i autistic 我自闭

What happens when an autism diagnosis was correct when it was made, but no longer describes the person standing in front of us today?

This is the question I believe the autism community, researchers and future diagnostic manuals need to confront. I am not trying to discredit the autism professionals who diagnosed us. A diagnosis can be entirely reasonable based on the evidence available at a particular point in a person’s development.

My argument is different: a diagnosis that was valid then should not automatically become an irreversible description of who that person must remain for the rest of their life.

Quick summary

1. I am not arguing that earlier autism diagnoses were mistakes. People change. A clinician may have correctly diagnosed someone at one stage of life while that person’s present clinical presentation is substantially different decades later.
2. The present diagnostic framework needs a clearer “out”. There should be a recognised way to record that someone has a genuine history of autism but no longer meets the full current clinical criteria or experiences clinically significant impairment attributable to autism.
3. This matters outside the clinic. An autism diagnosis can affect how employers, insurers and other institutions perceive a person. Someone who no longer requires autism-related accommodations may reasonably want their present abilities assessed rather than having assumptions made from an old label.
4. “Lifelong developmental condition” should not automatically mean “lifelong disability at the same clinical level”. A developmental history can remain lifelong even when a person’s abilities, difficulties and support needs change enormously.
5. Research should study people who improve substantially. Not to promise a cure, but because understanding successful developmental trajectories could reveal ways of expanding opportunities for other autistic people.

I am not questioning the professionals who diagnosed me

I want to make this point very clearly. If a professional assessed me years ago and concluded that I met the criteria for autism, I am not saying that professional was incompetent or that the diagnosis was inaccurate. They assessed the person I was at that point in time. But I have continued living, learning and developing since then.

That creates a problem when a diagnostic system is excellent at answering the question, “Has this person ever had the characteristics of autism?” but much less comfortable answering the equally legitimate question, “What is this person’s clinical condition now?” Those are not the same question.

The autism diagnostic system needs an “out”

DSM-5 deliberately allows autism characteristics to be established “currently or by history”. During its development, proposed criteria dealing with recovery were considered but ultimately excluded.

I understand why autistic advocates were concerned. They feared that people who had learned to cope better could suddenly lose accommodations, services and legal protections that they still genuinely needed. That was a legitimate concern, but solving that problem should not require creating the opposite problem.

We should protect autistic people who continue to need a diagnosis without trapping people who no longer want or clinically need autism to function as their present-day label.

These objectives are not mutually exclusive.

A future revision of the diagnostic framework should provide a recognised status for situations such as:

• autism spectrum disorder – historical;
• autism spectrum disorder – nonpersistent;
• previous autism spectrum disorder, no longer meeting full current criteria; or
• another carefully researched term that communicates both developmental history and present clinical reality.

I am not attached to the terminology. I am attached to the principle.

There must be a door in as well as a door out.

Why would anyone want to leave the diagnosis behind?

For some autistic people, the diagnosis is extremely valuable. It provides self-understanding, community, accommodations and access to services. They should keep it.

But different people have different interests. Some of us reach a stage where we require few or no autism-related accommodations. Some no longer wish to make autism an important part of our identity. Some wish to pursue careers involving management, negotiation, customer service, leadership or other intensive interpersonal responsibilities. In those situations, an old diagnosis can become baggage.

An employer should assess whether I can manage people by examining whether I can actually manage people. They should not infer my abilities from what they imagine an autistic person can or cannot do. This principle is already consistent with Singapore’s fair-employment approach: employees and jobseekers are supposed to be assessed on merit rather than disability-based assumptions.

But laws and guidelines cannot instantly remove stereotypes from people’s minds. A diagnosis can become shorthand:

• “poor people skills”;
• “cannot handle customers”;
• “cannot manage staff”;
• “too rigid”;
• “will require accommodations”; or
• “not suitable for this role”.

The individual may be none of those things. If someone has developed the abilities required for the job, their actual abilities should matter more than expectations attached to a diagnosis from years earlier.

Insurance makes this more than an identity debate

Insurance provides an especially concrete example of why present clinical status matters.

The Life Insurance Association of Singapore’s own 2024 guide to medical underwriting includes cases involving autistic applicants. The case illustrates that autism-related information can affect underwriting outcomes for different forms of coverage, and it specifically contemplates submitting fresh medical evidence for reassessment when the person’s condition is clearer at an older age. This is precisely why reassessment matters.

An insurer is entitled to assess risk according to applicable law, its underwriting standards and the evidence before it. But the evidence should be capable of being updated. If my functioning at age 40 is materially different from my functioning at age 15, there should be a clinically credible way of documenting that change.

A historical diagnosis should remain part of an honest medical history where relevant. Reassessment is not permission to give a false answer when someone is specifically asked whether they have ever received a diagnosis. But neither should “diagnosed once” automatically mean “clinically identical forever”.

We need to question what “lifelong” actually means

Autism is routinely described as a lifelong neurodevelopmental condition. I think this sentence has acquired far more philosophical weight than it deserves. There are several very different propositions hidden inside it:

• a person’s developmental history cannot be erased;
• some neurological differences may remain throughout life;
• some autistic traits may remain throughout life;
• a person will always meet every clinically significant requirement for autism;
• a person will always require autism-related support; and
• a person will always be substantially limited in the same areas.

The first proposition does not automatically prove all the others. A lifelong developmental history is not necessarily the same thing as lifelong clinical disability.

A person does not have to become perfectly NeuroTypical before we acknowledge meaningful developmental change. They may remain neurologically unusual. They may retain autistic traits. Their childhood history remains real. But they may also reach a point where they can form relationships, navigate unfamiliar social environments, manage employees, work with customers, cope with changes, understand other people’s intentions and live independently without the clinical impairments that once dominated their life.

If that happens, science should describe what actually happened rather than forcing the person back into what our theory says ought to have happened.

We already know that autism diagnoses are not perfectly persistent

This is not merely philosophical speculation. A 2023 study published in JAMA Pediatrics followed 213 children who had received clinical autism diagnoses between 12 and 36 months old.

At ages 5 to 7, 79 of those children (37.1%) did not continue to meet autism diagnostic criteria in the research assessment. The researchers called this nonpersistent autism spectrum disorder. This does not mean that 37% of autistic people can be cured, but it demonstrates something equally important: diagnostic persistence is an empirical question, not an article of faith.

Once researchers have documented people who genuinely met autism criteria at an earlier age and did not meet them later, “autism diagnoses can never become nonpersistent” is no longer a defensible universal claim. The interesting scientific question becomes: why?

My own development is why I care about this

My position comes from lived experience rather than theory alone. In June 2003, I scored 44 on the Autism Spectrum Quotient. In December 2024, I scored 23.

An AQ score is not a diagnosis, and I am not presenting these numbers as proof that autism disappeared. I present them because they are consistent with something much more important: my actual lived experience changed.

Before 2003, my emotional world seemed relatively elementary to me. Between approximately 2003 and 2006, my emotional experience became much richer and considerably more complicated. Later I began experiencing conflicting emotions towards the same person, something that previously seemed alien to how my mind worked.

My understanding of other people also changed.

Things that once required deliberate intellectual analysis increasingly became intuitive. I became better at sensing someone’s character, intentions and likely behaviour from ordinary interaction.

I participated in increasingly complicated social environments. I have worked for about one and a half years in a demanding job involving multitasking, customer service and management responsibilities without disclosing autism.

Applied Drama later helped me explore social situations in a structured but realistic environment. I am not claiming that Applied Drama treats autism. I am saying that it was useful to me.

I also dealt with chronic fatigue and other health issues over the years. Improving my physical health gave me greater cognitive and emotional capacity. I do not claim that treating fatigue treats autism; health problems should be treated as health problems.

The accumulated result is that I can now handle many situations that I previously could not.

This does not feel like masking

The standard explanation for an autistic person who becomes socially capable is often masking. I know what deliberate compensation feels like. It requires observation, intellectual processing and conscious control. That is not the whole of what happened to me.

Some things that once required conscious reasoning became increasingly automatic. I did not merely memorise a larger rulebook. My perception of people changed. Even my dreams became more socially complex, unpredictable and nonlinear over the years. My internal experience changed as well. That is why I disagree with explanations that reduce every improvement in an autistic person’s social abilities to better camouflage.

Sometimes compensation undoubtedly occurs. Sometimes masking occurs. But science should remain open to the possibility that genuine learning and developmental change occur as well.

Development does not have to mean becoming NeuroTypical

I do not believe that human development gives us only two possibilities: “still autistic in exactly the same way” or “completely transformed into a NeuroTypical person”. That is a false choice.

Someone who learns a second language as an adult may never speak exactly like a native speaker, yet they can become completely fluent enough to work, joke, argue, negotiate and build relationships in that language. I suspect some aspects of autistic development may be similarly complicated. An autistic developmental pathway may remain different while nevertheless reaching abilities that were absent earlier.

I am not presenting this as established neuroscience. It is the hypothesis that best fits my experience and deserves research rather than ridicule.

Study success instead of merely documenting deficits

This is where I believe autism research has a major opportunity. Research understandably concentrates on difficulties: early identification, disability, mental health, unemployment, sensory problems, communication barriers and support needs.

All of these subjects matter, but there is another scientifically valuable population: people who were substantially impaired earlier in life and later achieved unexpectedly strong outcomes. Find, study and compare them. Ask what changed.

Was it maturation? Education? Motivation? Cognitive development? Better physical or mental health? Structured practice? Changes in environment? Particular interventions? A combination? Something we have not yet considered?

Most importantly, distinguish between people who are successfully masking and people whose underlying abilities appear to have genuinely changed. There may be no single pathway. Different people may arrive at similar outcomes through completely different mechanisms. That is exactly why research is needed.

The purpose is not to create a new promise that every autistic child can “recover”. That would simply replace one rigid ideology with another. The purpose is to learn from unusually successful developmental outcomes and ask whether any part of that knowledge can expand opportunities for other autistic people.

Success does not invalidate disability

Whenever this subject is discussed, someone reasonably worries that recognising nonpersistent autism will be used against people who remain disabled; it should not be.

If someone requires lifelong substantial support, they should receive lifelong substantial support. If someone benefits from identifying as autistic, they should be free to identify as autistic. If someone needs workplace accommodations, those accommodations should not disappear merely because another autistic person achieved a different outcome.

My development proves nothing about what another individual “should” be capable of. But the reverse is also true. Another person’s continuing disability does not prove that my development did not happen.

Autistic people constantly ask society to recognise diversity within the autism spectrum. That principle must include diversity of developmental outcomes as well.

I am neither proud nor ashamed of autism

My position for many years has been that I am neither proud nor ashamed of autism. I do not regard autism as a moral achievement. Neither do I regard it as a moral failure.

I am pragmatic. If an autism diagnosis helps someone understand themselves, obtain accommodations and build a better life, use it. If someone no longer needs those benefits and the diagnosis instead creates practical disadvantages, I believe they should have a legitimate route to have their current clinical status reassessed.

Identity should remain voluntary. Medical history should remain truthful. Current clinical status should remain capable of change.

What I want changed

I am asking future reviewers of autism diagnostic criteria to confront a gap left unresolved by the present system.

• Do not invalidate earlier diagnoses.
• Do not remove support from people who still need it.
• Do not tell autistic people who embrace their identity that they must abandon it.
• But provide an equally legitimate pathway for people whose lives move in another direction.

Allow qualified clinicians to say, when the evidence supports it: This person had a valid autism diagnosis and developmental history. Their present functioning has changed substantially, and they no longer meet the full current clinical criteria or experience clinically significant impairment attributable to autism.

Then let employers, insurers and other institutions consider that updated evidence rather than pretending that a person’s clinical condition was frozen on the day of their first assessment.

A diagnosis should describe reality. Reality should not be forced to remain consistent with a diagnosis.

My challenge to autism research

I do not know how many people have followed developmental paths similar to mine; that uncertainty is precisely the problem. If researchers assume in advance that autism must always remain clinically persistent, people who no longer fit the expected pattern become anomalies to explain away rather than a population to investigate.

I want that to change.

• Study the people who exceeded expectations.
• Study the autistic adults who became managers, negotiators, teachers, counsellors, facilitators, entrepreneurs and highly capable social participants.
• Study those who say that relating to people became increasingly natural rather than increasingly well rehearsed.
• Study those whose support needs fell dramatically.
• And study those who genuinely no longer meet current diagnostic criteria.

Not because every autistic person should become like them, but because every autistic person deserves the benefit of everything we can learn from them.

Note: This article combines personal experience, commentary and discussion of published research. Individual developmental outcomes vary greatly. Screening questionnaires such as the AQ are not diagnostic tests, and professional reassessment is required for clinical conclusions. Nothing here recommends concealing a previous diagnosis where disclosure of medical history is lawfully required or contractually requested.