure the same outcomes — reduced visible divergence, increased compliance — using the same frameworks. The institutions that funded the research were the same institutions that benefited from the results.

The circularity is the point. The system defines what progress looks like. It funds the intervention that produces it. It uses the funding and the resulting research as evidence of legitimacy. Then it uses that legitimacy to justify more mandates, more coverage, more market growth. This is not science driving policy. It is policy producing the science that justifies itself. And the loop has been running for fifty years.

A growing body of research drawn from autistic adults — qualitative studies on burnout, masking, long-term trauma, and identity — tells a substantially different story about what intensive compliance training produces over time. That research was not funded by the insurance system. It was not used to design the mandates. It existed outside the loop, which is precisely why it took so long to be heard.

VII. What Autistic Adults Report

The clearest indictment of a childhood intervention system designed around institutional legibility is the adult outcomes it produces. Not the clinical metrics. The lives.

Autistic adults who underwent intensive behavioral intervention as children increasingly describe experiences of shame, loss of authentic self, and lasting psychological harm. Research on autistic burnout — a phenomenon characterized by chronic exhaustion, loss of skills, and reduced tolerance to stimulus — identifies prolonged masking as a primary risk factor. Masking refers to the conscious suppression of autistic traits in order to appear neurotypical: preparing scripts before social interactions, maintaining forced eye contact, suppressing stimming, performing emotions that do not match internal experience. These are precisely the behaviors that ABA has historically rewarded.

A 2020 study by Raymaker and colleagues, published in Autism in Adulthood, found that autistic burnout is a distinct and debilitating phenomenon — not occupational fatigue, not clinical depression, but a specific consequence of sustained masking and the chronic effort to pass as non-autistic in environments not designed for autistic people. The researchers explicitly noted the potential danger of teaching autistic people to mask their traits, and recommended including burnout education in suicide prevention programs. Subsequent research by Mantzalas and colleagues (2024) confirmed these findings, identifying masking as a central driver of burnout across the lifespan.

Suicide risk among autistic people is not marginal. A 2023 study estimated that one in four autistic people has planned or attempted suicide. Autistic adults are identified in UK national suicide prevention strategy as one of the highest-risk groups. The risk factors identified include masking, autistic burnout, social isolation, and lack of access to appropriate support. These are not random misfortunes. They are the predictable downstream effects of a system that has spent decades teaching autistic people that the way they naturally exist is unacceptable — and then withdrawing support when they reach adulthood.

Autistic adults who describe harm from behavioral intervention consistently distinguish between their critique of the system and criticism of their parents. The framing that appears repeatedly in the research is important: “They weren’t disrespectful. They just had a flawed paradigm for autism, and therefore, what they tried didn’t work. That doesn’t make them bad people.” The system shaped the options available to parents. Parents chose from what existed and what was covered. The flawed paradigm is not a parental failure. It is a structural one.

VIII. Who Is Missing From This Picture: Race, Gender, and Diagnostic Delay

The harms documented in this essay are not distributed equally. They compound along racial, gender, and socioeconomic lines in ways the industry has largely declined to examine.

Research consistently shows that autistic people of color are diagnosed later, referred for services less frequently, and when referred, are more likely to encounter compliance-based intervention with fewer cultural supports. The Autism and People of Color Collaborative has documented that autistic adults of color are less likely to be employed than white autistic counterparts, and that their exclusion from competitive employment increases risk of poverty, homelessness, and systemic harm. The ABA industry is concentrated in states with high autism prevalence and generous insurance mandates — but access within those states is not equal. The families with higher incomes, legal literacy, and flexible schedules are more likely to navigate the referral system successfully. Others face the same bureaucracy with fewer resources and more consequences.

Autistic girls and women are systematically underdiagnosed. Because diagnostic criteria were historically developed from research on young white boys, autistic girls who mask effectively are often missed entirely during the childhood window when services are most available. They are diagnosed later — sometimes not until adulthood — often after years of unexplained mental health struggles, burnout, and misattributed distress. The system designed to catch autistic children early was designed, implicitly, to catch a specific kind of autistic child. Many autistic people never fit that profile and spend their childhoods without either diagnosis or support.

Late-diagnosed autistic adults — particularly women and people of color — frequently report having spent decades masking without understanding why they were exhausted, developing complex mental health presentations that clinicians misread, and reaching adulthood with no services, no framework for understanding their own experience, and no pathway into the support systems that exist. The industry built around early childhood intervention does not reach them. The adult system barely exists. And the research base that drives policy was not built with them in mind.

IX. The School-to-Compliance Pipeline

The ABA industry does not operate in isolation. It is embedded in a larger institutional system that routes autistic children toward compliance-based intervention at multiple points — beginning in classrooms.

When an autistic child struggles in school, the institutional response follows a predictable pathway. Teachers identify behaviors that interfere with classroom management. The child is referred for special education evaluation. A Functional Behavioral Assessment is conducted — a document organized around identifying what maintains “problem behavior” and how to reduce its frequency. A Behavioral Intervention Plan is developed. If school-based supports are insufficient, a referral for outside ABA services follows. The child enters the insurance-funded therapy system.

At no point in this pathway is the classroom’s sensory environment evaluated as a potential cause of distress. At no point is the school’s demand for sustained stillness, eye contact, and neurotypical social performance examined as a design problem. The environment is treated as fixed. The child is treated as the variable to be adjusted.

The same traits that schools pathologize in autistic children are later used to filter autistic adults out of employment. Stimming, atypical eye contact, nonstandard communication style, sensory needs, monotropism, directness — these are flagged in childhood behavioral plans and then cited as professional deficiencies in adult hiring decisions. AI-based recruitment tools now analyze facial expression, vocal tone, and behavioral cues during video interviews, applying neurotypical behavioral norms at algorithmic scale. The pipeline from childhood compliance training to adult employment filtering is not broken. It is functioning exactly as designed.

X. The Labor Structure of the Industry

The economic logic of the ABA industry is most visible at its labor base.

The workforce delivering most ABA therapy consists primarily of Registered Behavior Technicians — entry-level practitioners who carry out treatment plans designed by Board Certified Behavior Analysts. The average RBT earns between $18 and $21 per hour. The industry bills insurers at rates many times that. The margin between what labor costs and what the service bills is where the industry’s profits live.

Turnover among RBTs runs at an estimated 35 to 40 percent annually. Approximately 72 percent of ABA professionals report moderate to high burnout. The ABA Workers Union has put it plainly: “RBTs are not failing at their jobs — their jobs are failing at being a reasonable job.” The same report documents that some RBTs are removed from cases when families feel “uncomfortable” with them — a dynamic that intersects with race and identity in ways the industry has not seriously addressed.

The children receiving therapy bear the cost of this labor model. High turnover means they regularly lose relationships with practitioners they have built connections with. Undertrained technicians deliver interventions with documented risks for compliance-based harm. The system is not designed around the child’s experience of continuity, trust, or emotional safety. It is designed around the billable hour.

Private equity’s preferred intervention when it acquires any sector is to reduce the cost of labor while maintaining or increasing the billing rate. In autism services, this means more children per technician, shorter training periods, higher productivity quotas, and faster turnover cycles. This is not a clinical program. It is a staffing model wearing clinical language.

XI. The Language of the System

How a system describes what it does is not incidental to what it does. The vocabulary of the autism treatment industry encodes a theory of autism — and that theory serves the industry by making its services appear necessary.

Consider the standard terminology. “Reduction of behaviors” frames autistic traits as categorically pathological — behaviors are things to be reduced, not expressions of a different neurology to be understood. “Improved compliance” explicitly defines institutional conformity as therapeutic progress, without asking what the child is complying with or at what cost. “Socially appropriate behavior” defines appropriateness by neurotypical standards, making neurotypical behavioral norms the invisible default of what “appropriate” means. “Challenging behavior” locates the challenge in the child rather than in the mismatch between the child and the environment. “Early intervention urgency” generates fear-based demand that accelerates referral before families have access to the full range of perspectives — including those from autistic adults who experienced the same interventions as children.

This language is not neutral. A system that defined autistic traits as variations requiring accommodation rather than behaviors requiring modification would not generate the same demand for compliance training. The vocabulary maintains a market. And it shapes what families understand to be at stake — which is why it appears not just in clinical literature but in insurance authorization documents, school IEPs, provider marketing, and the advice given by pediatricians at diagnosis.

The marketing language of the parent fear economy extends this framing into homes. Parents of newly diagnosed autistic children are frequently told that the intervention window is narrow, that without intensive early treatment their child risks a diminished future, that acting quickly is essential. This is not clinically neutral information delivered in a vacuum. It is emotionally loaded framing designed — consciously or not — to accelerate enrollment in the one available, insured, heavily marketed intervention before families have time to encounter the breadth of perspectives that exist. Fear, in this system, is a referral pathway.

XII. The Adult Cliff

If the childhood intervention system were genuinely oriented around long-term autistic flourishing, adult outcomes would look different. They do not.

Employment rates for autistic adults in the United States are estimated at between 15 and 23 percent — a figure that has not meaningfully improved in a decade. Some estimates place autistic unemployment as high as 85 percent among those with college educations. Among young autistic adults in their twenties, employment rates are lower than those of people with intellectual disabilities, learning disabilities, or emotional disabilities. This is not a pipeline failure. The children are being served. Then they turn eighteen, and the system largely disappears.

Scientific American reported in 2025 that an estimated 99 percent of autistic adults who need employment supports receive none. The federal government invests fifteen billion dollars annually in special education, with the stated goal of facilitating employment for people with disabilities. There is no mandated funding for services after high school. The investment in childhood is expected to mature on its own. It does not.

Housing is equally stark. Waitlists for licensed group homes or supported living arrangements can be years long in many states. More than half a million Americans with intellectual and developmental disabilities are currently waiting for government funding for long-term community-based care. Unlike special education — which is a federal entitlement — Medicaid waiver services for adults allow states to maintain waiting lists and limit the number of people served. The entitlement ends at graduation.

One in four families with an autistic child struggles with food or housing insecurity. Adult autistic people face elevated rates of poverty, underemployment, isolation, and mental health crisis. These are the outcomes of a system that concentrated its resources on making autistic children appear neurotypical in institutional settings — and then released those children into adulthood with no equivalent structural support.

The question this raises is simple: if the childhood system was working, why are the adult outcomes this bad? The answer is that the childhood system was not designed to produce adult flourishing. It was designed to produce institutional legibility in childhood. Institutions profit when autistic children can be managed in classrooms and counted as compliant. They do not profit from autistic adults thriving independently. So the investment stops where the institutional need stops.

XIII. The International Mirror

The American model is not universal. It is a specifically American market formation — shaped by specific legal structures, regulatory gaps, and financial incentives that do not exist in the same configuration elsewhere.

Ninety-four percent of all Board Certified Behavior Analysts globally reside in the United States. ABA is considered “treatment as usual” in North America. In most European countries, it is not. The UK’s National Health Service emphasizes early intervention through speech and language therapy and specialized schools. Australia takes a holistic approach to autism support that centers occupational therapy, sensory integration, and social participation. Germany, the Netherlands, and Sweden have developed inclusive education and employment frameworks that do not center compliance-based behavioral intervention as the default response to autism.

None of these countries have eliminated autistic hardship. Autistic people face barriers across all of them. But none of them built a private equity-backed, insurance-mandated, multi-billion-dollar compliance industry as their primary response. The existence of different approaches in other developed countries is evidence that the American model is a choice — not an inevitable response to a clinical reality. The choice America made was shaped by a for-profit healthcare system, a credentialing board established just ahead of the insurance mandate wave, a fragmented regulatory environment with minimal provider standards, and a political tradition of funding services to individuals rather than redesigning systems. Those are not universal features of modern healthcare. They are American ones.

XIV. What Redesign Would Actually Look Like

This essay is not an argument that support is harmful or that autistic people do not need assistance. Many do — some substantially. The argument is about what kind of assistance the system has been built to provide, and what it has been built to ignore.

A system genuinely designed around autistic flourishing would look different. It would invest in sensory-accessible schools not as a special accommodation but as a baseline design standard. It would fund autistic-led peer support networks as a recognized and reimbursable form of care. It would measure outcomes in terms of autistic people’s own assessments of their wellbeing, autonomy, and quality of life — not in terms of reduced visible divergence. It would maintain adult services at the same legal entitlement level as childhood special education. It would fund employment redesign — work-sample hiring, reduced social performance evaluation, flexible communication modalities — not just vocational rehabilitation programs that route autistic adults into subminimum wage sheltered employment.

It would also support families differently. Rather than flooding parents of newly diagnosed children with fear-based urgency that accelerates enrollment in the available, insured intervention, a redesigned system would give families time, information, and access to autistic adult perspectives before making decisions about intensive early intervention. Parents are not the problem. The architecture that shapes what they are told and what they are offered is.

The distinction that matters is not “support versus no support.” It is support designed around what autistic people say they need versus support designed around what institutions need from autistic people. Those are not the same project. For decades, the United States has funded the latter while calling it the former.

Conclusion

America did not build an eight-billion-dollar industry around autism because it cared deeply about autistic people’s quality of life. It built that industry because a specific set of legal mandates, insurance structures, and regulatory gaps made autistic normalization extraordinarily profitable — and because the institutions that needed autistic children to appear manageable had the power to define what “treatment” meant.

The fraud investigations, the private equity extraction, the workforce crisis, the adult outcomes — none of these are failures of the system. They are expressions of it. A system built to fund compliance produces an industry optimized for compliance. A system that measures success by the reduced visibility of autistic traits produces an industry that suppresses autistic traits. A system with no adult services mandate produces autistic adults with no services.

The autistic adults who came through this system and report lasting harm are not outliers. They are the evidence. When the same pattern appears across thousands of people — shame, burnout, masked identity, crisis in adulthood — that is not individual misfortune. That is structural outcome.

Changing it requires more than reforming individual practices. It requires acknowledging that the problem was never primarily clinical. It was always economic. And economic problems require economic solutions: different mandates, different reimbursement structures, different definitions of what counts as progress, and a genuine commitment to funding what autistic people say they need — not just what institutions say they can manage.

Support is not the problem. Support designed around normalization as the goal — and funded by a system that profits from that goal — is.

“When the same experiences are repeatedly dismissed across different people and contexts, that is not randomness. That is structure.”

Works Cited

I. Founding and Historical Documents

Baer, Donald M., Montrose M. Wolf, and Todd R. Risley. “Some Current Dimensions of Applied Behavior Analysis.” Journal of Applied Behavior Analysis 1, no. 1 (Spring 1968): 91-97. https://doi.org/10.1901/jaba.1968.1-91.

Lovaas, O. Ivar. “Behavioral Treatment and Normal Educational and Intellectual Functioning in Young Autistic Children.” Journal of Consulting and Clinical Psychology 55, no. 1 (February 1987): 3-9. https://doi.org/10.1037/0022-006X.55.1.3.

Rekers, George A., and O. Ivar Lovaas. “Behavioral Treatment of Deviant Sex-Role Behaviors in a Male Child.” Journal of Applied Behavior Analysis 7, no. 2 (Summer 1974): 173-190. https://doi.org/10.1901/jaba.1974.7-173.

II. Peer-Reviewed Research and Academic Sources

Appelbaum, Eileen, Rosemary Batt, and Quynh Trang Nguyen. “Pocketing Money Meant for Kids: Private Equity in Autism Services.” Center for Economic and Policy Research, June 2023. https://cepr.net/publications/pocketing-money-meant-for-kids-private-equity-in-autism-services/.

Arnold, Daniel R., et al. “Private Equity in Autism Services.” JAMA Pediatrics, January 5, 2026. https://doi.org/10.1001/jamapediatrics.2025.5443.

Autistic Self Advocacy Network. For Whose Benefit? Evidence, Ethics, and Effectiveness of Autism Interventions. Washington, DC: ASAN, 2021. https://autisticadvocacy.org/wp-content/uploads/2021/12/ACWP-Ethics-of-Intervention.pdf.

Kupferstein, Henny. “Evidence of Increased PTSD Symptoms in Autistics Exposed to Applied Behavior Analysis.” Advances in Autism 4, no. 1 (2018): 19-29. https://doi.org/10.1108/AIA-08-2017-0016.

[Note: As of late 2025, the publisher has issued an Expression of Concern for this article while an investigation into research standards is ongoing. Corroborate PTSD-related claims with Raymaker et al. (2020) and Mantzalas et al. (2024).]

Mantzalas, Jane, Amanda L. Richdale, and Cheryl Dissanayake. “Measuring and Validating Autistic Burnout.” Autism Research 17, no. 4 (2024): 750-765. https://doi.org/10.1002/aur.3129.

Pryke-Hobbes, Amber, et al. “The Workplace Masking Experiences of Autistic, Non-Autistic Neurodivergent and Neurotypical Adults in the UK.” PLOS ONE 18, no. 9 (September 6, 2023): e0290001. https://doi.org/10.1371/journal.pone.0290001.

Pyne, Jake. “’Building a Person’: Legal and Clinical Personhood for Autistic and Trans Children in Ontario.” Canadian Journal of Law and Society 35, no. 2 (2020): 341-365. https://doi.org/10.1017/cls.2020.8.

Raymaker, Dora M., Alan R. Teo, Nicole A. Steckler, Brandy Lentz, Mirah Scharer, Austin Delos Santos, Steven K. Kapp, Morrigan Hunter, Andee Joyce, and Christina Nicolaidis. “’Having All of Your Internal Resources Exhausted Beyond Measure and Being Left with No Clean-Up Crew’: Defining Autistic Burnout.” Autism in Adulthood 2, no. 2 (June 2020): 132-143. https://doi.org/10.1089/aut.2019.0079.

Zaks, Zosia. “Moving to a Neurodiversity-Affirming Paradigm in the Support System: Autistic Professionals as Paradigm Change.” Neurodiversity 3 (2025). https://doi.org/10.1177/27546330241294138.

III. Government Audits, Institutional Reports, and Legal Records

U.S. Department of Health and Human Services, Office of Inspector General. Medicaid Payments for Applied Behavior Analysis Services: Indiana. Washington, DC: HHS-OIG, December 2024.

U.S. Department of Health and Human Services, Office of Inspector General. Medicaid Payments for Applied Behavior Analysis Services: Wisconsin. Washington, DC: HHS-OIG, July 2025.

U.S. Department of Health and Human Services, Office of Inspector General. Medicaid Payments for Applied Behavior Analysis Services: Colorado. Washington, DC: HHS-OIG, 2026.

Autism Speaks. 2025 Autism by the Numbers Annual Report. New York: Autism Speaks, 2025. https://www.autismspeaks.org/science-news/2025-autism-numbers-annual-report-now-live.

Drexel Autism Institute. Housing Assistance for People with Autism: A Policy Report. Philadelphia: Drexel University, January 2025. https://policyimpactproject.org/housing-assistance/.

IV. Investigative Journalism and News Reporting

ABA Workers Union. “The Data Are Clear: The ABA Workforce Is in Crisis.” ABAWorkersUnion.org, March 25, 2026. https://www.abaworkersunion.org/post/the-data-are-clear-the-ABA-workforce-is-in-crisis/.

Bannow, Tara. “Federal Medicaid Audit Finds Massive Overpayment for Autism Therapy in Colorado.” STAT News, March 2, 2026. https://www.statnews.com/2026/03/02/hhs-medicaid-audit-finds-autism-therapy-overpayment-colorado/.

Behavioral Health Business. “ABA Volume Skyrocketed by 267% from 2019 to 2024.” December 22, 2025. https://bhbusiness.com/2025/12/22/aba-volume-skyrocketed-by-267-from-2019-to-2024/.

Brown University News. “Private Equity Firms Acquired More Than 500 Autism Centers in Past Decade, Study Shows.” January 7, 2026. https://www.brown.edu/news/2026-01-07/private-equity-autism-centers.

Devita-Raeburn, Elizabeth. “America’s Most Popular Autism Therapy May Not Work - and May Seriously Harm Patients’ Mental Health.” The 74 Million, March 6, 2024. https://www.the74million.org/article/americas-most-popular-autism-therapy-may-not-work-and-may-seriously-harm-patients-mental-health/.

Disability Scoop. “Private Equity Increasingly Taking Over Autism Therapy Centers.” February 23, 2026. https://www.disabilityscoop.com/2026/02/23/private-equity-increasingly-taking-over-autism-therapy-centers/31871/.

Groppe, Maureen. “Families Worry as Cost of Autism Therapy Comes Under State Scrutiny.” Stateline, November 25, 2025. https://stateline.org/2025/11/25/families-worry-as-cost-of-autism-therapy-comes-under-state-scrutiny/.

Leys, Tony. “The Wait for Government Disability Services Can Last Years. Some States Are Trying to Change That.” PBS NewsHour, May 15, 2023. https://www.pbs.org/newshour/nation/the-wait-for-government-disability-services-can-l

ast-years-some-states-are-trying-to-change-that.

Private Equity Stakeholder Project. “PE’s Failed Autism Bets Harm Workers and Consumers.” February 12, 2025. https://pestakeholder.org/news/pes-failed-autism-failed-bets-harm-workers-and-consumers/.

Private Equity Stakeholder Project. “Private Equity Health Care Acquisitions - February 2025.” March 2025. https://pestakeholder.org/news/private-equity-health-care-acquisitions-february-2025/.

Scott, Katherine. “Autism Services Blocked by Administrative Barriers.” DCReport, February 12, 2026. https://www.dcreport.org/2026/02/12/autism-services-blocked-by-administrative-barriers/.

Whistleblower Law Collaborative. “Medicaid Fraud Risks in ABA Therapy.” March 2026. https://www.whistleblowerllc.com/medicaid-fraud-risks-in-aba-therapy/.

V. Employment, Housing, and Adult Outcomes

Autism and People of Color Collaborative. Autism, Race, and Employment. Autistic People of Color Fund, April 2024. https://autismandrace.com/wp-content/uploads/2024/04/APOC-Employment-2024-04-07-1.pdf.

Autism Research Centre, University of Cambridge. “Suicidality and Autism.” 2023. https://www.autismresearchcentre.com/projects/autism-and-suicidality/.

Scientific American. “A Broken Employment System Leaves Autistic Adults Stranded.” February 24, 2025. https://www.scientificamerican.com/article/a-broken-employment-system-leaves-autistic-adults-stranded/.

VI. Market and Industry Data

Credence Research. Applied Behavior Analysis Market Size, Growth and Forecast 2032. 2025. https://www.credenceresearch.com/report/applied-behavior-analysis-market.

GMI Insights. U.S. Applied Behavior Analysis Market Size & Growth Outlook 2032. 2024. https://www.gminsights.com/industry-analysis/us-applied-behavior-analysis-market.

Marketdata LLC. “U.S. Autism Treatment Centers Market Report 2025.” May 2025. Via BusinessWire. https://www.businesswire.com/news/home/20250526138164/en/U.S.-Autism-Treatment-Centers-Market-Report-2025.

VII. Books

Silberman, Steve. NeuroTribes: The Legacy of Autism and the Future of Neurodiversity. New York: Avery, 2015.

© 2026 Eva Redford. All rights reserved.

Editor’s note: This essay draws on peer-reviewed research, meta-analytic literature, and published community surveys to examine the evidence for harm associated with Applied Behavior Analysis (ABA). It is written from a systems perspective, not a clinical one. The author is an autistic adult and institutional design researcher.

Applied Behavior Analysis is the most widely recommended, most heavily funded autism intervention in the United States. It is covered by nearly all state Medicaid programs. Pediatricians recommend it. School districts fund it. Insurance pays for it. And for decades, its status as the gold-standard treatment for autism has been treated as settled.

It is not settled.

This essay is not arguing that every ABA program is abusive, or that no autistic person has ever benefited from behavioral intervention. Some have. Modern, child-led implementations are meaningfully different from the compliance-driven models that generate the most criticism, and practitioners working in good faith within an evolved framework deserve that distinction. What this essay is arguing is narrower and, in some ways, harder to dismiss: a field that has consistently failed to look for harm cannot credibly claim it has found none. A growing body of research, including some of the most rigorous large-scale evidence syntheses in the field, is raising serious questions not just about whether ABA works, but about whether the research was ever designed to catch what it costs. Those are two different problems. This essay is concerned with both.

A Framework Built to Erase

To understand the current debate, you have to understand what ABA was designed to do.

Applied Behavior Analysis descends from the work of psychologist Ivar Lovaas, whose methods in the 1960s and 1970s included electric shock, food deprivation, and physical restraint as aversive tools to suppress autistic behavior. The 1965 Life Magazine photographs of Lovaas’s methods, depicting children being subjected to painful stimuli in the name of treatment, are not ancient history. They are the documented origin of the framework that underlies modern ABA.

Lovaas himself articulated the premise clearly: autistic children needed to be built from scratch. The goal was not to support an autistic child’s development. It was to produce a child who appeared neurotypical. That goal, critics argue, never fully left the field, even as the methods modernized and the explicit use of aversives declined.

Modern ABA programs typically emphasize positive reinforcement and describe their aims in terms of skill building. Many practitioners are genuinely working in good faith within an evolved version of the framework. But the foundational premise — that autism is a collection of behaviors to be extinguished rather than a neurotype to be supported — is still embedded in how ABA treatment goals are set, who sets them, and what counts as success.

The Research Was Not Designed to Find Harm

When a treatment is described as “evidence-based,” that phrase carries weight. It implies a body of research with consistent methodology, adequate controls, and systematic attention to both benefits and side effects. By those standards, the ABA literature has a problem.

Project AIM (Autism Intervention Meta-Analysis) is a multi-institution research collaboration that has conducted the most comprehensive systematic review of early childhood autism intervention research to date. Their updated 2023 analysis, published in the BMJ and covering studies from 1975 through 2021, did not find that ABA was ineffective. What they found was more structurally damaging: the quality of the evidence is too inconsistent to justify the confidence placed in it. Kristen Bottema-Beutel, a Boston College professor and one of the study’s principal authors, has stated plainly that “it doesn’t seem to be on anyone’s radar that these interventions could cause harm,” describing a field with a striking unwillingness to examine that question.

That unwillingness shows up in the data. Project AIM coded every included study for whether researchers monitored and reported adverse events. Among studies that described no monitoring procedures at all, 76% reported no adverse events. Among studies that did implement at least some monitoring procedures, only 50% reported no adverse events. When researchers looked for harm, they found it more often. When they did not look, they did not find it. This is not a minor methodological footnote. It is a structural feature of how ABA research has been produced.

The U.S. Department of Defense’s Autism Care Demonstration adds another layer of concern. This is not a small academic study with self-selected participants. It is a federal program’s own internal outcome data, covering thousands of military dependents receiving ABA services over multiple years. Its annual reports found weak evidence for ABA overall: in one analysis, approximately 76% of recipients showed no improvement after a year, roughly 9% showed measurable worsening, and only 16% showed improvement per therapist measures. Spending on ABA through the program increased significantly in the same period. Those findings do not prove that ABA caused harm in every case. But they make it very difficult to sustain the claim that the evidence base is robust.

The PTSD Association

In 2018, researcher Henny Kupferstein published a study in the journal Advances in Autism examining posttraumatic stress symptoms in autistic people who had been exposed to ABA. Surveying 460 respondents, she found that exposure to ABA predicted higher rates and more severe posttraumatic stress symptoms. Adult participants in particular showed significantly greater symptom severity with increasing exposure to the intervention.

This study has been contested, and the contestation is legitimate. The research relied on an online survey with self-selected participants, which introduces significant methodological limitations. Those limitations are real. But the response from much of the ABA research establishment was not to design better studies. It was to dismiss the finding. A 2025 letter to the editor in the same journal summarized the situation clearly: “Unless and until there is clear scientific evidence against Kupferstein’s preliminary finding of a link between ABA and PTSD, we think this finding should be taken seriously.” Seven years later, the rigorous follow-up studies that would resolve the question either way remain largely absent from the literature.

A contested finding that no one has tried to disprove is not proof of innocence. It is proof of a field that does not want to ask the question.

Compliance Is Not the Same as Wellbeing

ABA’s core mechanism is behavior modification: reinforcing desired behaviors and extinguishing undesired ones. What counts as “undesired” has historically included stimming, atypical eye contact, unconventional social responses, and other behaviors that are, from a neurodivergence perspective, natural autistic traits rather than deficits.

Critics argue that ABA often teaches masking — the suppression of visible autistic traits in order to appear neurotypical — rather than genuine skill development. The research on masking makes this concern concrete. A 2024 study published in Autism in Adulthood found that increased autistic masking is associated with increased depression, anxiety, burnout, and exhaustion. Learning to pass as neurotypical is not the same as being supported. In many cases, the research suggests, it works in the opposite direction.

There is also the question of what ABA studies actually measure. The majority of outcome research focuses on behavior change, compliance, language scores, and adaptive skill assessments — not on mental health, self-determination, or subjective wellbeing. A 2024 scoping review documented that many autistic individuals with lived experience of ABA describe their treatment as traumatic precisely because the intervention was oriented toward how they appeared rather than how they felt. Gains in eye contact and reduced stimming look like success in a study that only measures those variables. They do not tell you what the child experienced, or what they carry forward into adulthood.

Long-term outcome research compounds this problem. There is limited evidence that early intensive ABA improves adult quality of life, mental health, employment, or relationships over the long term. The field has invested heavily in measuring short-term behavioral compliance and comparatively little in following up on whether the people who received intensive intervention in childhood are doing well as adults. That gap is not accidental. It reflects what the system was designed to care about, and it tells you exactly who the research was built for.

Who Gets to Define the Problem

Running beneath the empirical debates is an ethical question that the research has largely avoided: who decides what autistic behavior needs to change, and who consents to the process?

ABA is most commonly delivered to children who have little or no meaningful assent in the treatment. Goals are typically set by clinicians and caregivers, and the autistic person’s own preferences and experience are rarely part of the formal design process. A 2025 paper in the Journal of Social Issues examined autistic adults’ lived experiences of ABA and found that the field’s focus on compliance to norms set by others — rather than self-determined goals — was central to participants’ negative experiences. The experience of being trained to perform behaviors one did not choose, in service of an appearance of normalcy one did not want, recurred across accounts.

This is a separate line of critique from whether ABA feels distressing in the moment. The ethical argument is about the structure of the intervention itself: a framework built to shape a child toward an external standard, without that child’s input, is not neutral by design. When the external standard is neurotypicality, the structural message is that the child’s natural way of being is the problem.

But there is a deeper institutional question beneath even that. When clinicians, insurers, school systems, and policymakers all define autism as a behavioral problem requiring correction, and autistic people are structurally excluded from that definitional process, the treatment goals that emerge will reliably reflect the priorities of everyone except the person receiving the treatment. This is not a failure of individual practitioners. It is how institutions that embed a narrow prototype of normal cognition produce treatment frameworks: by deciding, in advance, who the intervention is for and what success looks like, without asking the people whose lives are at stake.

Precision Over Polemic

It is worth being precise. The existing research does not prove that ABA causes PTSD in every recipient, or that no autistic person has ever benefited from behavioral intervention, or that all ABA practitioners intend harm. Some autistic adults report that ABA helped them develop skills they wanted. Modern, naturalistic, child-led ABA approaches are meaningfully different from the compliance-driven, high-intensity models that generate the most criticism. Harm is not uniformly distributed across all implementations of a framework with dozens of variations.

But what the research does establish is this: the evidence base for ABA is built on studies that largely did not look for harm. When researchers looked, they found it more often. The most cited foundational study in the field’s history has basic methodological problems that went unchallenged for decades. The largest real-world dataset shows weak outcomes and some measurable worsening. The association between ABA exposure and posttraumatic stress symptoms has not been resolved by better research. And the autistic community — the people with direct experience of the intervention — has consistently and clearly communicated concern that has been structurally excluded from the evidentiary record.

The strongest version of the harm argument is not that ABA is always abusive. It is that a field designed to measure behavioral compliance has systematically failed to measure the things that matter most to autistic people: autonomy, dignity, mental health, and the experience of being treated as a person rather than a set of behaviors to correct.

When the same experiences are repeatedly dismissed across different people and contexts, that is not randomness. That is structure.

Works Cited

Sandbank, M., Bottema-Beutel, K., et al. (2023). Autism intervention meta-analysis of early childhood studies (Project AIM): updated systematic review and secondary analysis. BMJ, 383, e076733. https://doi.org/10.1136/bmj-2023-076733

Kupferstein, H. (2018). Evidence of increased PTSD symptoms in autistics exposed to applied behavior analysis. Advances in Autism, 4(1), 19–29. https://doi.org/10.1108/AIA-08-2017-0016 [Expression of Concern issued by the journal; investigation ongoing as of 2025. Methodological limitations discussed in essay body.]

Evans, J.A., Krumrei-Mancuso, E.J., & Rouse, S.V. (2024). What you are hiding could be hurting you: Autistic masking in relation to mental health, interpersonal trauma, authenticity, and self-esteem. Autism in Adulthood, 6(2), 229–240. https://doi.org/10.1089/aut.2022.0115

Marshall, S. (2025). Autistic experiences of applied behavior analysis (ABA): Toward improved autistic-centered supports. Journal of Social Issues. Advance online publication. https://doi.org/10.1111/josi.70037

Najeeb, P. & Quadt, L. (2024). Autistic well-being: A scoping review of scientific studies from a neurodiversity-affirmative perspective. Neurodiversity, 2. https://doi.org/10.1177/27546330241233088

Wilkenfeld, D.A. & McCarthy, A.M. (2020). Ethical concerns with applied behavior analysis for autism spectrum disorder. Kennedy Institute of Ethics Journal, 30(1), 31–69. https://doi.org/10.1353/ken.2020.0000

Many women are not diagnosed with autism in childhood because the system was built to recognize a male prototype, which means they spend years being told they are anxious, depressed, too sensitive, too difficult, or simply fine when the real pattern was there all along; this is not a failure of the person, but a failure of recognition, and this guide is not asking you to prove anything, only to look at your actual experience and notice whether it has a name.

Social Communication

Autistic women often develop sophisticated workarounds. These experiences may look like shyness, introversion, or anxiety. They are not.

→ Take things literally; miss sarcasm, subtext, or implied meaning
→ Rehearse conversations beforehand or replay them afterward
→ Struggle to know when it is your turn to speak
→ Find group conversations harder to follow than one-on-one
→ Feel like you are performing “normal” rather than being yourself
→ Mirror the speech, accent, or mannerisms of people around you
→ Make friends easily as a child but struggle to keep them long-term
→ Prefer communicating in writing over speaking in real time
→ Feel exhausted — not just tired — after social interaction
→ Misread facial expressions or find eye contact uncomfortable or effortful

Masking and Camouflaging

Masking is the learned performance of neurotypical behaviour. It is one of the primary reasons autism in women is missed. It is also one of the primary sources of burnout.

The skill is real. The cost is real. The fact that you got good at it does not mean you were not autistic. It means you adapted to a system that gave you no other option.

→ Feel like a different person depending on who you are with
→ Study people to figure out how to act correctly in social situations
→ Cannot fully relax or be yourself in most social environments
→ Feel relief — not just preference — when plans are cancelled
→ Have been told you “seem fine” or “don’t seem autistic”
→ Experience a crash or shutdown after prolonged social performance
→ Know the rules of social interaction but feel like you had to memorise them
→ Have scripted phrases or responses you rely on in conversation
→ Hide stimming behaviours (fidgeting, rocking, tapping) in public
→ Feel like your “public self” and “private self” are very different people

Sensory Sensitivities

Sensory differences are frequently dismissed as fussiness or anxiety. They are a core feature of autism. They often go unidentified in women because women learn early to manage, minimize, and not complain.

If your body has always responded to the world more intensely than those around you seem to — that is information.

→ Clothing textures cause discomfort, distraction, or genuine distress
→ Certain sounds are physically painful or impossible to filter out
→ Strong smells are overwhelming in a way others do not seem to experience
→ Bright or flickering lights cause headaches, shutdown, or sensory overload
→ Crowded or loud environments drain you very quickly
→ Particular food textures are intolerable, not merely disliked
→ Sensitive to temperature — always too hot, too cold, or both
→ Unexpected or unwanted physical touch feels intensely uncomfortable
→ Background noise is difficult to separate from foreground conversation
→ You seek sensory input (pressure, movement, texture) to regulate or calm

Special Interests and Deep Focus

In girls and women, special interests are often dismissed as normal hobbies because they align with socially acceptable topics. The intensity is the signal. It is not about what you are interested in. It is about what that interest does to you — the depth, the pull, the way time disappears.

→ Areas of intense, absorbing interest that others find excessive
→ An unusual amount of detail known about specific subjects
→ Losing track of time when engaged with your main interest
→ Feeling misunderstood or “too much” when sharing things you care about
→ Interests that shift over time but always have this same quality of depth
→ Reading, research, and information-gathering that feels genuinely satisfying, not like work
→ Difficulty engaging with topics that do not interest you at all
→ Frustration when conversations move away from what you were discussing
→ Preferring to spend time pursuing your interest over most social activity

Executive Function

Executive function difficulties often look like laziness, irresponsibility, or low motivation from the outside.

They are structural, not attitudinal. The person who cannot start the task they want to do, who loses entire days to a single disruption, who has seventeen detailed plans and no completed actions — that person is not failing to try. Their brain is working differently.

→ Starting tasks is often harder than doing them once started
→ Tasks with many steps are difficult to initiate without external support
→ Struggle to switch between tasks, especially when interrupted
→ Time feels different — you lose track of it or it passes in unexpected ways
→ Rely heavily on routines; disruption causes disproportionate distress
→ Forget to eat, drink water, or take medication unless externally reminded
→ Procrastinate on tasks even when you want to do them and the stakes are high
→ Find open-ended tasks harder than tasks with clear rules and structure
→ Difficulty prioritising — everything feels equally urgent or equally unimportant
→ Make detailed plans but struggle to translate them into action

Emotional Regulation

Autistic emotional experiences are often intense and difficult to process in real time. This is frequently misread as emotional instability, BPD, or anxiety disorder. The diagnosis changes. The experience does not.

→ Emotions feel very intense — more so than other people seem to experience
→ Difficulty identifying what you are feeling in the moment (alexithymia)
→ Reaching a point of emotional overload — meltdown — that feels sudden to others
→ Alternatively: shutting down completely and going emotionally flat under stress
→ Recovering from emotional events more slowly than expected
→ Small frustrations or changes triggering disproportionately large responses
→ Struggle to communicate what you are feeling while you are feeling it
→ Ruminating on interactions, decisions, or events for days or weeks
→ Finding it hard to move on after conflict or perceived rejection
→ Being told you are too sensitive, too reactive, or too much

Interoception and Body Awareness

Interoception is the sense that tells you what is happening inside your body. Many autistic people have reduced or unreliable access to these signals. This makes it difficult to know when you are hungry, in pain, ill, or emotionally dysregulated — until it has already become a problem.

This is also why so many autistic women have been dismissed by medical professionals. When the internal signal is unclear, describing it accurately is nearly impossible. That is not exaggeration. That is not minimisation. That is a different kind of nervous system.

→ Often do not notice hunger or thirst until it becomes urgent
→ Difficulty identifying where pain is coming from or how serious it is
→ Frequently surprised to learn you are ill, tired, or unwell
→ Struggle to identify the physical sensations of emotions in your body
→ Difficulty knowing when you need to use the bathroom until it is urgent
→ Pain tolerance that is unusually high or unusually low
→ Difficulty describing symptoms to doctors — the internal signal is unclear
→ Being dismissed by medical professionals as exaggerating or minimising

Relationships and Belonging

Many autistic women describe the same experience: the persistent sense of being adjacent to belonging rather than inside it. Present in the room. Performing correctly. And still separate.

→ Always felt “different” without being able to explain why
→ Find it easier to connect with animals, younger children, or older adults
→ Struggle to maintain friendships without a shared activity or structure
→ Experienced intense, one-sided friendships — you more invested than them
→ Often do not understand why friendships fade without apparent cause
→ Find romantic relationships confusing — particularly unspoken expectations
→ Have been taken advantage of due to difficulty reading others’ intentions
→ Feel most comfortable with people who are direct and explicit
→ Prefer small groups or one-on-one — not crowds, parties, or gatherings
→ Deeply value fairness and consistency; find social “rules” that shift arbitrarily upsetting

Burnout History

Autistic burnout is distinct from general burnout. It results from sustained masking and overload, and it can last months or years. It is one of the most underrecognised features of autism in women — because it tends to arrive after years of looking functional.

By the time it hits, there is no reserve left. And it is frequently misidentified as depression, because the surface symptoms overlap. But the structure underneath is different.

→ Periods of complete functional collapse — unable to work, socialise, or manage daily tasks
→ These periods followed prolonged high-demand phases (new job, major transition, sustained stress)
→ Recovery required significant withdrawal from social and professional life
→ During these periods, things that were manageable before became impossible
→ Treatment for depression or anxiety that did not fully resolve
→ Chronic exhaustion that is not explained by sleep or activity levels
→ Capacity that fluctuates dramatically — high functioning, then not at all
→ Extended recovery time after events others handled without difficulty

Prior Diagnoses and Misattributions

Autism in women is frequently misidentified. The diagnoses listed below may describe real experiences. They may also be incomplete pictures — descriptions of symptoms without the structural cause.

→ Diagnosed with anxiety, depression, or both
→ Diagnosed with ADHD (ADHD and autism frequently co-occur)
→ Told you may have BPD or emotionally unstable personality disorder
→ A history of disordered eating
→ Told your symptoms are stress-related or psychosomatic
→ Sought help repeatedly and felt dismissed, disbelieved, or unseen
→ Responded differently to medications than expected
→ Symptoms that were explained by circumstances rather than neurology

If You Are Considering Seeking a Diagnosis

You do not need a diagnosis to be autistic. But if you are pursuing one, here is how to approach it.

Before your appointment:
Print this and mark items that apply to you. You do not need to check everything — patterns across categories matter more than totals. Bring specific examples, not general statements. “At this event, I did this” is clearer than “I struggle socially.” If you have a journal, old emails, or messages where you describe these experiences — bring them.

At your appointment:
Ask specifically about autism in adult women. Not all clinicians are trained in female presentation. You are allowed to request a referral to a specialist in adult autism assessment.

What this document is — and is not:
→ This is not a diagnostic tool. Checking many items does not confirm a diagnosis.
→ This is not an exhaustive list. Autism presents differently across individuals.
→ This is a starting point — for reflection, for conversation, for self-understanding.
→ The experiences listed here are real, regardless of whether you receive a formal diagnosis.
→ Diagnosis is one path. It is not the only path to knowing yourself.

I planned to go to medical school.

I completed a pre-med track, did what was expected, and followed the path that’s supposed to lead there. But along the way, something became increasingly clear: the structure of medical education wasn’t designed in a way that would support how my mind works—or give me a realistic path to succeed within it.

That realization wasn’t about a lack of ability or ambition. It was about pattern recognition. I could see the demands—timed high-stakes exams, rigid learning formats, sensory-heavy environments, unspoken social expectations—and I could see the mismatch.

So I stepped away.

When people step away from paths like medicine, we tend to label it as dropout, burnout, or a failure of resilience. But often, it’s something else entirely: a design mismatch that was visible long before any breaking point.

And that mismatch doesn’t just affect who becomes a doctor. It shapes how doctors are trained—and ultimately, how patients are understood.

Introduction

Walk into most medical school lectures, and you’ll hear exhaustive detail on cardiac physiology, pharmacology, and diagnostic algorithms. Ask those same students about neurodiversity—autism, ADHD, dyslexia, dyspraxia, and the full spectrum of cognitive variation—and you’re likely to get blank stares, outdated stereotypes, or overly clinical definitions detached from lived reality.

Despite growing societal awareness, medical education has not kept pace. Neurodiversity remains largely absent from formal curricula—not just in how physicians are trained to treat patients, but in how institutions support their own trainees.

This is no longer just an inclusion issue. It’s a patient safety issue, a workforce sustainability issue, and increasingly, a legal and ethical one.

A Curriculum Stuck in the Past

Medical training still relies on rigid, standardized structures: timed high-stakes exams, dense didactic lectures, overstimulating clinical environments, and unspoken expectations around social performance.

These systems assume a narrow cognitive norm.

They rarely account for sensory processing differences, executive functioning variability, or communication differences that are common across neurodivergent profiles.

At the same time, neurodiversity—when taught at all—is often confined to pathology-based frameworks. Students learn diagnostic criteria, but not how those traits actually show up in real clinical interactions.

A growing body of literature reflects this gap: minimal structured curriculum content, lack of evidence-based teaching strategies, and overreliance on reactive accommodations instead of inclusive design.

Medical education is still teaching about conditions—not people.

What’s Missing (And Why It Matters)

The consequences of this gap show up in clinical care every day.

Diagnostic blind spots persist, especially across gender and race, where neurodivergence presents differently and is often missed or misattributed.

Communication breakdowns are common. Directness may be misread as rudeness, reduced eye contact as disengagement, and processing delays as lack of understanding.

Sensory-inaccessible environments—bright lights, noise, rushed interactions—create unnecessary distress, yet few clinicians are trained to adapt them.

And critically, strengths are ignored. Neurodivergence is still framed almost entirely in terms of deficits, despite clear evidence of strengths like pattern recognition, deep focus, and systems thinking—traits that are highly valuable in medicine.

The Hidden Cost to Healthcare

This gap affects more than individual experiences—it shapes outcomes at scale.

Patients face higher rates of misdiagnosis, medical trauma, and avoidance of care. When communication fails, so does accurate assessment.

Neurodivergent medical students and residents often mask to survive training, leading to burnout and attrition. Medicine loses capable, insightful individuals before they ever reach practice.

Institutions also face growing legal and compliance risks under disability frameworks like the ADA when accessible design is not prioritized.

And at its core, this is a patient safety issue. Misinterpretation, missed context, and outdated assumptions directly affect quality of care.

Why the Gap Persists

The problem is structural.

Medical education prioritizes standardization and objectivity, often at the expense of cognitive diversity. The deficit-based medical model still dominates. Faculty are rarely trained in neurodiversity themselves. Accrediting bodies largely remain silent on the issue. And persistent myths about “lowering standards” continue to block progress.

The result is a self-perpetuating cycle: each generation of physicians inherits the same blind spots.

What Change Actually Looks Like

Progress doesn’t require a full overhaul—it requires integration.

Neurodiversity should be embedded across curricula, not isolated in a single lecture. Communication training should include real-world neurodivergent interaction. Sensory-aware care should be normalized. Universal Design for Learning should replace one-size-fits-all teaching.

Admissions and evaluation should reflect a broader understanding of competence—not just speed, memorization, or performance under pressure.

And most importantly, neurodivergent voices should be included as educators, not just subjects.

A Cultural Shift, Not Just a Curriculum Update

An estimated 15–20% of the population is neurodivergent. That means every physician is already treating neurodivergent patients—whether they recognize it or not.

Neurodiversity is not a niche topic. It is fundamental to modern, competent care.

The question is no longer whether this belongs in medical education.

It’s why it isn’t already there.

A Call to Action

This isn’t a theoretical gap—it’s already shaping who becomes a doctor and how patients are treated.

If you’re a medical student or physician, your perspective matters. Where have you seen neurodiversity misunderstood, dismissed, or unsupported in training or practice?

Share your experiences in the comments. The patterns won’t change unless they’re made visible.

If you believe medical education should reflect the full diversity of human cognition, advocate for change within your institutions. Push for curriculum reform, faculty training, and inclusive design.

And if you’re ready to take it a step further, support initiatives and petitions calling for the integration of neurodiversity into medical education standards. Systemic change happens when enough voices demand it.

References

• American Medical Association. (2021). Advancing health equity: Guide on language, narrative and concepts.

• Bolte, S. (2021). The power of neurodiversity. Journal of Child Psychology and Psychiatry.

• Botha, M., et al. (2021). Language in autism research. Autism.

• Crane, L., et al. (2019). Mental health experiences of autistic adults. Autism.

• Gordon, L., et al. (2022). Neurodiversity in medical education: A scoping review. Medical Education.

• Mahoney, W. J., et al. (2023). Supporting neurodivergent medical learners. Academic Medicine.

• Nicolaidis, C. (2012). Neurodiversity and medicine. AMA Journal of Ethics.

• Pellicano, E., & den Houting, J. (2022). Neurodiversity in autism science. JCPP.

• Stanford Neurodiversity Project. Stanford University.

• U.S. Department of Justice. ADA guidance.

• World Health Organization. (2022). Inclusive healthcare systems.

Editor’s note: This article is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare provider regarding your individual circumstances.

This essay is part of the Institutional Legitimacy series, which examines the structural conditions that produce systematic friction for people whose biology, cognition, or communication doesn’t match the prototype a system was built around. The focus here is medicine — specifically, what happens when the evidence base for treatment is built on a sample that excludes the people being treated.

I am autistic, diagnosed in adulthood. I write this series from a combination of lived experience and systems-level analysis. I know what it is to sit in a clinical room and watch a practitioner reach for the nearest available label — anxious, inconsistent, difficult to read — because the tool they were trained with was not built for the body in front of them.

That experience is not unusual. It is, as this essay will show, the statistically expected output of a system operating exactly as designed.

Medicine presents itself as a precision system.

It runs on evidence. Randomized controlled trials. Peer-reviewed data. Dosing protocols calibrated to body weight, age, metabolic function. It has reference ranges for what is normal, thresholds for what is pathological, and standard interventions for everything in between.

The problem is not the precision. The problem is the sample.

Every reference range has a population behind it. Every trial has inclusion and exclusion criteria. Every dosing chart was calibrated on someone. And in medicine, as in hiring, as in education, as in every institutional system examined in this series — the someone the system was built around is narrow. Specific. A prototype.

When your biology matches the prototype, the system reads you clearly. When it doesn’t, the system doesn’t upgrade its reading. It labels you.

What it labels you is: complex. Difficult. Anxious. Non-compliant. A poor responder.

This essay is about what that actually means — and what it costs.

Who Gets Left Out Before the Trial Starts

Clinical trials are designed to reduce variability.

That is their stated purpose. Reducing variability makes it easier to isolate the effect of a drug, measure an outcome, produce a clean result. The logic is sound. The consequence is not.

Reducing variability means excluding the people who introduce it. And the people who introduce the most variability into a medical trial are the people whose biology, cognition, communication, or history doesn’t conform to the expected pattern.

That means: people with psychiatric diagnoses. People on multiple medications. People with complex comorbidities. People whose symptom presentation doesn’t follow the textbook. People whose communication style is atypical. People whose bodies process stress, sensation, and medication differently.

That means, systematically: neurodivergent people.

In a recent analysis of 46 clinical trial protocols, 78.3% included exclusion criteria targeting people with psychiatric or neurological conditions. The most common justification was “compliance concerns.” Nearly 30% offered no justification at all.

In 129 behavioral clinical trial articles, health conditions were excluded 697 times. One hundred and fifteen of those exclusions had no stated reason.

“Compliance concerns” is worth examining as a phrase. It sounds like a safety rationale. What it actually describes is a design verdict: this person doesn’t fit the protocol. They are not excluded because the drug is unsafe for them. They are excluded because the trial was built for someone else, and this person doesn’t match.

The exclusion is then incorporated into the evidence base. The evidence base shapes the prescribing guidelines. The prescribing guidelines are applied universally — including to the people who were excluded. The cycle is self-sealing. The gap never closes because the system never acknowledges that the gap exists.

This is not a fringe issue or a subset of poorly designed trials. These exclusion patterns are standard practice across clinical research, regulatory approval, and guideline development. The system is producing exactly what its structure predicts. The structure predicts exclusion.

The Diagnostic Manual Did It Too

The exclusion wasn’t only happening at the trial level.

Until 2013 — when the DSM-5 replaced DSM-IV — a diagnosis of autism was an explicit exclusion criterion for ADHD. You could not hold both diagnoses simultaneously. The diagnostic framework made co-occurrence definitionally impossible.

This matters because ADHD medication trials were, by definition, conducted on populations from which autistic people were barred. The evidence base that defines standard ADHD treatment was built without the group that represents an estimated 50–70% of autistic adults.

This wasn’t a gap someone missed. It was written into the manual.

When the DSM-5 finally allowed co-diagnosis, researchers noted that many studies performed under DSM-IV had explicitly excluded these patients — and that this created a knowledge gap that is only now beginning to be addressed.

The drugs had already been prescribed for decades. The guidelines had already been written. The gap was inherited by every clinician trained on that evidence, every formulary built from it, every prescription written since.

The Approval Was Thin Before You Even Get to Neurodivergence

Before examining what happens when neurodivergent people take these medications, it is worth establishing what the evidence base actually consists of.

The median number of participants studied per ADHD drug before FDA approval was 75. Fifty-five percent of ADHD drugs were approved after fewer than 100 people were studied. Seventy percent were approved after fewer than 300. The median trial length before approval was four weeks. Six drugs were approved with requests for specific post-marketing studies. Two of those studies were actually completed.

Seventy-five people. Four weeks. A standard derived primarily from male, neurotypical, childhood presentations, applied to adults, to women, to people with complex neurodevelopmental profiles — without stratification, without adjustment, without acknowledgment that these are different bodies responding to a protocol that was never calibrated for them.

Over half — 53% — of real-world ADHD patients would be ineligible for the medication trials their prescriptions are based on, due to strict exclusion criteria. For adults, that figure reaches 74%.

The prescription is evidence-based. The evidence was built on someone else’s body.

What Happens When the Body Isn’t the Body in the Trial

The largest controlled trial of methylphenidate in autistic children was conducted by the Research Units on Pediatric Psychopharmacology — the RUPP Autism Network.

In this trial, methylphenidate showed a positive effect in 49% of autistic participants. The trial was discontinued in 18% of subjects due to side effects — primarily irritability, increased anxiety, and behavioral activation.

The established response rate for methylphenidate in neurotypical ADHD is 70–80%. In the NIMH’s large-scale ADHD treatment study of neurotypical youth, nearly three-quarters of participants were successfully maintained on methylphenidate with a 4% dropout rate.

Same drug. Same diagnostic label. One population studied for decades, dosed to a calibrated protocol, with a 70–80% response rate and 4% discontinuation. Another population studied in a single major trial, with a 49% response rate and 18% discontinuation.

That is not a non-response. That is not a complex patient. That is the predictable consequence of applying a protocol to a body it was never built for — and then having no framework, when it fails, except to call the body the problem.

The reason the numbers split this way is not random variation. It is biology. Neurodivergent people have systematically different medication responses because their biology is different from the neurotypical baseline used to design and evaluate those medications. Heart rate and autonomic tone operate at different set points. Gastrointestinal function alters absorption. Neurochemistry presents a different starting profile for drugs to act on. These are not individual anomalies. They are population-level, measurable, reproducible differences — and they were not in the sample.

The picture is not better for the second major class of ADHD medication. Amphetamine-class compounds — Adderall, Vyvanse, Dexedrine — are considered a co-equal first-line treatment alongside methylphenidate. In autistic populations, their use has been actively discouraged by researchers, not because the drug has been shown to be harmful, but because the data does not exist to say what it does. No adequately powered safety and efficacy trial has been conducted. The drug is still prescribed. The evidence was never gathered.

There is an additional layer to the side-effect data that is rarely reported. The figures above come from trials that pre-screened participants. Researchers selected autistic individuals considered most likely to tolerate the study conditions before a single dose was given. The published numbers are already the best-case scenario. The real-world rates are higher.

The Biology Is Not a Footnote

The biology is systematically different in ways that affect regulation, signaling, and drug response. These are not edge-case differences. They are baseline shifts in how the body regulates, signals, and responds — applied to a system calibrated for a different baseline. This is not a gap in knowledge. It is a measurement error built into the system itself.

Autistic adults show resting heart rates approximately 10–15 beats per minute higher, and heart rate variability approximately 30–40% lower, than matched neurotypical controls. Heart rate variability — the variation in time between heartbeats — is a standard clinical marker of autonomic function. Lower HRV indicates reduced parasympathetic (”rest and digest”) activity and elevated sympathetic (”fight or flight”) tone.

In standard medicine, reduced HRV is a marker of pathology. When a neurodivergent patient presents with reduced HRV, the clinical record reads: elevated sympathetic tone. Possible anxiety. Possible cardiovascular risk.

What the clinical record does not record is that this is that patient’s regulated baseline. The reference range was built from neurotypical samples. The threshold for “normal” was set without this body in the room. The instrument isn’t broken. The baseline is wrong.

Research shows that autism and ADHD are associated with opposite autonomic arousal profiles — autistic individuals tend toward hyper-arousal under cognitive demand, while people with ADHD tend toward hypo-arousal at rest. These are not variations on the same nervous system. They are distinct regulatory patterns. Drugs that work by modulating arousal — stimulants, SSRIs, norepinephrine reuptake inhibitors — are acting on different biological starting points when they enter a neurodivergent body.

The SSRI data makes the same point from a different angle. In placebo-controlled trials of SSRIs in autistic children, activating side effects appeared at rates that dwarf what the same drugs produce in neurotypical populations: disinhibition at 19.2% versus 6.6%, increased inattention and hyperactivity at 12.3% versus 2.6%, gastrointestinal effects at 26% versus 11.8%. These are not rare adverse events. They are systematic, reproducible patterns of a drug acting on a biology it was not tested on.

Interoception adds another layer. Autistic individuals can experience allodynia — extreme sensitivity to normally non-painful touch — alongside hyposensitivity to pain from injury or illness, sometimes simultaneously. These differences trace to neurological variation at the level of unmyelinated small fiber density and thalamocortical signaling. They are not psychological. They are structural.

The clinical consequence: pain and distress in autistic patients often does not present the way clinical intake instruments expect. It may arrive as behavioral change — agitation, withdrawal, increased stimming — that a clinician reads as emotional dysregulation rather than physical distress. GI symptoms, which affect 70–90% of autistic individuals and are more severe than in neurotypical populations, are particularly subject to this misreading. Distress expressed through behavior is charted as a psychiatric concern. The physical cause goes uninvestigated.

This is how the system produces what it calls “complex patients.” The patient is not complex. The signal is valid. The system was not built to receive it.

The neurotransmitter picture reinforces this. ADHD is associated with higher norepinephrine and lower serotonin relative to neurotypical norms. Autism often shows the reverse. A drug that recalibrates these systems in a neurotypical direction is not moving both bodies to the same place. It is moving them from different starting positions, by the same amount, toward different destinations.

Approximately 35% of autistic people take at least one psychotropic medication. Most of those medications are oral formulations whose absorption depends on consistent gastrointestinal transit. Between 70–90% of autistic individuals report gut dysfunction — constipation, irregular motility, altered gastric emptying. When transit is irregular, blood-level peaks are irregular. The drug does not behave as the dosing chart predicts. Formal pharmacokinetic studies adjusted for autistic GI physiology are rare. Dosing is extrapolated from neurotypical data.

The enzyme-level problem adds a final layer. Atomoxetine is primarily metabolized by the CYP2D6 enzyme — one of the most genetically variable enzymes in human metabolism. Ultra-rapid metabolizers clear the drug before it reaches therapeutic concentration. Poor metabolizers accumulate it and face higher risk of adverse events. Neurodivergent populations show broader heterogeneity in this enzyme activity. Standard dosing does not account for it.

There are currently no reliable biomarkers to predict how any individual will respond to long-acting stimulants. Dosing is calibrated through behavioral observation — which depends entirely on the clinician’s ability to accurately read atypical presentation. When a neurodivergent patient’s presentation is itself atypical, behavioral observation is not a reliable instrument. The system has no other instrument to offer.

Beyond Neurodivergence

This pattern is not unique to neurodivergence.

It appears in autoimmune disease, chronic illness, and other conditions defined by high diagnostic delay, dismissal, and frequent misattribution to anxiety — particularly in women. Endometriosis. Lupus. Ehlers-Danlos Syndrome. Conditions where the symptom presentation departs from the textbook, where the patient population is predominantly female, where the standard diagnostic model was not built to read the signal in front of it.

The common factor is not the condition. It is the mismatch between the patient’s signal and the system’s model.

The autoimmune overlap with neurodivergence compounds this further. Research shows that maternal autoimmune disease is associated with a meaningfully higher risk of ADHD in offspring — hazard ratio approximately 1.30, higher for specific conditions including type 1 diabetes. Autoimmune conditions share a patient population with ADHD and autism at rates above what chance would predict. When a neurodivergent patient presents with overlapping autoimmune symptoms, the system is not equipped to read multiple atypical signal patterns simultaneously. It routes each one separately to the nearest available category.

The Prototype Is Also in the Diagnostic Criteria

The medication problem does not begin at the prescription pad. It begins at diagnosis, where the same prototype logic operates.

Autistic women are diagnosed on average seven years later than autistic men. One large analysis found males had a 38% decreased chance of late diagnosis compared to females. They are three times more likely to receive a mental health misdiagnosis before receiving an autism diagnosis. Nearly 80% report a prior misdiagnosis — most commonly anxiety, borderline personality disorder, eating disorders, or mood disorders.

ADHD diagnosis in women typically arrives in the late 30s to early 40s. For boys, the average age of diagnosis is around 7.

This is not diagnostic error in the conventional sense. It is prototype fidelity. The diagnostic tools for ADHD and autism were historically built on male, externalizing, childhood presentations — hyperactivity, disruption, visible non-compliance. Female presentations — internalizing, camouflaged, socially exhausted, apparently functional — do not trigger the pattern-match. The system routes them to anxiety, to personality disorder, to the category of “she’s just a lot.”

The DSM-5 still carries documented gender bias. This is not a contested claim. It is a finding in the research literature. The reference instrument encodes prototype assumptions at the definitional level.

A 2022 UK government study found that 84% of women felt NHS medical professionals were not listening to them. Thirty-one percent had waited over ten years for a diagnosis.

Ten years of waiting is not a diagnostic backlog. It is a prototype failing to recognize a signal it was never designed to read.

What This Looks Like in the Chart

The language medicine uses to describe these failures is precise. It is also precisely wrong.

When a neurodivergent patient’s symptom presentation doesn’t match the expected pattern, it is described as “atypical presentation.” The word atypical places the variance in the patient. It does not describe what atypical means relative to: the presentation the diagnostic criteria were built around, the population that was included in the trial, the sample that defined the reference range.

When a neurodivergent patient has difficulty reporting symptoms in the expected format — verbal, linear, scaled from 1 to 10 — it is charted as “poor historian.” Poor historian describes a communicative failure. It does not describe a mismatch between the intake instrument and the patient’s interoceptive and communicative profile.

When a neurodivergent patient’s elevated heart rate or atypical HRV registers during a clinical encounter, it is recorded as “anxious.” Anxious attributes the reading to a psychological state. It does not note that the patient’s autonomic baseline was never accounted for in the reference range being applied.

When a drug doesn’t work at the standard dose, the patient is a “non-responder.” Non-responder locates the failure in the patient’s body. It does not ask whether the dosing protocol was ever validated in a body like this one.

When a neurodivergent patient appears in a trial protocol’s exclusion criteria, the rationale is “compliance concerns.” Compliance concerns sounds like safety reasoning. It is a design verdict about whose variability the trial was built to accommodate.

These terms are not neutral descriptions. They are the vocabulary of a system that has one explanation available when a patient doesn’t fit: the patient.

What the Labels Actually Do

The consequence of being labeled is not just administrative. It shapes what happens next.

When a neurodivergent patient is charted as a poor responder, the clinical record carries that forward. The next clinician inherits it. The next prescribing decision is made in its shadow. The patient often internalizes it. They have been told, in the language of medicine, that the system’s tools work — and that they are the reason those tools didn’t. That is a precise and lasting form of harm.

Many patients describe this as structural gaslighting: not deliberate manipulation, but a system that presents its own instrument as neutral and objective, attributes the misread to the patient, and offers no mechanism for the patient to demonstrate that the instrument was miscalibrated. The patient cannot argue with a reference range. They cannot appeal a dosing protocol. They are told the evidence says the drug works. They are not told who the evidence was built on.

What follows is often a prescribing cascade. When a first-line medication fails on a body it wasn’t designed for, the clinical response is rarely to question whether that body was ever in the trial. It is to add another drug. Then another. Each addition attempts to compensate for the failure of the previous one — without naming the upstream reason any of them failed. Neurodivergent patients, who tend toward higher rates of comorbidities and greater sensitivity to side effects, are disproportionately exposed to the cumulative risks of polypharmacy as a result. The complexity the system attributed to the patient was produced, in part, by the system’s own sequential interventions on a biology it misread from the start.

This has direct clinical consequences: side effects misread as psychiatric symptoms, ineffective treatments prolonged because the failure is attributed to the patient rather than the protocol, patients cycled through successive interventions that were never calibrated for them — often for years, sometimes for decades. The harm is documented in the discontinuation rates, the diagnostic delays, the prescribing cascades, and the records of patients labeled complex long before anyone asked whether the measurement model fit.

Why the System Stays This Way

It is worth being precise about why this pattern persists. The clinical research system is not uniformly malicious. It is optimized — and what it is optimized for is not representativeness.

Clinical trials are expensive. Large, heterogeneous samples cost more, take longer, and produce results that are harder to bring to statistical significance and harder to publish. Regulatory frameworks reward clean, fast, reproducible outcomes. Academic incentives favor controlled conditions and narrow populations. Drug approval timelines create pressure to reach a finding quickly on the smallest sample that will support the claim.

Small samples, short trials, and homogeneous participants are not oversights. They are the rational output of a system whose incentives are structured around efficiency and fundability rather than generalizability. The narrow prototype is what “clean results” looks like when you need them fast.

This does not make the consequences acceptable. It makes them predictable. And predictability is what the Institutional Legitimacy framework is built to examine — not whether individual actors are malicious, but whether the structure produces harm reliably, and whether that harm falls on the same people every time.

What Correction Would Require

The path forward is not adding neurodivergent people to existing trial designs and running the same protocol. That would produce better demographic representation inside a measurement model still built for the prototype. Better headcount is not a corrected baseline.

Correction requires something prior to enrollment. Trials would need to stratify by neurotype and account for known biological differences before they begin — adjusting vital sign thresholds, accounting for GI variability in pharmacokinetic modeling, acknowledging that “standard dose” is a calibration made on a specific sample and not a biological universal. The absence of safety data in a population would need to be treated as a prescribing caution, not a gap to quietly inherit.

The DSM lesson needs to be absorbed as a standing principle: when a diagnostic framework excludes a population by definition and drugs are approved in that window, the resulting evidence gap does not resolve on its own when the manual changes. It requires active remediation — funded trials, stratified re-analysis, explicit acknowledgment in prescribing guidelines that existing evidence does not cover this population.

Real-world evidence — data collected from actual clinical practice rather than controlled trials — offers one mechanism for correction. It captures populations that trials exclude, documents outcomes in the bodies actually receiving treatment, and surfaces the patterns that small, homogeneous samples miss. It is not a substitute for well-designed prospective trials. It is a correction for the ones that were never designed well to begin with.

The deeper correction is institutional. It requires medicine to develop the language to say, inside clinical records and prescribing guidelines and training programs: the baseline was wrong. The sample was incomplete. The tool was built for a different body. The evidence is not evidence-based for this population.

A design verdict can be changed. But only after it has been named.

The Institutional Legitimacy Frame

This series has argued that institutions are not neutral structures. They are built around a prototype — a model of the person they were designed to serve, measure, and accommodate. The prototype is usually not named. It is embedded in the tools: the diagnostic criteria, the trial design, the reference range, the intake form, the dosing chart.

When a person matches the prototype, the institution reads them clearly. It produces an accurate diagnosis, an appropriate treatment, a legible record. The system works.

When a person doesn’t match the prototype, the system doesn’t revise its reading. It routes the person into a secondary category — complex, atypical, non-compliant, poor responder — and records the mismatch as a property of the person.

This is what Institutional Design Mismatch looks like in medicine. The system was not designed for every body. It was designed for a specific body, and the design was then treated as universal. When it fails — when the drug doesn’t work, when the diagnosis takes seven years, when the vital signs register as pathological — the system produces an explanation that protects the design and locates the failure in the patient.

The patient is told they are complex. What complex means is: the measurement model wasn’t built for them.

The patient is told the medication doesn’t work for them. What that means is: the trial was conducted on someone else.

The patient is told their presentation is atypical. What atypical means is: the prototype didn’t include them.

This is not a clinical judgment. It is a design verdict. And the difference matters — because a design verdict can be changed, but only if it is named.

What Would Have to Be True

If a population was systematically excluded from the data used to define a treatment, in what sense is that treatment evidence-based for them?

That question is not rhetorical. It has a direct answer: it isn’t. And the clinical, regulatory, and prescribing systems that rest on that evidence have inherited that gap without naming it.

If the standard of care for ADHD was derived from trials that excluded autistic people, and 50–70% of autistic adults also have ADHD, then the standard of care for ADHD was derived from a sample that excluded the majority of the people currently receiving it.

If the diagnostic criteria for autism and ADHD were built on male, externalizing, childhood presentations, then the 80% of autistic women who receive a prior misdiagnosis are not failures of clinical judgment. They are the expected output of a system running a pattern-match against the wrong prototype.

If the reference ranges for heart rate and heart rate variability were calibrated on neurotypical samples, then applying those ranges to autistic patients does not produce a diagnosis. It produces noise — which the system then attributes to the patient.

These are not inflammatory claims. They are the logical implications of the data that already exists. The research does not hide them. They emerge plainly from trial results, exclusion criteria, approval records, and diagnostic delay statistics.

What is missing is the institutional language to say: the baseline was wrong. The sample was incomplete. The tool was built for a different body. The evidence is not evidence-based for this population.

Until that language exists inside clinical practice — not just in research footnotes, not just in advocacy — the system will continue to produce the same output. The patient will be complex. The medication will not work. The vital signs will be abnormal. The presentation will be atypical.

And the design will remain unchanged.

When the same experiences are repeatedly dismissed across different people and contexts, that is not randomness. That is structure.

Works Cited

Clinical Trial Exclusion — Psychiatric and Neurological Conditions Exploring Barriers to Inclusivity: Systematic Analysis of Exclusion Criteria and Potential Bias in Clinical Cancer Trials for Psychiatric and Neurological Conditions in European Protocols. PMC/NCBI, 2024. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12068421/

Clinical Trial Exclusion — Disability and Behavioral Trials Underrepresentation of Adults and Older Adults With Disabilities in Behavioral Clinical Trials: A Scoping Review. PMC/NCBI, 2020. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7742415/

DSM Co-Diagnosis Exclusion / ADHD–ASD Knowledge Gap Effectiveness of Pharmacological Interventions for Managing ADHD Symptoms in Individuals with Autism Spectrum Disorder: A Systematic Review and Meta-Analysis. ScienceDirect, 2024. https://www.sciencedirect.com/science/article/abs/pii/S027858462400157X

DSM Co-Diagnosis and ASD/ADHD Comorbidity Management An Update on the Comorbidity of ADHD and ASD and Its Clinical Management. Expert Review of Neurotherapeutics, 2025. https://www.tandfonline.com/doi/full/10.1080/14737175.2025.2599856

FDA Approval Sample Sizes — ADHD Medications Premarket Safety and Efficacy Studies for ADHD Medications in Children. PMC/NCBI, 2014. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4090185/

ADHD Trial Exclusion Rates — Real-World Patients (53%; 74% adults) Children and Adolescents with Co-Occurring ADHD and ASD: A Systematic Review of Multimodal Interventions. PMC/NCBI, 2025. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12156790/

RUPP Autism Network — Methylphenidate Response Rate (49%) and Discontinuation (18%) Stimulant Medication Treatment of Target Behaviors in Children with Autism: A Population-Based Study. PMC/NCBI, 2008. https://pmc.ncbi.nlm.nih.gov/articles/PMC2585372/

RUPP Methylphenidate Response — Genetic Moderators Positive Effects of Methylphenidate on Hyperactivity Are Moderated by Monoaminergic Gene Variants in Children with ASD. The Pharmacogenomics Journal, 2014. https://www.nature.com/articles/tpj201323

Neurotypical ADHD Methylphenidate Dropout Rate (4%) — NIMH MTA Study Eye Tracking as a Predictor of Methylphenidate Response in Autism with ADHD. ClinicalTrials.gov Protocol Document, NCT02874690. https://cdn.clinicaltrials.gov/large-docs/90/NCT02874690/Prot_SAP_000.pdf

Methylphenidate in ASD — Long-Term Naturalistic Study; Dosing at 0.3 mg/kg/day Methylphenidate in Autism Spectrum Disorder: A Long-Term Follow Up Naturalistic Study. PMC/NCBI, 2020. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7463802/

Amphetamine Data Gap in Autistic Populations Evaluation of the Effect and Safety of Lisdexamfetamine in Children Aged 6–12 With ADHD and Autism. ClinicalTrials.gov, NCT03337646. https://clinicaltrials.gov/study/NCT03337646

ADHD Medication Response — ASD Comorbidity, Discontinuation Rates The Effect of Autistic Traits on Response to and Side-Effects of Pharmacological ADHD Treatment. Journal of Neurodevelopmental Disorders, 2022. https://link.springer.com/article/10.1186/s11689-022-09424-2

Pre-Screening Bias in AuDHD Stimulant Trials AuDHD and Stimulant Medication. Embrace Autism, 2024. https://embrace-autism.com/audhd-and-stimulant-medication/

Swedish Registry — Methylphenidate Dosing Patterns, ADHD+ASD vs. ADHD Only Medications for ADHD in Individuals with or without Coexisting ASD: Swedish Prescribed Drug Register Analysis. PMC/NCBI, 2021. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7758935/

No Reliable Biomarkers for Stimulant Response; Pharmacokinetic Variability An Update on the Pharmacokinetic Considerations in ADHD Treatment with Long-Acting Stimulant Formulations. Expert Opinion on Drug Metabolism and Toxicology, 2019. https://www.tandfonline.com/doi/full/10.1080/17425255.2019.1675636

Atomoxetine — CYP2D6 Enzyme Variability and Pharmacokinetics Growing Evidence of Pharmacotherapy Effectiveness in Managing ADHD in Young Children With or Without ASD. Frontiers in Psychiatry, 2024. https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2024.1408876/full

Heart Rate Variability — Autism and ADHD Opposite Arousal Profiles Heart Rate Variability in Children and Adolescents with Autism, ADHD and Co-occurring Autism and ADHD. PMC/NCBI, 2022. https://pmc.ncbi.nlm.nih.gov/articles/PMC9556357/

HRV — Elevated Resting Heart Rate in Autistic Individuals Indices of Heart Rate Variability and Performance During a Response-Conflict Task in ADHD and Autism. PMC/NCBI, 2021. https://pmc.ncbi.nlm.nih.gov/articles/PMC8785294/

HRV — Reduced HRV in Autistic Children; Autonomic Dysfunction Marker Heart Rate Variability in Children with Autism Spectrum Disorder and Associations with Medication and Symptom Severity. PubMed/NCBI, 2020. https://pubmed.ncbi.nlm.nih.gov/33225622/

Interoception — Atypical Pain Perception, Allodynia, Structural Neurological Basis Interoception in Autism: A Narrative Review of Behavioral and Neurobiological Findings. Dove Medical Press / Psychology Research and Behavior Management, 2024. https://www.dovepress.com/interoception-in-autism-a-narrative-review-of-behavioral-and-neurobiol-peer-reviewed-fulltext-article-PRBM

Pain in Autism — Paradox of Hyper- and Hyposensitivity The Current View on the Paradox of Pain in Autism Spectrum Disorders. PMC/NCBI, 2022. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9352888/

GI Dysfunction Prevalence in Autism (70–90%) and Interoception in Assessment Prevalence of Gastrointestinal Symptoms among Autistic Individuals. PMC/NCBI, 2023. https://pmc.ncbi.nlm.nih.gov/articles/PMC10527131/

Gut Microbiome Differences in Autism Comparing the Gut Microbiome in Autism and Preclinical Models: A Systematic Review. PMC/NCBI, 2022. https://pmc.ncbi.nlm.nih.gov/articles/PMC9286068/

Psychotropic Medication Use in Autism (35%); GI Absorption Concerns How Autism’s Gut Problems May Be Linked to Brain and Behavior. Science Daily, 2014. https://www.sciencedaily.com/releases/2014/09/140902114731.htm

Diagnostic Delay — Autistic Women (7-Year Gap; OR 0.62; 3x Misdiagnosis Rate) The Hidden Face of Neurodiversity: Why Women Are Overlooked and Misunderstood. The Psychology Group, 2026. https://www.tpgroup.co.nz/blog/hidden-face-of-neurodiversity

DSM-5 Gender Bias — Diagnostic Tools Built for Male Presentations IWD2022: Neurodivergent Women and Girls Misdiagnosed, Misunderstood and Missed Altogether. Creased Puddle, 2022. https://www.creasedpuddle.co.uk/iwd2022-neurodivergent-women-girls-misdiagnosed-misunderstood-and-missed-altogether/

UK Government Study — 84% of Women Not Listened To; 31% Waited 10+ Years When Doctors Don’t Listen: How Neurodivergent Women Are Falling Through the Cracks. The New Feminist, 2025. https://thenewfeminist.co.uk/2025/07/when-doctors-dont-listen-how-neurodivergent-women-are-fallingthrough-the-cracks/

Medical Gaslighting — Neurodivergent Patients, Misdiagnosis Patterns Medical Gaslighting and Neurodivergence: Why Autistic and ADHD Patients Aren’t Believed. Fish in a Tree NOLA, 2025. https://www.fishinatreenola.org/blog/dup-test-8ezrr-eer8f-r8s98-sgdm5-23826-8l6d8-nk54f-hw7pa-57bga-2n578-3ndfk-5b2xf-c8lg4-7d9er-27rjd-bfecx-4fkwj-8zwkx-5cz3k-afxcx-38a5x-xddkz-dt3mn-zkxec-df9ty-bn7e4-8zfgk-mh8tw

Autoimmune Comorbidity with ADHD — Hazard Ratio 1.30 Long-Term Medication for ADHD (LMA) Trial: 2-Year Prospective Observational Study. PMC/NCBI, 2024. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11127865/

Neurodiversity Prevalence Statistics Neurodiversity Statistics and Neurodivergent Research Data. Neurodiversity Directory, 2026. https://neurodiversity.directory/neurodiversity-statistics/

Editor’s note: This essay applies the Institutional Design Mismatch lens to ADHD — examining how diagnostic criteria built around a specific prototype have shaped who gets identified, when, and what story they’re told in the meantime.

The diagnostic criteria for ADHD were built from field trials in which only 21% of participants were girls. Then we spent decades wondering why women kept getting missed.

That is not an oversight. That is a design decision with consequences.

The system measured what it was built to measure. It found what it was built to find. And everyone who fell outside that prototype was handed a different verdict: anxious, sensitive, dramatic, lazy. Try harder. Pay attention.

There is a version of this story you already know.

Can’t focus. Can’t follow through. Can’t sit still. A brain that fails to do the basic things adults are supposed to do without being asked twice.

That story has been told millions of times. It has been diagnosed, medicated, and managed. It has become so familiar that most people accept it as fact.

It isn’t.

What the research actually says

ADHD is officially defined as a deficit in attention. But that framing contains a hidden assumption: that the standard environment’s demand for sustained, volitional, externally-regulated attention is a neutral baseline, and any brain that can’t meet it on demand is broken.

That assumption does not survive contact with the evidence.

Dr. William Dodson — a board-certified psychiatrist and one of the first clinicians to specialize in adult ADHD — describes the hallmark of the ADHD nervous system not as an attention deficit, but as inconsistent attention: a brain that works well, using its own set of rules.

He coined the concept of the interest-based nervous system. The challenge for most people with ADHD, he notes, lies not in performing tasks but in getting motivated and engaged. Once engaged, many report they can do almost anything.

That is not a deficit. That is a different attentional architecture.

A 2025 study published in European Psychiatry found that 68% of adults with ADHD reported frequent hyperfocus episodes — states of intense, prolonged concentration lasting from several hours to days. The same nervous system that struggles to initiate a low-stakes administrative task can produce extraordinary focus under the right conditions.

The condition is not the problem. The conditions are.

ADHD is not a failure of attention. It’s a failure of environments to engage it.

This is an Institutional Design Mismatch

The systems we built — schools, offices, bureaucratic processes — were designed around a specific attentional prototype. Someone who can initiate independently. Sustain focus across low-stimulation periods. Self-regulate without external structure or environmental signal.

When you don’t match that prototype, the system produces a verdict: disordered, lazy, unreliable.

The Institutional Legitimacy framework asks a different question: what human prototype was this environment built for, and what happens when someone falls outside it?

For ADHD, the answer is clear. The environment was not built for interest-responsive attention. When that attention shows up, it gets called a disorder. The friction is real. But it is not located in the individual.

The prototype was always a boy

The design mismatch runs deeper than attention systems. It runs through the diagnostic criteria themselves.

In the field trials used to establish the diagnostic criteria for ADHD in both the DSM-5 and ICD-11, only 21% of participants were girls.

The result is a diagnostic instrument built primarily on how ADHD presents in boys — externalizing, hyperactive, visible. Girls and women tend toward inattentive presentations: quieter, more internalized, more easily read as something else.

According to CDC data, 15% of boys between ages 3 and 17 have been diagnosed with ADHD, compared to just 8% of girls — even though research now shows girls experience ADHD at comparable rates, with symptoms that tend to look different.

A Swedish population register study of over 85,000 individuals with ADHD found that women received their first ADHD diagnosis nearly four years later on average than men — and that prior to diagnosis, women with ADHD were roughly twice as likely to have been diagnosed with both anxiety disorders (50% vs. 26%) and mood disorders (38% vs. 20%) compared to their male counterparts.

That gap is not biology. It is not because girls are less affected. It is because the measurement tool was built for a different prototype. When you don’t match the prototype the system was designed to recognize, you don’t get diagnosed. You get misdiagnosed with anxiety. You get told you’re sensitive. You get told to try harder.

The delay is not random. It is structural.

Masking is not coping. It is labor.

Many people with ADHD — especially women and girls — develop elaborate strategies to appear neurotypical in environments that demand it. This is called masking. And it is routinely mistaken for evidence that there is no problem.

The system calls it inconsistency. The nervous system calls it context.

ADHD has been called the “moral diagnosis” — because its symptoms are so consistently read as personal failings rather than the result of a neurodevelopmental condition. The misattribution is so complete that many people internalize it long before they ever receive a diagnosis.

What looks like laziness is often executive dysfunction, stress-based motivation patterns, or energy depletion from constantly compensating in a neurotypical world.

The institution sees someone who occasionally underperforms and draws the obvious conclusion: character problem.

Treatment built for the wrong environment

Most first-line ADHD treatment focuses on making the ADHD brain perform like a neurotypical one in neurotypical environments. That is a short-term accommodation. It is not a reframe.

When people with ADHD respond poorly to standard interventions, it deepens the shame about not being able to simply do the thing. Rarely is the question asked: what would it look like to build the environment differently?

Environments with external structure. Novelty. Stakes. Output evaluated on result, not process. In those environments, the traits called symptoms frequently become assets — fast pattern recognition, deep focus, crisis performance.

ADHD is not a failure of attention. It’s a failure of environments to engage it.

The structural diagnosis

ADHD is not a broken attention system.

It is a different attention architecture operating inside institutions designed around a prototype it was never built to match. The diagnostic criteria were shaped by observations of boys. The treatment model targets compliance with neurotypical environments. The delay in diagnosis for girls and women is not randomness — it is the predictable outcome of measuring the wrong things on the wrong population.

What gets called a disorder is, in most cases, a design mismatch. The friction is real. The suffering is real. But the source is not located where we have been told to look.

That matters. Because where you locate the source determines what you try to fix.

When the same experiences are repeatedly dismissed across different people and contexts, that is not randomness. That is structure.

Sources

1. DSM-5/ICD-11 field trial participant composition (21% girls): Diouf, A. et al. (2025). “Bias by gender: exploring gender-based differences in the endorsement of ADHD symptoms and impairment among adult patients.” Frontiers in Global Women’s Health. https://pmc.ncbi.nlm.nih.gov/articles/PMC11965619/ (citing Lahey et al., 1994 field trials)

2. Interest-based nervous system / inconsistent attention: Dodson, W. (2025). “Secrets of the ADHD Brain.” ADDitude Magazine. https://www.additudemag.com/secrets-of-the-adhd-brain/ — Dr. Dodson is a board-certified adult psychiatrist and Life Fellow of the American Psychiatric Association.

3. Hyperfocus prevalence (68%): [Author TBC]. (2025). “Hyperfocus in ADHD: A Misunderstood Cognitive Phenomenon.” European Psychiatry, Vol. 68, Suppl. 1. https://pmc.ncbi.nlm.nih.gov/articles/PMC12437476/ — Study of 50 adults with ADHD; 68% reported frequent hyperfocus episodes lasting hours to days.

4. CDC diagnosis rates (15% boys, 8% girls): CDC. “Data and Statistics on ADHD.” https://www.cdc.gov/adhd/data/index.html — Based on National Health Interview Survey data (ages 3–17).

5. Swedish population register / diagnostic delay and comorbidity: Skoglund, C. et al. (2023). “Time after time: failure to identify and support females with ADHD — a Swedish population register study.” Journal of Child Psychology and Psychiatry. https://acamh.onlinelibrary.wiley.com/doi/10.1111/jcpp.13920 — N=85,330; average diagnosis age 23.5 for females vs. 19.6 for males; anxiety disorders 50% vs. 26%, mood disorders 38% vs. 20%.

Editor’s note: This essay examines Applied Behavior Analysis not as a verdict, but as a case study in institutional design — specifically, in how systems define progress, who bears the cost of adaptation, and what tends to get left unmeasured. This is not a settled conversation. It’s one that is still evolving.

Applied Behavior Analysis (ABA) is one of the most widely used early interventions for autistic children.

Over time, it has evolved. Many programs today are more play-based, individualized, and relationship-driven than they once were. They often focus on communication, daily living skills, and increasing independence.

That evolution matters.

But most conversations about ABA focus on how it is delivered. Very few examine what it is designed to produce — and how those outcomes are defined in the first place.

What Research Tells Us

Research on ABA tends to measure observable change: improvements in communication, adaptive behavior, and reduction of behaviors considered disruptive. These outcomes can be meaningful. They can improve access, safety, and daily functioning.

But they are also externally defined and externally measured.

Which raises an important question: what do these measures capture — and what do they leave out?

What Often Goes Unmeasured

I’m sharing this as someone who is autistic. Not to dismiss ABA, but to add a perspective that is often missing from how it’s discussed.

Because alongside visible progress, there is often an internal experience that is harder to quantify.

Learning to tolerate environments that feel overwhelming. Adjusting natural communication styles. Responding in ways that are more socially expected. Even when done with care, the direction of change often flows one way — from the individual toward the environment.

More recent research, driven largely by autistic adults asking questions that behavioral frameworks hadn’t thought to ask, has begun naming longer-term patterns. Studies on masking and camouflaging consistently show that the effort required to suppress natural autistic behavior in order to appear neurotypically legible is not cost-free — it is associated with anxiety, depression, identity strain, and what researchers now call autistic burnout. Raymaker et al. (2020) formally defined autistic burnout as a syndrome resulting from chronic life stress and a mismatch of expectations and abilities without adequate supports — characterized by prolonged exhaustion, loss of skills, and reduced tolerance to stimulus. Critically, their findings explicitly warned against the dangers of teaching autistic people to mask or suppress their traits in order to meet mainstream expectations.

These outcomes don’t always surface early. And they are rarely part of how success is measured at the point when intervention decisions are made.

Research tells us a great deal about what changes externally. It tells us far less about the long-term experience of being the one who had to change.

The Direction of Adaptation

This is where the conversation becomes less about method, and more about design.

Every intervention is built around an implicit definition of progress. And that definition is never neutral. It reflects assumptions about what counts as functional, what counts as successful, and what kind of person the system was designed to optimize for.

If progress is defined as increased compliance, improved tolerance, and more “appropriate” behavior — we have to ask: appropriate for whom?

This question has a structural answer. Damian Milton’s Double Empathy Problem (2012) reframes the entire premise: difficulties in social communication between autistic and non-autistic people are not one-sided deficits. They are bidirectional. Both groups struggle to understand each other. Yet intervention has historically only targeted one side. As researchers studying camouflaging have noted, autistic people invest significant time and energy learning to fit into the neurotypical world — often to the detriment of their mental health — while the reverse adaptation is rarely expected or measured.

When the environment stays the same and the individual is expected to adapt, support can still mean learning to navigate mismatch more efficiently rather than reducing it. That’s not a neutral outcome. It’s a design decision.

A Systems Lens on Support

From a design perspective, every institution embeds assumptions about who it was built for. The diagnostic question I keep returning to is this: what human prototype was this system designed around?

In the case of most behavioral interventions, the prototype is neurotypical. Communication means reciprocal verbal exchange in the expected format. Appropriate behavior means behavior that doesn’t disrupt social legibility. Success means fit.

This isn’t a clinical failure. It’s a design feature. Early ABA in particular was often evaluated on its ability to make autistic individuals appear, in the field’s own language, “indistinguishable from peers” — a goal that research has since linked to long-term harm including shame, loss of agency, and identity erosion. And when the measure of progress is how well the person fits the system — rather than how well the system fits the person — the intervention that follows will, by design, target the individual.

If the environment stays the same, and the individual is expected to adapt, progress will require ongoing effort from the person. That can work in the short term. But over a lifetime, that effort accumulates.

Support shouldn’t only help individuals navigate the world as it is. It should also work to reduce the mismatch between the person and the environments they move through.

What Parents Should Consider

None of this is about rejecting support. It’s about expanding how we define it.

Some questions worth sitting with:

What outcomes are we actually aiming for — and who defined them as outcomes? Is this building autonomy, or primarily increasing compliance? Is my child becoming more themselves, or better at performing expectations? And what parts of the environment could change alongside them, rather than requiring them to absorb all the change?

Because support doesn’t happen in isolation. It happens within a system. And systems have designs — whether or not we name them.

Closing

Skill-building matters.

But so does direction.

The goal isn’t just to help someone fit into the world as it currently exists. It’s to create conditions where they don’t have to continuously reshape themselves in order to belong.

That requires asking what we are measuring. What we are not measuring. And what kind of person we are implicitly asking a child to become in order to be considered successful.

When the same experiences are repeatedly dismissed across different people and contexts, that is not randomness. That is structure.

We have built entire systems around the assumption that the autistic person should do the adapting.

Not because we examined that assumption. Because we never had to.

The environment doesn’t get referred for evaluation. The environment doesn’t receive intervention hours. The environment doesn’t get measured for outcomes.

Only the person does.

That is not a clinical finding. It is a cultural default. And defaults are not neutral — they just feel that way when no one has named them yet.

We talk a lot about helping individuals adapt.

We talk much less about why adaptation is always theirs to do.

I’d be curious how others are sitting with that.

Works Cited

Hull, L., Mandy, W., Lai, M. C., Baron-Cohen, S., Allison, C., Smith, P., & Petrides, K. V. (2019). Development and validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). Journal of Autism and Developmental Disorders, 49(3), 819–833.

Hull, L., Petrides, K. V., Allison, C., Smith, P., Baron-Cohen, S., Lai, M. C., & Mandy, W. (2017). “Putting on my best normal”: Social camouflaging in adults with autism spectrum conditions. Journal of Autism and Developmental Disorders, 47(8), 2519–2534.

Milton, D. E. M. (2012). On the ontological status of autism: The ‘double empathy problem.’ Disability & Society, 27(6), 883–887.

Pearson, A., & Rose, K. (2021). A conceptual analysis of autistic masking: Understanding the narrative of stigma and the illusion of choice. Autism in Adulthood, 3(1), 52–60.

Raymaker, D. M., Teo, A. R., Steckler, N. A., Lentz, B., Scharer, M., Delos Santos, A., Kapp, S. K., Hunter, M., Joyce, A., & Nicolaidis, C. (2020). “Having all of your internal resources exhausted beyond measure and being left with no clean-up crew”: Defining autistic burnout. Autism in Adulthood, 2(2), 132–143.

Sparrow, M. (2016). Therapy should make your child better, not traumatize them. Unstrange Mind.

Wilkenfeld, D. A., & McCarthy, A. M. (2020). Ethical concerns with applied behavior analysis for autism spectrum disorder. Kennedy Institute of Ethics Journal, 30(1), 31–69.

The mental health system has a number it should have to answer for.

85 percent.

That’s the proportion of autistic children who carry at least one comorbid psychiatric diagnosis. Not a tendency toward distress. Not elevated risk. A majority — already in possession of an additional clinical label before they’ve finished primary school.

The system’s response to that number has been to add more labels, prescribe more medication, and develop more intervention frameworks aimed at the person carrying the burden.

What it has not done is ask what produced the burden.

The Pattern the System Keeps Naming and Never Explaining

Depression affects 26 percent of autistic people. In the general population, it’s 10 percent.

Anxiety: 30 percent of autistic people. General population: 10 percent.

Bipolar disorder: 11 percent versus 2.

ADHD co-occurs in 50 to 70 percent of autistic people — so frequently that researchers have begun treating the overlap as its own clinical category.

In children with ADHD alone, comorbidity rates run between 40 and 80 percent. Researchers studying the data have noted that an ADHD diagnosis without comorbidities is the exception, not the rule.

And when both autism and ADHD are present — the group researchers now sometimes call AuDHD — anxiety reaches 72 percent in autistic females and 69 percent in autistic males. Depression hits 38 to 39 percent. Both figures are up to ten times the prevalence seen in adolescents without either diagnosis.

The mental health system has documented this pattern meticulously. It has named every condition in it. It has built treatment protocols for each one.

What it has not done is ask why the same people keep arriving with the same constellation of diagnoses, across countries, across decades, across every demographic group studied.

When a pattern is that consistent, the diagnosis belongs to the pattern — not only to the people inside it.

This is not a comorbidity problem. It is a systems problem.

What the Mental Health System Was Built to See

Every institution embeds a prototype. A default human being it was designed to evaluate, treat, and serve.

The mental health system’s prototype presents distress in legible ways. Describes internal states with precision and consistency. Responds to talking therapies that assume a particular relationship to language, narrative, and self-reflection. Arrives at appointments on time, in the right affect, with the right level of insight. Tolerates the sensory and social demands of clinical environments without that tolerance itself becoming a source of exhaustion.

When patients don’t match this prototype — when they communicate differently, mask their distress so thoroughly it doesn’t register, or arrive burnt out in ways the diagnostic frameworks weren’t designed to recognize — the system doesn’t adapt its instruments. It questions the patient’s insight. It notes poor therapeutic engagement. It adds another diagnosis.

This is not a failure of individual clinicians. It is a feature of how the system was built. The instruments were designed for someone. When someone else shows up, the instruments keep running — and what gets measured is the distance from the prototype, not the distress being produced by that distance.

Mismatch gets labeled as pathology. The system calls this treatment.

The Misdiagnosis Record

There is a way to test whether a diagnostic system is seeing clearly. You look at what it gets wrong — and whether it gets it wrong in patterns.

One in four autistic adults, and one in three autistic women, reported at least one psychiatric diagnosis obtained prior to their autism diagnosis that they perceived as a misdiagnosis. These are not isolated cases of clinician error. Personality disorders, anxiety disorders, and mood disorders are consistently the most common misdiagnoses — with autistic women particularly susceptible to misdiagnosis of personality disorder.

Research found an average ten-year delay between autistic women having first contact with mental health services and receiving an autism diagnosis — considerably longer than for men. The most common misdiagnosis for women at first evaluation was personality disorder, at 36 percent.

Ten years. In the mental health system. Being treated for the wrong thing.

Nearly 80 percent of women with autism are misdiagnosed — often with borderline personality disorder, eating disorders, bipolar disorder, and anxiety. The treatment those women received during that decade was real treatment — medication, therapy, clinical frameworks — built on an incorrect understanding of what was actually happening.

There is increasing clinical recognition that recurrent and treatment-resistant depression may reflect an underlying neurodevelopmental condition such as autism, missed due to overlapping clinical symptoms. The system’s own literature is beginning to acknowledge what patients have been saying for years: the diagnosis was wrong because the instrument couldn’t see past the presentation.

This is what a system looks like when it was built to recognize one kind of person and encounters another.

Masking Is the Gap the System Keeps Missing

There is a mechanism that sits between neurodivergent experience and mental health outcomes, and the clinical system has been largely unable to see it because it is, by definition, invisible.

Masking is the sustained, often unconscious suppression of neurodivergent traits in order to appear closer to what the environment — including the clinical environment — expects. It works well enough that it fools the people doing it, sometimes for decades. It is why so many autistic people, particularly women, receive diagnoses of anxiety, depression, and personality disorders for years before anyone asks whether the underlying presentation is something else entirely.

A 2021 study found that autistic employees who masked more heavily reported significantly higher levels of workplace stress, anxiety, and burnout. That relationship — more masking, more distress — is not incidental. It is the mechanism made visible.

But the mental health system is often the last place masking gets named. Because the same performance that passes in a job interview also passes in a clinical intake. The patient presents as articulate, self-aware, mildly anxious. The depression gets treated. The masking that produced it doesn’t.

And the environment that demanded the masking is never in the room at all.

Open-plan offices, fluorescent lighting, unpredictable schedules, constant context-switching — these are design choices made for a specific nervous system. When an environment is itself a source of sustained demand, the cost of being present in it is a tax paid before the work begins. The mental health system receives the invoice. It calls the invoice a diagnosis. It does not look for the source.

72 percent of autistic adults scored above the psychiatric cut-off for suicide risk — compared to 33 percent of the general population. Camouflaging significantly predicted suicidality even after controlling for depression, anxiety, employment, and other factors. Masking isn’t just a coping strategy the system fails to address. It is a measurable path toward the most serious outcomes the system is supposed to prevent.

The system is treating the destination. It is not looking at the road.

The Diagnostic Framework Is Part of the Problem

The DSM categorizes. It names conditions. It draws boundaries between what counts as pathology and what doesn’t.

What it does not do is ask what produced the conditions it names.

For neurodivergent people, this omission has structural consequences. Burnout becomes a depressive episode. The sensory exhaustion of operating in an environment built for a different nervous system becomes generalized anxiety disorder. The social isolation that follows years of failed attempts to navigate unspoken rules becomes avoidant personality traits. The grief of late diagnosis — of decades of being told you were difficult, lazy, too sensitive, not trying hard enough — gets coded as dysthymia.

Each of these is a real condition. The treatment for each is real treatment. But when the framework stops at the label, it leaves the cause intact. The person is treated and returned to the same environment that produced the distress. The cycle restarts. The system documents the relapse.

Autistic adults are more likely to receive psychiatric diagnoses compared with non-autistic adults — but significantly less likely to agree with those diagnoses. That gap between what the system assigns and what patients recognize as true is not a compliance problem. It is a measurement problem. The instrument is producing results the people it is measuring do not recognize as accurate.

This is what a system looks like when it was not built to ask the right question.

The right question is not: what is wrong with this person? The right question is: what human prototype was this system designed for — and what does it do to the people who don’t match it?

The Global Picture Removes the Excuse

This is not an American problem with the American healthcare system.

The WHO estimates autism affects approximately 1 in 100 children worldwide. A cross-national study of adults with ADHD across 20 countries found anxiety disorders present in 34 percent of cases and mood disorders in 22 percent — consistent regardless of where the research was conducted.

When the same clinical outcomes appear across different countries, different healthcare systems, different cultural contexts, different diagnostic traditions — that consistency is evidence. It means the variable producing the outcomes is not located in any one system’s particular failures. It is located in something all of these systems share.

What they share is a framework that treats neurodivergent distress as a property of the person, not a response to conditions. What they share is a prototype — and the habit of measuring everyone against it, and calling the gap a disorder.

That is not a healthcare finding. That is a design finding.

What the System Owes Instead

The principle the mental health system has not yet applied to itself: institutions are evaluated not by what they claim to value, but by who bears the cost of their design assumptions.

The mental health system claims to treat distress. What it has largely done, for neurodivergent people, is document it. Name it. Return the person to the conditions that produced it, with better tools for managing those conditions — tools that require the person to do more work, not the institution to do different work.

35 percent of autistic children with a psychiatric diagnosis are already on at least one psychotropic medication. The system that prescribed that medication did not design itself differently. It prescribed.

When you change the instrument, different people succeed. When you redesign the environment, the distress diminishes. When clinicians are trained to recognize masking, late-diagnosed autistic adults stop spending a decade in the wrong treatment. The research on all of this is consistent.

Which means the system has had the information. What it has not had is the framework that makes structural redesign the obligation — rather than the aspiration.

Here is what that framework makes visible: what the mental health system has been calling a diagnosis is often a design mismatch. A person whose nervous system was not built for the environment they are required to inhabit. A person whose communication, regulation, and sensory experience diverge from what the institution was designed to accommodate. The friction that produces is real. The distress that follows is real. But the location of the problem has been misidentified.

The mismatch is between the individual and the environment. The diagnosis has been assigned to the individual.

That is not a clinical error made by one practitioner in one appointment. It is a structural error built into a system that was never designed to ask whether the environment itself was the problem.

Neurodiversity is not the crisis. It is the signal.

The system has been diagnosing the signal instead of reading it.

The burden belongs to the institution. Not to the person who has been carrying it.

When the same experiences are repeatedly dismissed across different people and contexts, that is not randomness. That is structure.

If this framework is useful to you:

If you want tools for understanding and naming the systems you move through — frameworks, reframe cards, and workbooks built for neurodivergent people — they live at LegitimacyLab on Etsy: etsy.com/shop/LegitimacyLab

because it was never just you.