Autism in Women, Why So Many Are Diagnosed in Adulthood

A great many autistic women move through childhood, adolescence and much of adult life without ever being identified. They do not match the picture most people carry of autism, they learned to compensate very early, and the clinical tools were built on samples that were overwhelmingly male. The result is a diagnosis that often arrives at 25, at 35, at 50, and almost always after a long detour through other explanations.

What follows covers the female autism profile, how the signs shift across life stages, the diagnoses that usually come before the right one, and what an autism assessment involves in Quebec.

Why autism in women is still underdiagnosed

Autism was first described in the 1940s, from observations of what were almost entirely boys. For decades afterward, the criteria, questionnaires and observation scales were calibrated on a male presentation. A woman whose traits expressed themselves differently could meet the diagnostic criteria without ever triggering clinical suspicion.

The numbers tell the rest. In the United States, the Centers for Disease Control (CDC) surveillance network has tracked the share of 8-year-olds identified as autistic rising from 1 in 150 in 2000, to 1 in 68 in 2010, to 1 in 31 in 2022, in the report published in April 2025. The CDC ties that rise first to better identification, particularly among very young children and in groups that had been underserved, along with more systematic screening and better access to evaluation. Less stigma attached to discussing these traits is also often put forward. In Quebec, the Institut national de santé publique reports the same upward trend among people aged 1 to 24.

That same US report counts 3.4 boys for every girl among 8-year-olds identified in 2022, down from 4.2 in 2018. The CDC warns against reading the narrowing too quickly. Over the same period the absolute gap between boys and girls kept widening. A ratio that shrinks does not mean girls are now being found.

A 2017 meta-analysis in the Journal of the American Academy of Child and Adolescent Psychiatry compared studies that screen a whole population against studies that recruit only people who already carry a diagnosis. In already-diagnosed samples, the ratio runs at roughly four boys to one girl. Among children who genuinely meet criteria, it sits closer to three to one. The authors concluded that girls who meet criteria for autism are at disproportionate risk of never receiving a clinical diagnosis.

The gap between four to one and three to one accounts for a whole population of women who meet the criteria and fall through.

Three mechanisms stack on top of each other. The presentation differs. Masking hides whatever remains visible. And when a difficulty finally does get named, it usually gets attributed to something else.

The female autism profile, what actually differs

There is no male autism and female autism. There is a spectrum, and a statistical tendency in how traits express themselves and how the people around them read those traits.

Special interests do not raise flags

Intense, focused interests are part of the diagnostic criteria. In a boy, they often land on subjects that stand out, like transit schedules or mineral classification. In a girl, they frequently land on socially expected subjects, like animals, books, a television series, a musician, psychology.

The criterion is met either way. What counts is the intensity, the structure and the function of the interest, not its topic. But when the topic looks ordinary, nobody flags it.

Social skills are learned rather than intuitive

Many autistic women have friendships, sometimes an active social life. The difference is in the method. Where social reading is automatic for most people, here it is analyzed, decoded and rebuilt. Some observe at length before acting, model themselves on a popular classmate, rehearse conversations in advance, and replay interactions for hours afterward.

The social competence is real. It simply costs far more energy to produce.

Sensory differences are turned inward

Heightened sensitivity to noise, light, texture or smell is common. The difference shows up in the response. Instead of a visible reaction, you tend to see quiet avoidance, forced tolerance, and then a collapse in private once the front door closes.

Emotional overload goes inward too

Overload more often shows up as withdrawal, temporary loss of speech, anxiety or distress turned on the self than as visible opposition. A child who disrupts gets referred. A quiet, conscientious, slightly separate girl gets described as shy, sensitive or a daydreamer, and the file closes there.

Autism masking, the central mechanism

Masking, also called camouflaging, covers the conscious and unconscious strategies used to hide autistic traits and meet social expectations. Forcing eye contact through the discomfort. Suppressing a repetitive movement that would have been regulating. Copying other people’s facial expressions and speech rhythm. Preparing conversational scripts. Building a separate persona for each setting.

It is the single factor that best explains why diagnosis arrives so late for women, and it is also the most expensive one.

Sustained over years, masking exhausts. It produces chronic fatigue out of all proportion to actual activity, social anxiety, a persistent sense of being an impostor, and real difficulty answering the question of who you are when you are not performing. This is what is meant by autistic burnout. It tends to hit at the points where load spikes, such as starting university, a first demanding job, or the arrival of a child.

Many women seek an assessment at exactly that moment, not because they are wondering whether they are autistic, but because a strategy that had worked for twenty years has just stopped working.

Signs of autism in women by life stage

The presentation does not read the same way at seven, at fifteen and at forty. That is one reason a serious assessment always reconstructs the full trajectory rather than photographing the present moment.

Autism in girls

A child often described as well behaved, serious and quiet. One close friend at a time, in an intense and exclusive relationship. Play that faithfully reproduces observed scenarios rather than freely invented ones. Marked sensitivity to clothing, tags, cafeteria noise. A strong need for clear rules and for fairness. Often solid academic results, which is frequently enough to close the question.

In adolescence

This is where the gap becomes noticeable. Social codes get more complex and shift faster, and the learned strategies stop keeping up. What shows up is exhaustion by end of day, withdrawal, anxiety settling in, sometimes eating or sleep difficulties. The link to an autistic profile is rarely made at this stage.

In young adulthood

University or a first job remove the structure that had been carrying everything. The load of organizing, planning and improvising socially rises all at once. Many women describe a performance that holds, but at the cost of full recovery every evening and every weekend.

Autism in adult women after 40

Assessment requests at this age are rising. The trigger is usually external. A child receives an autism diagnosis and the mother recognizes her own history in the report. Or a long stretch of overload finally brings the mask down. The question becomes retrospective, rereading forty years of history against a framework that, this time, explains things.

The diagnoses that usually come first

Very few autistic women arrive at an assessment with no history. Most have already been given one or more other explanations.

In adults whose autism was not identified in childhood, autistic traits are frequently read as belonging to some other psychiatric condition, long before autism is considered at all. The diagnoses that come up most are anxiety disorders, social anxiety among them, mood disorders, personality disorders, eating disorders and ADHD.

A 2024 study in eClinicalMedicine measured how often this happens. Across 1,211 autistic adults, 24.6% reported at least one previous psychiatric diagnosis they considered to have been a misdiagnosis. Personality disorders were the most frequent of those, followed by anxiety disorders, then mood disorders.

Among autistic women the figure rises to 31.7%, against 16.7% of men.

Borderline personality disorder, the hardest differential

Borderline personality disorder is an important differential diagnosis for some autistic people, because a number of features overlap, particularly around social interaction, emotional regulation and interpersonal relationships. Qualitative studies have documented pathways in which autistic adults carried a borderline personality disorder diagnosis, sometimes for years, before autism was recognized.

It is also why that diagnosis comes up so often for women specifically. The relational difficulties and emotional dysregulation that follow years of masking and mutual misunderstanding look, from the outside, a great deal like what borderline personality disorder describes.

An earlier diagnosis is not automatically an error

A psychiatric diagnosis given before an autism diagnosis should not be treated as an error by default. Anxiety, depressive, eating and personality disorders can be present at the same time as autism, and often are.

Anxious and depressive symptoms in particular deserve to be taken seriously in their own right. They frequently accompany autism, in women and men alike, and recognizing the autistic profile does not make them go away. They need their own follow-up.

So the problem is not having received another diagnosis. It is the analysis stopping there. The symptom gets named, treated, and the question of what has been generating it since childhood never gets asked.

What the differential assessment has to settle

The assessment works out whether what is being observed is better explained by autism, by a psychiatric condition, by both conditions coexisting, or by an interaction between them. Those four answers do not lead to the same follow-up. Dr. Diab puts that untangling at the centre of her autism assessments, working through the diagnoses already on file.

Quebec data points the same way. A 2025 study in the Journal of Neurodevelopmental Disorders, covering 2,799 adults who received a first autism diagnosis in Quebec between 2012 and 2017, reconstructed their diagnostic pathways from health administrative data. Anxiety and depressive disorders were among the conditions most often recorded before the autism diagnosis. More importantly, the pathways varied enormously from one person to the next, to the point that the authors identify five distinct types.

There is no typical route to a late autism diagnosis. An assessment therefore has to reconstruct the developmental history in full.

ADHD, giftedness and autism overlap

Three combinations come up frequently in women assessed as adults.

ADHD and autism frequently co-occur. Since the DSM-5, both diagnoses can be given to the same person, which was not permitted before. In women, ADHD often presents in an inattentive form with no visible restlessness, and it too gets identified late. An assessment that looks at only one of the two profiles risks seeing half the picture. The subject is covered on the ADHD assessment page.

Giftedness and autism also overlap. High intellectual potential provides considerable compensatory resources. A gifted autistic woman can construct, through analysis alone, a social functioning that holds until the load becomes too heavy. It is a frequently missed profile. See giftedness assessment and, for the combination of the two, twice-exceptional.

The principle is the same in all three cases. When two profiles overlap, they mask one another, and the clinical picture only makes sense if the assessment examines them together.

What free online autism tests can and cannot do

Free online questionnaires are often the first step, which is understandable. Some are genuinely validated screening instruments, such as the autism quotient questionnaire and the camouflaging traits questionnaire.

What they can do is give shape to an intuition. They provide vocabulary, they indicate that a line of thinking is worth pursuing, and they help prepare for a first appointment.

What they cannot do is establish a diagnosis. A high score can reflect anxiety, trauma, ADHD or a period of burnout. A low score says nothing reassuring about someone who has been masking effectively for thirty years, since masking targets exactly the behaviours the questionnaire is trying to measure. These tools are a starting point, never a conclusion.

How an autism assessment works for adult women in Quebec

In Quebec, diagnosing autism spectrum disorder is a reserved activity. It can be issued by a physician, and by psychologists and neuropsychologists who are members of the Ordre des psychologues du Québec. Since 7 November 2024, when the act amending the Professional Code came into force, section 37.1 refers to diagnosing mental disorders, intellectual disability and neuropsychological disorders, where it previously referred to evaluating them. The Ordre describes the change as formal recognition of a practice that already existed, not as a new activity.

An assessment with an adult woman normally involves an in-depth clinical interview, reconstruction of the developmental history, standardized instruments, and an integrated analysis that accounts for previous diagnoses. The full process is described on the autism assessment page and, more broadly, on the neuropsychological assessment page.

Two things are worth knowing before starting.

The first, standardized observation instruments have documented limitations with women who mask. An observation scale measures what is visible during the session, which is precisely what thirty years of compensation have trained into invisibility. This is why the developmental history, the person’s own account and, where possible, an outside perspective on childhood carry as much weight as the testing itself. An assessment resting on direct observation alone would miss part of the picture.

The second, developmental history can be hard to reconstruct in adulthood. Parents are not always available, and memories reorganize themselves. Report cards, teacher comments, photographs, health records and siblings’ accounts all have real clinical value. They are worth gathering before the first appointment.

On the financial side, an assessment in private practice may be partly covered by a group insurance plan, and the fees may qualify for the medical expense tax credit. It is worth checking with your insurer before beginning.

What changes after a late diagnosis

Being diagnosed at 35 or 50 does not change the person you were the day before. It changes what she understands about herself.

The most commonly reported effect is the reread. Episodes attributed for years to a lack of willpower, to being too sensitive, to some flaw of character finally receive a coherent explanation. That reattribution has clinical value in itself, because it lifts a weight of guilt that has often been carried since childhood.

Then there is the practical side. The report documents a specific cognitive profile, with its strengths and its points of vulnerability. It allows work or study environments to be adjusted, supports a request for accommodations, and above all directs psychological follow-up toward what is actually driving things rather than toward the surface symptom.

And there is the relational side. Naming what is happening makes it possible to explain it to a partner, a family, an employer, and to ask for adjustments without having to justify them each time.

A diagnosis stays a framework, not a label and not the end of the road. For many women it is the first one that holds.