GOLDEN AUTISM &
THE CRUELTY OF CARE

1INTRODUCTION: THE GOLD THAT WAS PAID FOR

Golden Autism is this article’s name for autism first recognized in the second half of a human’s life, the golden half that holds what we call the golden years, in adults who survived decades of doing what the world asked of them, going to school, working, building organizations, raising children, keeping relationships, learning to read rooms, anticipate expectations, suppress discomfort, perform competence and keep moving. The name carries a second meaning that is worth examining as a society. The world treated the autistic performance of masking as gold, praising it, promoting it and building its expectations around it. Adults who reach diagnosis this way are no longer rare. Ginny Russell and colleagues, working from primary care records in the United Kingdom, found a 787 percent rise in recorded autism diagnoses between 1998 and 2018, with the steepest increases among adults and women, and they attributed most of that rise to better recognition of people who had been autistic all along (Russell et al., 2022). The second half of life brings its own reckonings, and researchers have documented autistic women reassessing a lifetime of masking during the menopausal transition, when the old performance becomes harder to sustain (Karavidas & de Visser, 2022).

From the outside the record of these lives looked like resilience. We call it resilience when an adult continues doing it successfully, and it is worth asking whether much of it was adaptation, a different thing from health, and whether one of the cruelest things that can happen to a person is to become so good at adapting to conditions that hurt them that everyone, the person included, starts to take the adaptation as evidence that nothing is wrong. In that arrangement the mask becomes evidence against the person wearing it.

A late diagnosis changes the past. The events and details all stay the same, and the framework for understanding them moves, so the exhaustion, the vigilance, the slow recovery, the relationships that became impossible to sustain and the careers that came apart each acquire a new history. Beside the question of whether the diagnosis is correct sits a harder one, whether some of the systems meant to care for these adults were part of the problem. Plenty of hard experiences have other causes, relationships end without anyone abusing anyone, and institutions cause harm with no malice at all, which is exactly what makes this kind of harm so hard to see.

Harm can arrive looking like care. It shows up as treatment or repair, as accommodation that requires endless disclosure, as a workplace asking for resilience from a nervous system already exhausted, as a family reading dysregulation as selfishness, as a professional deciding the behavior is the problem without asking what the behavior is communicating. And sometimes the person who has spent almost fifty years adapting to everyone else’s expectations finally cannot do it anymore. We call that failure. Collapse may be evidence, and the families, employers, clinicians and friends around a late-diagnosed adult can learn to read it as information they can use.

What happens when a society systematically requires autistic people to adapt themselves to neurotypical norms, rewards them when they successfully disappear their differences, and then interprets the consequences of that adaptation as evidence of individual dysfunction?

The article answers that question in two registers and keeps them apart. Where it reports research, it names the study, the sample and the limits of what the design can show, separating clinical samples from population records, prevalence from odds, and risk markers from causes. Where it moves past the research into theory, it numbers the move as a proposition, four in all, so a reader can accept the evidence and still argue with the theory, or accept the theory and demand more evidence, without confusing one for the other. Sections 2 through 6 carry the evidence, on masking and stigma, suicide risk and the course of masking over time, the history and structure of professional help, adult diagnosis, and the double empathy problem. Sections 7 and 8 carry the theory, the construct of neurotypical abuse and the framework of Meta Macro Social Work. Section 9 sets the pieces inside a single life, and Section 10 turns to repair.

2THE PRICE OF BELONGING: MASKING, STIGMA AND BURNOUT

The research has begun to give this cruel evisceration of one’s personhood a name. In 2020, Dora Raymaker and a team of autistic and nonautistic researchers working through the Academic Autism Spectrum Partnership in Research and Education published a qualitative study in Autism in Adulthood, built from 19 interviews and 19 public online accounts, describing autistic burnout as a state of chronic exhaustion, lost skills and lowered tolerance for sound, light and social demand, brought on by life stress that outpaced a person’s capacity and arrived without adequate support (Raymaker et al., 2020). The people in that study described losing abilities they had held for most of their lives, the ability to speak easily in a meeting, to cook a meal, to answer email, to walk through a grocery store, and they described the people around them reading that loss as laziness, depression, or a sudden change in character. Participants named masking among the roads into burnout, the steady work of hiding autistic traits and performing the expected version of oneself, work that looks effortless from outside and costs a great deal inside. They also described what burnout did to their health, their independence and their will to live, including suicidal behavior, and the authors recommended that burnout education become part of suicide prevention.

Masking is learned under pressure. Amy Pearson and Kieran Rose argued in 2021 that masking is best understood as a response to stigma, and they questioned how much of it can honestly be called a choice, since a person who is punished for being seen has little real freedom to be seen (Pearson & Rose, 2021). Radulski went further, describing the social advantages that flow to people whose neurology matches the norm as neurotypical privilege, and proposing a minority group model of neurodiversity in which masking is the cost a minority pays to move through a world arranged for the majority (Radulski, 2022). Survey evidence points the same way. Perry and colleagues found, among 223 autistic adults, that higher perceived autism stigma predicted more camouflaging (Perry et al., 2022), and Monique Botha and David Frost, applying the minority stress model first developed for sexual minorities (Meyer, 2003), found that everyday discrimination, expectations of rejection, internalized stigma and concealment each predicted lower wellbeing and greater psychological distress in 111 autistic adults (Botha & Frost, 2020). Nonautistic people also adjust themselves to fit in, and researchers asking autistic and nonautistic adults about masking heard some version of “masking is life” from both (Miller et al., 2021). What sets autistic masking apart is its scale and its compulsion, a performance sustained across every setting for decades under the threat of being read as defective. In this literature masking is what a person does when being seen as autistic carries a price.

That changes how a late-diagnosed life is understood. The undiagnosed autistic adult reads every room correctly, sets aside discomfort and pain so a meeting can go smoothly, and offers a performance at work, in the community and at home so that others can stay comfortable. Society takes these gifts from the undiagnosed adult without that person’s knowledge or informed consent, because consent requires knowing the price, and the price is exactly what an undiagnosed person has no language to name. Each gift extracted this way leaves a wound that is deep and invisible, and the people who benefited had little reason to check on the giver, because the painful performance was the very thing they valued. An adult diagnosed at forty or fifty has usually been paying this way since childhood, long before anyone, including themselves, had a word for what it cost.

Most people hear the word masking and picture something mild, a social habit on the order of stifling a yawn in a long meeting, and the people trained into it describe something far heavier. When Laura Hull and colleagues at University College London asked 92 autistic adults from fifteen countries about camouflaging, participants described exhaustion that required long stretches alone to recover, constant anxiety from monitoring every exchange, and a slow loss of any sense of who they were underneath the performance (Hull et al., 2017). One participant said her head raced during ordinary conversation “as if I’m interpreting another language.” Another, a woman of 48, described “the weight of a black cloud” that came with “having to be this fake version of me.” Imagine translating every sentence you hear and every sentence you speak, all day, for forty years, while being graded on how natural it looks.

The quantitative picture is consistent and incomplete. A 2024 meta-analysis pooling sixteen studies found that camouflaging correlated with depression at r = .36, with generalized anxiety at r = .27 and with social anxiety at r = .24, and inversely with wellbeing, while noting that none of the included studies followed people over time and that many were too small to detect modest effects (Khudiakova et al., 2024). Most of this research draws on online and clinic samples in which women, white participants and highly educated adults are overrepresented. What the literature establishes firmly is the direction of pressure. Stigma predicts camouflaging, camouflaging travels with distress, and burnout is described by the people who live it as the bill that eventually comes due.

3SUICIDE RISK AND THE COURSE OF MASKING OVER TIME

Three kinds of evidence speak to suicide risk among autistic adults, and each says something different. The first comes from a clinic. In 2014 Sarah Cassidy and colleagues at the University of Cambridge surveyed 374 adults diagnosed with Asperger syndrome in adulthood at a specialist diagnostic service, and 66 percent reported having thought about suicide, against a lifetime rate of 17 percent in the general population of England and higher than the 59 percent reported among people with psychosis, while 35 percent had planned or attempted suicide (Cassidy et al., 2014). Clinic samples draw people who are struggling enough to seek help, so these figures describe adults who sought diagnosis, many of them in crisis, which makes them a fair portrait of the Golden Autism population at the moment it becomes visible and a poor estimate for autistic adults as a whole.

The second kind comes from community surveys. In 2018 Cassidy’s team compared 164 autistic adults with 169 adults from the general population and found that 72 percent of the autistic group scored at or above the clinical threshold for suicide risk, against 33.7 percent of the comparison group. Two risk markers particular to autistic adults, camouflaging and the number of unmet support needs, predicted suicidality after accounting for age, gender and diagnosed mental health conditions (Cassidy et al., 2018). The design was cross-sectional, so these are markers that travel with risk, and the study cannot show that either one causes it. In that sample camouflaging was unrelated to the age at which people were diagnosed, a reminder that masking belongs to autistic people diagnosed early as well as late.

The third kind comes from population registers. Tatja Hirvikoski and colleagues compared 27,122 people with an autism diagnosis in Swedish national records against 2,672,185 matched controls and found that the autistic group had 7.55 times the odds of dying by suicide during the study period, and among autistic people without intellectual disability, the group from which Golden Autism is drawn, the odds were 9.40 times higher (Hirvikoski et al., 2016). The same study is widely misquoted as proof that autistic people die sixteen years early. That figure comes from the mean age at death among the 2.6 percent of autistic participants who died during follow-up, 53.87 years against 70.20 among controls who died, and a 2025 analysis in JAMA Network Open found that seven in ten citing articles misrepresented the study, most often by turning that figure into a claim about life expectancy (Hand et al., 2025). The accurate reading needs no exaggeration.

The most important test of the masking argument comes from the Netherlands. Wikke van der Putten and colleagues followed 332 autistic adults between the ages of 30 and 84, diagnosed on average at 44, which describes the Golden Autism population almost exactly, across two measurements about two years apart, and found that higher camouflaging at the start predicted slightly fewer mental health difficulties two years later, while lower camouflaging predicted more, and that mental health difficulties at the start did not predict later changes in camouflaging (van der Putten et al., 2025). The effect was small, and two measurements cannot settle direction. An earlier analysis of the same Dutch cohort found that camouflaging was strongly tied to mental health difficulties for only about one participant in ten, and weakly or not at all for most others (van der Putten et al., 2024), and a 2026 study of autistic adults with social anxiety found that camouflaging no longer predicted distress once social anxiety was taken into account, raising the possibility that camouflaging questionnaires partly measure anxiety itself (Roisenberg et al., 2026).

The literature present a clear pathway for the Golden Autistic. Over a span of two years masking can work, holding a life together while sustaining biologically and behaviorally marked traumatic abuse to the brain, and one reading of the Dutch finding, consistent with autistic adults’ accounts of the mask slipping during burnout (Bradley et al., 2021), is that the adults whose masking falls away are the ones whose difficulties climb, because they have depleted most of their emotional, social, economic, and physical resources for repair. It is similar to effects observed in communities colonized by Europeans. Masking, unmet support needs, stigma and discrimination, isolation, co-occurring depression and anxiety, and relentless environmental demand travel together in these data, and Perry’s team found that stigma harmed wellbeing through routes beyond camouflaging alone (Perry et al., 2022). Several of those markers share a source in the environment, in what the person is asked to hide, what support is withheld, and how often the person is told that their way of being is wrong. The harm this article describes therefore lives in the demand to mask and in the collapse that follows when the adaptation fails, a synthesis drawn across studies that calls for research long enough to follow that bill across decades. A paper can be careful about cause and still state plainly that autistic adults without intellectual disability, the ones most often praised for coping, die by suicide at many times the rate of their neighbors.

4ILLICH AND THE HARM OF HELP

Ivan Illich gave us language for harm that comes from the helper. In Medical Nemesis he described iatrogenesis on three levels (Illich, 1975). Clinical iatrogenesis is the damage done directly by treatment. Social iatrogenesis is the damage done when professional systems take over capacities people once held in common, turning ordinary life into a series of conditions requiring professional management and teaching people to depend on the institutions that took those capacities away. Cultural iatrogenesis is the deepest, the erosion of a people’s own ability to understand and bear their suffering on their own terms. Two years later, in Disabling Professions, Illich and his coauthors described how experts come to hold the authority to define a person’s needs, so that the person’s own account of their life counts for less than the professional’s reading of it (Illich et al., 1977).

Each level has an autistic form. Clinical iatrogenesis appears when an autistic adult in burnout is misdiagnosed and treated for the wrong condition, and in a Dutch study of 1,211 autistic adults, one in four reported a psychiatric misdiagnosis before autism was recognized, most often a personality disorder (Kentrou et al., 2024). Social iatrogenesis appears when support can be reached only through a professional gate, a diagnosis, a letter, a documented limit, so that the person’s account of what they need is inadmissible until an expert repeats it. Cultural iatrogenesis appears when autistic people absorb the deficit language of the systems around them so thoroughly that they come to read their own exhaustion as a character flaw, which is the internalized stigma Botha and Frost measured.

The history of treatment shows how the goal was set. The diagnostic manual describes autism through deficits in social communication and interaction (American Psychiatric Association, 2013), and much of twentieth century treatment aimed at making autistic children look less autistic, rewarding eye contact, extinguishing rocking and hand movement, and measuring success by resemblance to peers. In his landmark 1987 study, Ole Ivar Lovaas reported that 9 of 19 young children who received about forty hours a week of behavioral treatment, a program that included contingent aversives such as a loud “no” or a slap on the thigh, reached normal educational and intellectual functioning, and he described them as indistinguishable from their classmates (Lovaas, 1987). The study assigned children to groups without randomization, and a later randomized trial reported more modest results (Smith et al., 2000). Behavioral practice today varies widely, and many practitioners have abandoned aversives and the language of recovery. The goal of indistinguishability, however, still shapes outcome measures, and much of that work, offered with love, taught children the exact skill that the burnout research now links to collapse, hiding what they feel so others can be at ease. Adults diagnosed late often learned the same skill with no program at all, from teachers, parents, bosses and peers who rewarded the performance and punished the rest.

PROPOSITION 1 · ADAPTIVE IATROGENESIS.Adaptive iatrogenesis is harm produced by care that improves a person’s performance inside an injuring environment while leaving the environment untouched, so that the success of the care is measured by how well the person tolerates the injury. It is this article’s extension of Illich, and it names a pattern the evidence above makes visible without yet measuring directly. The Dutch longitudinal finding sharpens it, since care that props up masking may produce real short-term stability while deferring the cost to the moment the adaptation fails.

Illich’s questions can now be put to autism systems directly. Helping becomes harmful when success is measured by how closely a person resembles peers, because the measure rewards the very concealment that the suicide data mark as dangerous. Expertise begins defining the person’s needs more than the person does when a test score outranks a life history, which Section 5 takes up. Treatment becomes adaptation to an unhealthy environment when a therapist teaches an exhausted adult to tolerate open offices, fluorescent light and back to back meetings and nobody asks the office to change. Accommodation becomes surveillance when the person must document their limits, disclose them to managers, and renew the evidence at every review in order to keep a quiet room, so that support arrives bundled with a file. And care teaches a person to tolerate the conditions injuring them whenever the treatment plan lists the person’s coping as the goal and lists the environment nowhere at all.

5BEING AUTISTIC, BEING DIAGNOSED, BEING RECOGNIZED

Five different conditions are often collapsed into one. A person can be autistic, which is a fact about their neurology whether anyone has noticed it or not. A person can be diagnosed, which is a clinical and administrative event. A person can be recognized, which happens when the people around them, and the person themselves, understand their life in light of autism. A person can demonstrate autistic traits during an assessment, which depends on the conditions of the session. And a person can have learned compensatory strategies so thoroughly that their traits are hard to see under exactly those conditions. Golden Autism describes people who were autistic for decades before they were diagnosed or recognized, and whose skill at compensation is the reason.

Diagnosing a child and diagnosing an adult are different undertakings. A child usually comes to an evaluation because a parent or teacher noticed something, the child has had little time to learn to hide, and the clinician can watch the child play, interview parents about early years that are still fresh, and compare the child against developmental milestones everyone can still name. The standard instruments grew up in that setting. The Autism Diagnostic Interview, Revised, is a long structured interview with a parent or caregiver about the person’s early development, and the Autism Diagnostic Observation Schedule, second edition, is a set of structured activities in which a trained examiner watches for autistic patterns of social communication over roughly an hour.

An adult of forty or fifty meets those tools under very different conditions. The parents who could describe the early years may have died, may remember a childhood four decades gone through the hopes they held for it, or may have been the source of the pressure to hide. The adult has spent a lifetime rehearsing the very behaviors the observation schedule scores, eye contact, back and forth conversation, gesture and shared attention, and a person skilled at camouflage can perform them through an hour in an office and pay for it afterward at home. An hour of skilled observation identifies autism in many adults, and the danger lies in treating a single performance as the whole person. Scotland’s National Autism Implementation Team advised clinicians in 2022 that the observation schedule should never be used alone, that it is not required for a robust diagnosis, and that a score below threshold may reflect a person who is masking, for whom a camouflaging measure should be considered (NAIT, 2022). Researchers at the Centers for Disease Control and Prevention named the same problem in their own study, writing that “there are no psychometrically validated tests of ASD for adults, which leads to uncertainty for studies using tests designed for children, such as the Autism Diagnostic Observation Schedule” (Dietz et al., 2020). The diagnostic manual itself anticipates the adult case, stating that symptoms must be present in early development but may not become fully manifest until social demands exceed limited capacities, or “may be masked by learned strategies in later life” (American Psychiatric Association, 2013).

The adults most likely to be missed are the ones the older picture of autism never described. A 2017 meta-analysis by Rachel Loomes, Laura Hull and William Mandy found the true ratio of autistic males to females closer to three to one than the four to one usually quoted, and concluded that girls who meet the criteria are at disproportionate risk of never receiving a clinical diagnosis (Loomes et al., 2017). Meng-Chuan Lai and Simon Baron-Cohen called the adults who grew up before autism without intellectual disability was widely recognized a lost generation, many of whom reached adulthood carrying other labels (Lai & Baron-Cohen, 2015). Those other labels are common. In the Dutch study, 24.6 percent of autistic adults reported at least one psychiatric misdiagnosis, rising to 31.7 percent among women, most often a personality disorder, followed by anxiety disorders, mood disorders and burnout, and misdiagnosis was most common among people diagnosed between the ages of 18 and 54 (Kentrou et al., 2024). Seeming fine can itself delay recognition. In a small study of 40 autistic adults, Belcher and colleagues found that autistic men who made better first impressions on strangers in brief video clips had been diagnosed later in life (Belcher et al., 2022). In the United States, researchers at the Centers for Disease Control and Prevention estimated that 5,437,988 adults between the ages of 18 and 84, about 2.21 percent or one in 45, were autistic in 2017, including 111,067 in Indiana, a figure that counts diagnosed and undiagnosed adults alike and had to be modeled from childhood data because no surveillance system in the country counts autistic adults (Dietz et al., 2020).

Adult assessment works best when it integrates developmental history where it exists, functioning traced across years and settings, direct inquiry into masking and compensatory strategies, the person’s own account, and clinical judgment, treating each as one line of evidence. This is why a clinician who has treated a person over years can hold evidence the testing room cannot produce. A therapist or physician who has walked with a patient through job changes, grief, illness and recovery has seen how that person handles transitions, what happens after a crowded week, how sensory load and social demand show up in the body, and what the patient’s account of their inner life sounds like when there is no performance left to give. Long clinical observation of that kind can reach behind the mask in a way a single session rarely does, and it fits the manual’s own standard, which rests diagnosis on clinical judgment of the whole history, with instruments serving as supports. Families, employers and insurers who accept only a test score as proof are asking for the one form of evidence that masking was built to defeat.

A clinician sees behavior. A family member sees behavior. An employer sees performance. A diagnostic instrument captures behavior under specific conditions. The autistic person experiences the cumulative physiological, cognitive, sensory, emotional and social cost of producing that behavior, and no outside observer has direct access to that cost. Self-report is fallible, and a meta-analysis found alexithymia, difficulty identifying and describing one’s own emotions, in 49.93 percent of autistic participants against 4.89 percent of nonautistic participants (Kinnaird et al., 2019), though the authors noted that the measure itself depends on self-report and that few studies controlled for depression and anxiety, so self-report needs structure and support like any other evidence. It remains essential evidence all the same, and it is systematically undervalued in one predictable situation, when the observer is rewarded for seeing successful adaptation. The employer profits from the performance, the family from the stability, the clinic from a clean outcome, and each has a reason to trust what it sees over what the person reports. That incentive is this article’s synthesis, and it explains why the Golden Autism population can go unseen by people who see them every day.

PROPOSITION 2 · GOLDEN AUTISM.Golden Autism names autism first recognized in the second half of life in adults whose sustained, socially rewarded adaptation concealed their autism from others and often from themselves. Its defining mechanism is attributional reversal, the process by which successful adaptation is read as proof that the disability is mild or absent, so that the mask becomes evidence against the person wearing it, and the eventual collapse is read as a new personal failure instead of the cost of the old adaptation.

6DOUBLE EMPATHY AND THE DISTRIBUTION OF CONSEQUENCES

Damian Milton, an autistic sociologist in England, named the other half of the problem in 2012. He called it the double empathy problem, observing that communication breaks down between autistic and nonautistic people in both directions, and that the nonautistic side, holding the power to diagnose, tends to read the break as proof of the autistic person’s deficit (Milton, 2012). The key word in Milton’s account is power. The nonautistic majority sets the norms, writes the diagnostic criteria, staffs the clinics, runs the hiring committees and defines what counts as a good conversation, so the autistic person’s difficulty reading the majority is recorded as a symptom while the majority’s difficulty reading the autistic person goes unrecorded.

The experimental evidence is strong on some points and mixed on others, and the article’s argument rests on the strong ones. Catherine Crompton and colleagues ran a version of the telephone game in 2020, arranging 72 adults into nine chains of eight, and found that all autistic chains passed a story along as accurately as all nonautistic chains, while mixed chains lost more detail and reported lower rapport (Crompton et al., 2020). A larger preregistered replication in 2025, with 311 participants across Edinburgh, Nottingham and Dallas, confirmed that autistic people share information with one another as successfully as nonautistic people do, found no extra loss of information in mixed chains, found that nonautistic pairs reported the highest rapport, and found that disclosing a diagnosis improved rapport (Crompton et al., 2025). The finding that holds across both studies is the one that matters most here, that autistic communication works among autistic people, which locates the difficulty in the meeting of styles and removes it from the autistic person alone.

The majority’s judgment forms in seconds. Noah Sasson and colleagues showed nonautistic observers clips of autistic and nonautistic people as short as two to four seconds, and even single still frames, and observers consistently rated the autistic people less favorably and were less willing to interact with them. When observers read transcripts of what the same people said, with no voice or image, the bias disappeared, and the authors concluded that style, and never substance, drove the negative impressions (Sasson et al., 2017). Kerrianne Morrison and colleagues paired 125 adults for five minute conversations and found that autistic adults paired with autistic partners disclosed more about themselves and felt closer to their partners (Morrison et al., 2020). When Desiree Jones and colleagues showed videos of such conversations to 102 nonautistic observers, the observers rated mixed conversations as less smooth, rated autistic participants as more awkward and less likable, trustworthy and intelligent, and reported far less interest in the autistic participants than their actual conversation partners did (Jones et al., 2024). The observers could see the gap and reproduced it in the same moment, which is the exact position of a hiring committee, a human resources office or a diagnostic clinic.

A misunderstanding can be mutual while its consequences land on one side. Interpersonal misunderstanding is a problem between two people, and institutional power decides whose account of that problem becomes the record. When a job interview goes badly, the autistic candidate loses the job. When a performance review goes badly, the autistic employee receives the improvement plan. When a clinical encounter goes badly, the autistic patient receives the personality disorder. The double empathy problem, taken seriously, requires institutions to stop treating the majority’s reading of an encounter as neutral.

Because the gap sits in the majority’s perception as much as in autistic communication, the work of closing it belongs to the majority too, and the research points to practical steps. Desiree Jones, Kilee DeBrabander and Noah Sasson gave 238 nonautistic adults a short autism acceptance training built from facts and autistic first person accounts, and the trained group formed more positive impressions of autistic adults and held higher expectations of their abilities, though implicit bias did not move (Jones et al., 2021). Kristen Gillespie-Lynch and colleagues found that a training designed with autistic university students improved knowledge, reduced explicit stigma and improved attitudes toward inclusion more than one written by nonautistic professors alone (Gillespie-Lynch et al., 2022). The 2025 finding that disclosure improved rapport cuts both ways, since Botha and Frost found that being open about autism was associated with lower wellbeing, a sign that disclosure helps only where the listener is prepared to receive it. The civic task, then, is to prepare the listener, teaching the majority to read autistic communication, letting autistic people design and lead that teaching, judging people by the substance of what they say, and making room in congregations, workplaces, neighborhoods and schools for autistic people to find one another.

7NEUROTYPICAL ABUSE: A PROPOSED CONSTRUCT

When a person who held everything together for decades can no longer hold it, the collapse can carry evidence of something that existing vocabulary names only in pieces. This article proposes the term neurotypical abuse for it, offered as a theoretical construct for testing and not as an established clinical or legal category.

PROPOSITION 3 · NEUROTYPICAL ABUSE.Neurotypical abuse describes the cumulative harm produced when neurotypical norms of communication, sociability and sensory tolerance become compulsory conditions of participation, and when autistic people are repeatedly corrected, disciplined, pathologized, excluded or abandoned for failing to perform those norms, so that sustained masking becomes the price of belonging and the resulting harm is attributed back to the autistic person.

The construct is defined by five conditions, each observable. A neurotypical norm is enforced as a condition of participation in a family, school, workplace, clinic or community. Departures from the norm are met with repeated correction or sanction. The enforcing party holds more relational or institutional power than the autistic person. Harm accumulates over time, in exhaustion, burnout, distress or suicidality. And the harm is attributed back to the autistic person as a deficit, a disorder or a choice. Intent is absent from the list on purpose. Johan Galtung’s concept of structural violence describes harm built into social arrangements with no individual actor required (Galtung, 1969), and Carly Parnitzke Smith and Jennifer Freyd define institutional betrayal as harm an institution causes to a person who trusts or depends upon it (Smith & Freyd, 2014). Neurotypical abuse belongs in that family. The word abuse is chosen deliberately, because the pattern it describes has the structure of abuse, repeated harm inside a relationship of dependence and unequal power, followed by blame placed on the one harmed, and the burden now falls on researchers to test whether the pattern holds.

The construct overlaps with established concepts without collapsing into any of them, and the distinctions matter.

The construct can be tested with instruments already in use. Camouflaging can be measured, minority stressors can be measured, unmet support needs can be counted, and the attribution of harm can be read in clinical records, performance reviews and family accounts. If neurotypical abuse is a useful construct, measures of compulsory norm enforcement should predict burnout and suicidality beyond what autistic traits and co-occurring conditions predict, and reducing enforcement, through accommodation granted without proof, for example, should reduce the harm. The Dutch longitudinal finding adds a specific prediction: where enforcement stays high, the adults whose capacity to mask declines should show the steepest rise in difficulties. If those predictions fail, the construct should be revised or retired.

8META MACRO SOCIAL WORK: FOLLOWING HARM ACROSS SYSTEMS

PROPOSITION 4 · META MACRO SOCIAL WORK.Meta Macro Social Work is a framework that takes as its unit of analysis the relationship between a person and their environment across six levels, person, relationship, institution, professional system, culture and social norm, and asks two questions at each level: who is required to change, and to whom is the resulting harm attributed. Harm is assessed cumulatively, as the sum of adaptation demanded across all six levels.

The framework builds on a long tradition. Urie Bronfenbrenner described human development as nested within systems from the family outward to culture (Bronfenbrenner, 1979), and Alex Gitterman and Carel Germain’s life model taught social work to see problems as poor fit between person and environment (Gitterman & Germain, 2008). Klein and Macoun have already applied person-environment fit theory to camouflaging, arguing that autistic people achieve fit by suppressing autistic traits and displaying neurotypical ones, which succeeds between people and fails within the person, and that fit should instead be achieved by environments adjusting to authentic autistic traits (Klein & Macoun, 2025). Meta Macro Social Work extends that argument in three ways. It separates the professional system and the social norm into levels of their own, since the Golden Autism case shows that clinics and unwritten rules act on people with a force distinct from the institutions that employ or educate them. It measures the direction of adaptation cumulatively, asking at each level whether the person or the environment was required to change and summing the demand across all six. And it tracks attributional reversal, the point at which harm produced by the environment is recorded as a property of the person. The table below traces Golden Autism across the six levels.

Read down the middle column and the cumulative weight becomes visible. No single level asks for very much, and every level asks, so the person carries the total, and every level has a way to record the result as the person’s own failure. Social work has long studied people in environments. Meta Macro Social Work studies where the burden of adaptation lands across environments, and that is the contribution this article offers to the field.

9WHEN CARE BECOMES CRUELTY

Put these findings together and the shape of the cruelty becomes visible in practices commonly understood as care. Requiring eye contact asks a person to spend attention on the performance of listening instead of on listening. Rewarding socially typical behavior teaches concealment and makes the person dependent on a performance that may hold for years and fail at the worst moment. Treating withdrawal as rejection turns recovery into an offense. Requiring disclosure before accommodation makes support contingent on exposure, and Botha and Frost found outness associated with lower wellbeing. Interpreting autistic distress as attitude, and burnout as personal failure, completes the attributional reversal. Treating successful masking as evidence that disability is mild punishes the most skilled adapters with the least support. Teaching coping strategies without changing harmful environments is adaptive iatrogenesis. Pathologizing autistic communication ignores the evidence that it works among autistic people. Requiring autistic people to explain their needs again and again, and to repair every misunderstanding that differences in communication create, places the whole labor of the double empathy problem on one side. And treating institutional norms as neutral hides every one of these choices behind the word normal.

The workplace numbers are stark. In the United Kingdom, the Office for National Statistics found in 2020 that 22 percent of autistic adults were in any kind of employment, the lowest rate of any disability group reported, against roughly half of disabled people overall and more than 80 percent of nondisabled people (Office for National Statistics, 2021). The accommodations that might keep a person in work are cheap. Across employer surveys collected from 2019 through 2024, the Job Accommodation Network found that 61 percent of employers reported accommodations that cost nothing, and among those with a one time cost the median was $300 (Job Accommodation Network, 2025). A quiet room, a changed schedule and an agenda sent the day before cost less than the meeting called to approve them, and the process that guards them asks the exhausted employee to prove, in writing and in person, the very limits they spent a whole career hiding.

The clinic can deepen the wound. One in four autistic adults in the Dutch study carried a psychiatric misdiagnosis first, and roughly one in eight was told they had a personality disorder, a label that places the problem in the person’s character (Kentrou et al., 2024). An adult who arrives in burnout, depleted, shut down, unable to cook or answer email, can be treated for depression with medication and with techniques aimed at changing their thoughts, while the load that produced the collapse goes unexamined and unrelieved. Cassidy’s team found that the number of unmet support needs was associated with suicidality in autistic adults beyond the effects of diagnosed mental health conditions (Cassidy et al., 2018), which means every support withheld, delayed or wrapped in paperwork adds weight to a scale that may already be tipping.

Set the pieces inside one life and the picture is terrifying. A person spends forty or fifty years performing and is praised for it, and the performance holds, until the day it no longer can. The person burns out, is told by a clinician that the trouble is a personality disorder or a mood disorder, is told by an employer that support requires disclosure, documentation and a meeting, is told by family that the change looks like a choice, and joins a population in which, in the British data, fewer than one in four hold a job, and in which, in the Swedish registers, adults without intellectual disability die by suicide at more than nine times the odds of their neighbors. Nobody in that story set out to cause harm, and every one of them offered help. The harm gathers in the space between their good intentions, and it lands on one person, who is usually alone when it does.

10TOWARD CONVIVIAL CARE: CORRECTIVE INSIGHTS AND NEXT STEPS

Illich hoped for tools for conviviality, arrangements that enlarge what people can do for themselves and one another without waiting for a professional to authorize it (Illich, 1973). For an autistic adult coming out of burnout those tools are plain, a friend who accepts a text in place of a phone call, a job that lets a person work in a quiet room without filing a form to earn it, a family that asks what the silence means before deciding what it proves, a clinician who treats the person’s account of their own nervous system as evidence. Each costs almost nothing, and each returns to the person some of the authority to say what they need. Read through Meta Macro Social Work, each is also a transfer of the burden of adaptation from the person back to the environment.

Recognition itself is a form of repair. Autistic adults interviewed by Louise Bradley and colleagues described camouflaging less after diagnosis and among people who understood and accepted them (Bradley et al., 2021), and a 2025 study of autistic and ADHD adults found that accepting one’s diagnosis was associated with less masking and better quality of life (Wurth et al., 2025). Both studies are cross-sectional, and both point the same direction, toward environments where the mask can come down safely before it falls down on its own.

Families and friends can read burnout after years of compelled adaptation as a matter of human capacity, and respond by lowering demand before asking for explanation, accepting the unmasked person as the same person they have always loved, learning the person’s preferred ways of communicating, whether text, writing or quiet company, and treating rest and withdrawal as recovery. The question that opens the most doors is what would make today easier, asked without a deadline for the answer.

Employers can grant ordinary accommodations on request, without requiring a diagnosis or a medical file, since most cost nothing and the median of the rest is a few hundred dollars. Better still, they can build quiet workspace, written agendas, flexible hours, remote options and clear instructions into the workplace for everyone, so that no one has to disclose to receive them, and they can judge people by the substance of their work, the same standard that erased the bias in Sasson’s transcripts.

Clinicians can ask every autistic patient directly about camouflaging and unmet support needs, since both travel with suicide risk, and can watch most closely for the moment masking begins to fail, since the Dutch data suggest that is when difficulties climb. They can consider autism when adults present with burnout, treatment resistant depression, anxiety or a suspected personality disorder, especially women, integrate developmental history, longitudinal functioning, masking, compensatory strategies and the patient’s own account with any instrument, and measure progress by the patient’s rest, safety and self understanding, never by how well the patient has learned to look typical.

Communities and institutions, including congregations, schools, neighborhood groups, social work and public health, can offer acceptance training designed and led by autistic people, host regular spaces where autistic adults can meet one another, fund peer support and adult assessment so that diagnosis does not depend on money or luck, and add burnout education and unmet support needs to suicide prevention work, as Raymaker’s team recommended.

Researchers can test the propositions offered here, following late-diagnosed adults across many more than two waves to learn whether compulsory norm enforcement predicts burnout and suicidality beyond autistic traits and co-occurring conditions, whether the loss of capacity to mask marks the turn toward crisis, and whether changing the environment reduces harm. Those studies would move this article’s theory into evidence, or correct it.

Late-diagnosed adults can treat rest as part of recovery, unmask first in the places and with the people where it is safe, seek out other autistic adults, and trust their own long record of experience as evidence, because the research now says what many of them have felt for decades, that the performance was expensive and the exhaustion was earned.

A late diagnosis gives a person a history. The adults who arrive at it, and the people who love them, deserve to see that history clearly, to read the collapse as the evidence it is, and to build care that asks less of the people who have already given the most.