Autism-pedia / Diagnosis / Rising prevalencebrianspringer.com
Diagnosis

Rising prevalence

How autism is counted, how far the numbers have risen in the US, the UK, Scandinavia and elsewhere, what the rise is made of, and what the few studies with a steady method say about whether the underlying rate has changed.

How scientists count autistic people, why the numbers have gone up so much, and what the best studies say about whether autism itself has become more common.

Last reviewed 12 October 2026 · one person's reading of the research · 44 research and 8 community sources

How autism is counted Established

There is no blood test or scan for autism, so a prevalence figure is a count of people who meet a definition, found by some method. Studies use four main methods: registers and insurance claims that hold existing diagnoses, review of health and school records, surveys that ask parents, and screening a whole population and assessing the children who screen positive.

The method moves the number. A meta-analysis of 79 studies from 1994 to 2019 pooled them at 0.72%. Studies that reviewed records gave 1.22%, and studies based on health insurance data gave 0.35%. Estimates were also higher in North America and in high-income countries.1 A review of 99 estimates published from 2012 to 2021 found a median of 100 per 10,000 (1%), with single estimates from 1.09 to 436 per 10,000.2

The US CDC's Autism and Developmental Disabilities Monitoring (ADDM) Network reviews records of 8-year-olds in selected areas. A child counts if the records show a diagnosis of autism in an evaluation, eligibility for special education under autism, or an autism diagnostic code.3 A US national household survey (the National Health Interview Survey) gave 3.42% of 3 to 17 year olds in 2021 and 2022.4 When researchers screened children aged 7 to 12 across one South Korean community, they estimated 2.64%, and two-thirds of the autistic children they found were in mainstream schools, undiagnosed.5

There is no blood test or brain scan for autism. So a count of autistic people depends on how you look for them. Some studies count people who already have a diagnosis. Some read through doctors' and school records. Some ask parents. Some test every child in an area.

The way of counting changes the answer. A big review put many studies together and got 0.72%, or about 7 in every 1,000 people. Studies that read records got 1.22%. Studies that used health insurance data got only 0.35%. Numbers were also higher in North America and in richer countries.1 Another review found that most studies landed near 1 in 100. Single studies ranged from about 1 in 10,000 to about 4 in 100.2

In the US, a network run by the CDC reads the records of 8-year-olds in chosen areas. A child counts if a record shows an autism diagnosis, school help for autism, or a medical code for it.3 A US home survey found autism in 3.42% of children aged 3 to 17 in 2021 and 2022.4 In one South Korean community, researchers tested children aged 7 to 12. They estimated 2.64%. Two out of three autistic children they found were in ordinary schools with no diagnosis.5

Suggest a correction
Established

A meta-analysis and a systematic review agree that the method of counting changes the estimate, though they pool many of the same studies. The single surveys and the Korean screening study illustrate this; they do not carry the tier on their own.

Word
Prevalence

The share of a population that has a condition at a given time. Incidence is the rate of new diagnoses over a period.

Why counts of diagnoses run low or high

A register or insurance count can only find people who have already been diagnosed. Screening finds people who were never referred, but depends on who agrees to be assessed. Record review falls in between and depends on what the local clinics and schools wrote down.

See also
Established

Many studies, pulled together in big reviews, agree on this. The way you count changes the number you get.

Word
Prevalence

How common something is: the share of people who have it at one time.

See also

How far the numbers have risen Established

In the ADDM Network, 6.6 per 1,000 8-year-olds (1 in 152) met the definition in 2002, across 14 sites.6 In 2020 the figure was 27.6 per 1,000 (1 in 36) across 11 sites.7 In 2022 it was 32.2 per 1,000 (1 in 31) across 16 sites, ranging from 9.7 in Laredo, Texas, to 53.1 in California. In 2022 the rate was lower among White children (27.7 per 1,000) than among Asian or Pacific Islander, American Indian or Alaska Native, Black, Hispanic and multiracial children, though the American Indian or Alaska Native estimate was imprecise. Children born in 2018 had been diagnosed or found eligible for autism special education by 48 months at 1.7 times the rate of those born in 2014.3

In UK primary care records, new autism diagnoses rose by 787% from 1998 to 2018. The rise was steepest among adults and steeper in females than in males.8 In Swedish national registers covering people born from 1988 to 2016, registered prevalence rose by about 800% from 2001 to 2020.9 Of the 2.76 million people born in Sweden from 1985 to 2020, 2.8% had received an autism diagnosis by 2022, the younger ones after only a few years of follow-up.10 Pooled across countries, estimates rose over time.1

In the US network, about 1 in 152 8-year-olds counted as autistic in 2002.6 By 2020 it was 1 in 36.7 By 2022 it was 1 in 31. The areas checked were not the same each time. Some areas in 2022 had five times the rate of others.3

In the UK, new autism diagnoses in family doctors' records went up almost ninefold from 1998 to 2018. The biggest rise was in adults. The rise was bigger in girls and women than in boys and men.8 In Sweden, recorded autism went up about ninefold from 2001 to 2020.9 Of the people born in Sweden from 1985 to 2020, about 3 in 100 had an autism diagnosis by 2022.10 Across many countries, the numbers went up over time.1

Suggest a correction
Established

Independent record systems in the US, the UK and Sweden, and a meta-analysis, show the same steep rise in recorded autism. Each number is what one system counted; the ADDM case definition changed after 2002. These figures describe diagnoses and records, not the underlying rate.

Reading the ADDM series

The ADDM sites are not the same areas in every cycle, and the 2022 sites differed more than fivefold from each other. In 2002 clinicians reviewed the records against the diagnostic criteria; by 2022 a recorded diagnosis, special education eligibility or diagnostic code was enough. The series tracks records in the areas surveyed that year, not a fixed national sample.

Established

Record systems in three countries all show the same big rise. These numbers count diagnoses. They do not tell us how many people are autistic underneath.

Who is being counted Emerging

Where it has been measured, a sizeable share of the rise traces to changes in how autism is defined and recorded. In Denmark, for children born from 1980 to 1991, the 1994 change in diagnostic criteria and the addition of outpatient diagnoses to the national register in 1995 together accounted for an estimated 60% of the increase (33% and 42% when taken separately, with wide margins of error).11 In California, an estimated quarter (26.4%) of the growth in the state's autism caseload from 1992 to 2005 came through one route alone: people first diagnosed with "mental retardation" who later received an autism diagnosis.12 These measured changes leave part of the rise unexplained: by subtraction, about 40% in Denmark, with a wide margin of error, while the Californian figure covers one route only.11,12

Criteria can also narrow. When the DSM-5 criteria of 2013 were applied to ADDM records from 2006 and 2008, 81.2% of the children counted under the earlier DSM-IV-TR criteria still qualified, and the 2008 estimate would have fallen from 11.3 to 10.0 per 1,000.13

In Sweden, the people being diagnosed have changed. In Stockholm from 2001 to 2011, autism among 2 to 17 year olds rose almost 3.5-fold, almost all of it autism without intellectual disability (from 0.14% to 1.10%). Autism with intellectual disability rose only from 0.28% to 0.34%.14 The ADDM records show a different picture: intellectual disability in 40 to 62% of autistic 8-year-olds at the three sites with enough test data in 2000, and in 39.6% of those with test data across 15 sites in 2022, though the two years are not directly comparable.15,3 Across studies published since 2012 the median was 33%.2

The threshold for a diagnosis also seems to have moved. In a parent survey of 28,118 Swedish twins, among the 271 diagnosed before age 13, the autism symptom score reported by parents fell by about 30% across birth cohorts from 1992 to 2002. Among those diagnosed at ages 7 to 12 it fell by about 50% from 2004 to 2014, with no fall among those diagnosed before age 7.31 Across 11 meta-analyses covering 27,723 people, the differences between autistic and non-autistic groups shrank in later publications, significantly so for five of seven measures, including emotion recognition, theory of mind and planning. The analysis shows a link with publication year, not its cause.32

Part of the rise comes from changes in how autism is defined and written down. In Denmark, two changes in 1994 and 1995 explained an estimated 60% of the rise. One was new rules for diagnosis. The other was counting children seen at clinics, not just in hospitals.11 In California, about a quarter of the rise came from people who first had another diagnosis and later got autism.12

The rules can also get stricter. When newer rules from 2013 were tried on old US records, about 8 in 10 children still counted. The 2008 number would have dropped a little.13

In Sweden, the kind of person being diagnosed has changed. In Stockholm, almost all the rise from 2001 to 2011 was in autistic people without an intellectual disability.14 US records look different. In 2000, 4 to 6 in 10 autistic 8-year-olds in three areas had an intellectual disability. In 2022 it was about 4 in 10. The two years are hard to compare.15,3 Across many studies since 2012, the middle value was about 1 in 3.2

It may also take fewer autistic traits to get a diagnosis now. In Sweden, children diagnosed at ages 7 to 12 had fewer traits in later years.31 Studies comparing autistic and non-autistic groups also find smaller differences than they used to.32

Suggest a correction
Emerging

Two independent studies, from Denmark and California, agree that changes in diagnosis and reporting account for a sizeable share of the rise, but each estimate is single and the Danish one has wide margins of error. The shift toward autism without intellectual disability rests on Swedish registers alone; the US records are not comparable enough across years to test it.

New 2026

In Swedish national registers covering people born from 1988 to 2016, the share of each year's newly diagnosed autistic people who also had an intellectual disability diagnosis fell from 55.8% in 2001 to 6.7% in 2020, by a similar amount in females and males.9 In an editorial, Uta Frith argues that loosened criteria have diluted the clinical meaning of autism.16

Word
DSM

The Diagnostic and Statistical Manual of Mental Disorders, the American Psychiatric Association's manual of diagnostic criteria. DSM-IV appeared in 1994 and DSM-5 in 2013.

Word
Intellectual disability

Significant limits in intellectual functioning (usually an IQ of about 70 or below) and in everyday adaptive skills, beginning in childhood.

Emerging

Two studies from different places agree on this, but each gives only one estimate. The change in Sweden shows up in Swedish records only.

Word
Intellectual disability

Having a lot of difficulty with learning and reasoning, and with everyday skills, from childhood.

New 2026

In Swedish national registers covering people born from 1988 to 2016, the share of each year's newly diagnosed autistic people who also had an intellectual disability diagnosis fell from 55.8% in 2001 to 6.7% in 2020, by a similar amount in females and males.9 In an editorial, Uta Frith argues that loosened criteria have diluted the clinical meaning of autism.16

Label switching and services Emerging

Some of the rise reflects children moving from one label to another. In US special education from 1994 to 2003, autism among 6 to 11 year olds rose from 0.6 to 3.1 per 1,000. Over the same years, the categories for mental retardation and learning disabilities fell by 2.8 and 8.3 per 1,000, and states with more autism had larger falls. California did not clearly follow this pattern.17 In British Columbia schools, autism codes rose from 12.3 to 43.1 per 10,000 from 1996 to 2004, and children switching from other categories accounted for at least a third of the increase.18

How support is funded may also matter. In Australia, the staggered introduction of the National Disability Insurance Scheme, which funds support for individuals with a qualifying diagnosis, went with a 32% increase in reported autism prevalence. In this one study's model, the scheme accounted for an estimated 47% of new diagnoses since it began. Diagnoses from government-subsidized health professionals fell and diagnoses from disability service providers rose. The authors judged a lower threshold for recognizing autism more consistent with their results than catching up on groups missed in the past.19

Some children moved from one label to another. In US schools from 1994 to 2003, more children were listed as autistic. At the same time, fewer were listed under some other learning labels.17 In British Columbia, Canada, at least a third of the rise came from children switching labels.18

How help is paid for may also matter. Australia started a scheme that pays for support for people with a diagnosis. Reported autism went up by about a third after it began. The researchers think a lower bar for diagnosis fits their results better than finding people who were missed before.19

Suggest a correction
Emerging

Each finding comes from one administrative system (US special education, British Columbia schools, Australia's disability scheme), and the size of the effect differs from place to place.

Word
Diagnostic substitution

When people who would once have received one diagnosis receive a different one instead, so that one count rises while another falls.

Emerging

Each finding comes from one school or support system. How big the effect is changes from place to place.

Adults and women Emerging

Many autistic adults have never been diagnosed. A 2007 survey of English households assessed adults directly and estimated 9.8 per 1,000, with a wide margin of error. The rate did not differ by age, and the autistic adults found tended to be unrecognised.20 Adding adults with intellectual disability from case registers, the same research program estimated 11 per 1,000 among adults of all ages, again with wide margins of error.21

A second English count asks adults to report their own autism rather than assessing them. In a repeated survey series of 5,999,433 adults registered with a general practice, run from 2018 to 2025, the weighted share who reported being autistic rose from 0.9% in 2018 to 2.6% in 2025, a rise of 188.9%, most of it between 2021 and 2025. Over the same years the share reporting a learning disability rose 26.7% and the share reporting a mental health condition 51.6%, so the autism rise outpaced the other conditions the survey asked about.42 This counts something different from the 2007 survey: people who say they are autistic, not adults assessed against diagnostic criteria, so it cannot separate a change in how common autism is from a change in who reports it.

New Zealand's national databases tell a similar story for diagnosed adults. Across three linked databases covering 2009 to 2023, diagnosed autism rose about threefold. By 2023 the rate was 1.3% among children and 0.14% among adults, up from about 0.04% among adults in 2009, and diagnosed autism stayed more common in males.43

The recent growth in diagnoses has come partly from these groups. In UK primary care, new diagnoses rose most among adults and more in females than in males.8 In Sweden, the ratio of males to females was lower among people diagnosed at older ages.10 A preprint, not yet peer reviewed, reports that adult diagnoses in the same UK primary-care database climbed fastest among 16 to 20 year olds and among females, and that the UK rise sped up after 2020; the authors say whether this reflects better recognition or a broadening of diagnostic practice is unresolved.44

Many autistic adults have never been diagnosed. In 2007, researchers tested adults in English homes. They estimated about 1 in 100. The rate was about the same at every age. Most of the autistic adults they found had not been recognized.20,21

Another English survey asked adults if they were autistic. It ran from 2018 to 2025. The share who said yes rose from 0.9% to 2.6%. That is nearly three times as many. Most of the rise came after 2021. The rise was much smaller for learning disability and mental health conditions.42 These are people's own answers. They are not the same as a diagnosis.

New Zealand keeps national records of autism diagnoses. From 2009 to 2023, the number of diagnosed people rose about threefold. In 2023, about 1 in 80 children and 1 in 700 adults were diagnosed. More males than females were diagnosed.43

Some of the recent rise in diagnoses comes from these groups. In the UK, diagnoses rose most in adults, and more in women and girls than in men and boys.8 In Sweden, the gap between males and females was smaller in people diagnosed when older.10 A preprint is a study that is not yet checked by other scientists. One preprint says adult diagnoses rose fastest in 16 to 20 year olds and in females. It also says the UK rise sped up after 2020. The authors are not sure if that means better recognition or wider diagnosis.44

Suggest a correction
Emerging

The assessed English figure is one survey program with a wide margin of error. The self-report trend and the New Zealand figures are each a single source; the UK and Swedish adult trends are single register analyses, and the two UK analyses draw on the same primary-care database. Self-reported and diagnosed counts measure different things.

See also
Emerging

The adult numbers come from one survey program. The other numbers come from single studies. Numbers people report about themselves are different from numbers based on a diagnosis.

See also

Is there a real increase? Emerging

The few studies that kept their method steady over time mostly find little change, though not all. In Swedish twins born from 1993 to 2002, the share with autistic traits, measured by the same parent interview, stayed stable over ten years, while registered diagnoses among all Swedish children of the same ages rose steadily.22 In South Korea, 12 consecutive birth cohorts born from 1998 to 2010 were screened and assessed at age 7 by the same method. Prevalence stayed between 2.0% and 3.4%, with no significant rise or fall. This is the same research group as the earlier Korean study, and fewer than half of the 2,077 children who screened positive completed an assessment.23 A UK comparison of two cohorts of 7-year-olds, in Avon in 1998 to 1999 and across the UK in 2007 to 2008, found otherwise. Parent-reported diagnoses rose from 1.09% to 1.68%. The share of children with high scores on a composite of autism-related social and communication traits, a measure the authors note is not validated for autism, also rose (risk ratio 1.61), and allowing for it made the rise in diagnoses no longer significant. The authors say the higher diagnoses may partly reflect a real rise in these traits, greater recognition by parents and teachers, or both.30

The Global Burden of Disease study, which models prevalence using only studies that actively searched for cases, estimated 788.3 autistic people per 100,000 in 2021, or about 1 in 127 (61.8 million people), after adjusting for the age structure of each country. The comparable figure for 1990 was 773.2. Its authors caution that a shortened data window may have limited their ability to detect trends.24 The 2010 round of the same study also found no clear change from 1990 to 2010,25 and the 2023 round listed notable rises in several mental health conditions but not in autism.26

The English adult survey is weakly consistent with little change: its authors reasoned that if autism had become more common, older adults would have a lower rate, and they did not.20 That reasoning rests on one survey with a wide margin of error, and assumes the assessment works as well in older adults.

Some of the rise could still be real, though the evidence here is indirect. In a meta-analysis of 27 studies, older parental age was associated with a higher likelihood of autism (18% and 21% higher for each 10 years of maternal and paternal age).33 Whether this adds measurably to the trend has not been tested in any study found as of October 2026.

In 2025 and 2026, official US sources lent partial support to the claim that vaccines contribute to the rise in autism. The CDC's autism and vaccines page, dated November 2025 and revised in July 2026, says the statement "Vaccines do not cause autism" is "not an evidence-based claim," that scientific studies "have not ruled out the possibility that infant vaccines contribute to the development of autism," and that "the rise in autism prevalence since the 1980s correlates with the rise in the number of vaccines given to infants."39 Federal research policy has also shifted emphasis toward environmental contributors, alongside genetics: a draft 2026-2028 strategic plan from the Interagency Autism Coordinating Committee, considered in August 2026, says it has not settled whether a true increase in incidence has occurred, and concludes that the usual explanations for the rise, such as awareness, earlier identification and broadened criteria, do not, between them, account for the full size of the increase.40 The vaccine claim is the one tested in the callout below, where the pooled cohort and meta-analytic evidence finds no link. The claim that autism is spreading is not well supported by the counts in this entry: much of the rise follows changes in how autism is defined and counted, and the fixed-method studies above mostly find little change, though one UK comparison found traits rising with diagnoses.

In the United Kingdom, the government commissioned an independent review in December 2025, applying to England, into the prevalence and support for mental health conditions, ADHD and autism, including the question of overdiagnosis. It was chaired by the clinical psychologist Peter Fonagy. Its final report, published on 9 October 2026, found "little evidence of a substantial change in underlying population prevalence" of autism, though it notes there has been no recent England-wide prevalence study; the recorded (administrative) prevalence among 6 to 11 year olds in England rose from 0.24% in 2000 to 4.58% in 2025, which the report describes as more than a 20-fold rise.41 The report did not back the strongest form of the claim: it said it could not conclude that overdiagnosis was occurring "at scale," and it found no fall in severity among those presenting for assessment, so that they are no less affected than those diagnosed earlier. It also warns that recorded diagnoses are rising past older population estimates, which "should not be treated as fixed biological ceilings," and that the balance of risks "may be changing" toward misdiagnosis and overdiagnosis. That reading of the underlying rate agrees with the fixed-method studies above, and the report calls for regular population prevalence surveys.

Often saidThe rise in autism is an epidemic caused by vaccines.
Evidence

A large part of the rise is explained by changes in who is diagnosed. A meta-analysis of 5 cohort studies (over 1.2 million children) and 5 case-control studies found no link between vaccines and autism.27 A study of 657,461 Danish children found no higher rate of autism after the measles, mumps and rubella vaccine (hazard ratio 0.93).28 Later studies agree. In a US health-record cohort of 2.56 million children, a first measles, mumps and rubella dose before age 2 was not linked to autism (adjusted hazard ratio 0.97; 99% confidence interval 0.91 to 1.03).36 In a Danish cohort of 1.22 million children, the aluminum in vaccines was not associated with any higher rate of autism or the other neurodevelopmental conditions studied (adjusted hazard ratio 0.93 for each 1-milligram increase in aluminum; 95% confidence interval 0.90 to 0.97).35 In Israel, influenza vaccination during pregnancy was not linked to autism in 153,321 children once other factors were accounted for; the unadjusted rate was higher (hazard ratio 1.22), and the adjusted hazard ratio was 0.97 (95% confidence interval 0.91 to 1.05).37 A 2026 review of 59 studies reached the same conclusion for aluminum-containing vaccines, resting on the randomized trials and large cohorts among them.38

A few studies used the same way of counting each year. Most found little change, but not all. In Sweden, parents of twins answered the same questions for ten years. The share of children with autistic traits stayed the same. Over those years, autism diagnoses in Swedish children kept going up.22 In South Korea, children were tested at age 7 for 12 years in a row. The rate stayed between 2 and 3.4 in 100. Fewer than half of the children flagged by the test were fully checked.23 A UK study found something different. Both diagnoses and autistic behavior went up between 1998 and 2008. The authors say parents and teachers noticing autism more may play a part too.30

A world study estimated that about 1 in 127 people were autistic in 2021. Its estimate for 1990 was about the same. Its authors warn that their method may miss slow changes.24 Earlier and later rounds of the same study also found no clear rise.25,26

Some of the rise may still be real. Having older parents goes with a slightly higher chance of being autistic. Nobody has checked yet whether this changes the overall numbers.33

Two governments have given opposite answers to this question. In the United States, the CDC now says it cannot rule out that vaccines play a part. It points to the way the rise in autism lines up with the rise in the number of vaccines.39 US autism research advisers also want more research on what in the environment might play a part. They say no single reason explains the whole rise.40 In the United Kingdom, the government asked experts to look at diagnosis and overdiagnosis. Their report came out in October 2026. It found that the share of children who are autistic has not changed much. But far more children are now written down as autistic in doctors' records.41 The report said it could not tell yet whether overdiagnosis is happening on a large scale. But it warned that the risk may be growing. It also found that children coming forward now have just as much difficulty as children diagnosed before.

Often saidVaccines caused the rise in autism.
Evidence

A large part of the rise comes from changes in who gets diagnosed. Big studies of millions of children have found no link between vaccines and autism. This includes the MMR shot, the aluminum added to some vaccines, and flu shots given during pregnancy.27,28,35,36,37,38

Suggest a correction
Emerging

Studies with a fixed method disagree. Swedish twins show stable traits and repeated Korean screening a stable rate (with fewer than half of screen-positive children assessed), but a UK comparison found traits rising along with diagnoses. The Global Burden of Disease rounds are one modeling program whose authors warn about detecting trends.

New 2026

The Korean screening study's authors conclude that changes in detection, rather than in the underlying rate, may explain much of the reported increase. Girls made up a disproportionate share of the previously undiagnosed children they found in ordinary classrooms.23

Emerging

Scientists are not sure yet. Studies that counted the same way every year do not all agree.

New 2026

The Korean screening study's authors conclude that changes in detection, rather than in the underlying rate, may explain much of the reported increase. Girls made up a disproportionate share of the previously undiagnosed children they found in ordinary classrooms.23

Boys, girls and a narrowing ratio Established

More males than females are diagnosed, but the size of the gap depends on how cases are found. A meta-analysis of 54 studies found a pooled male-to-female odds ratio of 4.20, close to the commonly quoted 4 to 1. In higher-quality studies it was 3.32, and in studies that screened a whole population it was 3.25, against 4.56 in studies that counted only existing diagnoses.29 A review of studies since 2012 found a median ratio of 4.2.2 In the ADDM Network the ratio was 3.8 in 2020 and 3.4 in 2022 (49.2 per 1,000 boys against 14.3 per 1,000 girls).7,3 The Global Burden of Disease models gave age-standardized estimates of 1,064.7 per 100,000 males and 508.1 per 100,000 females in 2021.24

In the Korean screening study published in 2011, the ratio was 2.5 to 1 among children found in mainstream schools and 5.1 to 1 among children already in special education or on disability registers.5 In English adults, being male went with a much higher likelihood of autism only among those with no or mild intellectual disability, an estimate with a very wide margin of error (odds ratio 8.5, 95% interval 2.0 to 34.9); the authors suggest that the measures used may miss autistic women.21

Part of the difference may be biological. In Danish families, siblings of autistic girls were more often autistic than siblings of autistic boys, and in two US collections the mothers of autistic children carried slightly more common genetic variants linked to autism than the fathers (a difference of 0.08 standard deviations). The authors, one research group, read this as a female protective effect.34

More boys than girls are diagnosed with autism. A big review found about 4 boys for every girl. When researchers tested everyone in an area, the gap was smaller, about 3 to 1.29,2 In the US network in 2022, it was about 3 to 1.3

In Korea, the gap was smaller among children found in ordinary schools than among children already getting special help.5 Part of the gap may come from genes. One study suggests girls may need more autism-linked genes than boys before they become autistic.34 In England, the researchers think their tests may miss some autistic women.21

Suggest a correction
Established

A meta-analysis of 54 studies, a systematic review, US surveillance and the Global Burden of Disease models agree that more males than females are diagnosed; the meta-analysis shows the gap is smaller when a population is screened. The Korean and English findings are single studies, and the Swedish trend toward parity is a single register study and a projection.

New 2026

In Swedish registers, the male-to-female ratio fell with older age at diagnosis and, for people diagnosed after age 10, over calendar time. By 2022 the cumulative ratio by age 20 was 1.2, and the authors projected parity by age 20 by 2024.10

See also
Established

Many studies agree that more boys than girls are diagnosed. One big review shows the gap is smaller when everyone is tested. The Swedish trend is from one study.

Word
Ratio

A way of comparing two numbers, such as 3 boys for every 1 girl.

New 2026

In Swedish registers, the male-to-female ratio fell with older age at diagnosis and, for people diagnosed after age 10, over calendar time. By 2022 the cumulative ratio by age 20 was 1.2, and the authors projected parity by age 20 by 2024.10

See also

In the community Community

Autistic writers mostly treat the rise as recognition catching up with people who were always there. Eric Garcia, an autistic journalist, recounts how Steve Silberman came to write NeuroTribes: struck by the gap between missing services and media coverage focused on vaccines, Silberman noticed that a hockey-stick graph of rising diagnoses, used by organizations such as Autism Speaks, "led to panicked news articles" about an epidemic.c1 In Garcia's account, Silberman then found scholarly articles debunking the idea.c1

Silberman, who is not autistic, tells the history through Lorna Wing, who had expected that widening the definition of autism would raise the estimates. "These people have always existed," she said.c2 Silberman argues that fears of an epidemic also "skewed the direction of autism research" toward causes and risk factors, while work on autistic people's quality of life went underfunded.c2

Sol Smith dismisses both "the 1990s fear of vaccinations" and the idea that autism is a TikTok trend. He puts more people being diagnosed and self-identifying down partly to better language: "our vocabulary surrounding autism has become much better", as autistic people describe their own experience, and he describes posting a TikTok about meltdowns that drew hundreds of comments from people who recognized themselves.c3 Devon Price reads the rising US figures as a sign of underdiagnosis and estimates that "at least half of all Autistic people in the United States currently fail to get diagnosed", reasoning from how many people lack access to mental-health care.c4 That is Price's estimate, not a measured figure.

Clinicians, educators and parents writing for others add their own readings. Donna Henderson, Sarah Wayland and Jamell White, in a guide for clinicians, trace published estimates from about 1 in 2,222 children in 1966 to 1 in 44 in a 2018 survey, and pass on John Elder Robison's suggestion that only 3 to 10% of autistic adults are diagnosed; Robison's figure is his reading of one study, not a measurement.c5 Judith Gainsborough, an educational psychologist, notes that many researchers think the rise of the 1980s and 1990s "cannot be explained completely by the broadening of the category", and points, after Wing and Potter, to better-informed practitioners, sharper identification and better assessment tools.c6 Nick Chown, writing on autism theory, asks whether the difference between the sexes reflects tests that fail to identify autism in girls and women, and suggests calling it an "apparent gender disparity".c7 Carrie Cariello, a parent, recalls hearing "One out of one hundred and fifty-eight children" at her son's diagnosis and writes of today's higher figures: "There is much speculation as to this increase, but when asked about it, I have no answer."c8

The autistic writers above treat the rise as almost wholly a matter of recognition. The research above finds that changes in who is diagnosed explain a large part of it, and that most studies with a steady method find little change in the underlying rate, though one UK comparison found traits rising with diagnoses.

Some people call the rise in autism an "epidemic". Autistic writers mostly disagree. They say more autistic people are being noticed, not that more exist. Eric Garcia says scary graphs led to panicked news stories.c1 Steve Silberman is a writer who is not autistic. He quotes a doctor who said, "These people have always existed."c2 He argues that fear sent research money toward causes. Less went to making autistic lives better.c2

Sol Smith says vaccines and TikTok do not explain the rise. He says people now have better words for autism.c3 Devon Price guesses that at least half of autistic people in the US are not diagnosed.c4 That is a guess, not a measured number.

Some clinicians, teachers and parents also write about the rise. Donna Henderson and two other clinicians list how the counts have grown since 1966.c5 They pass on a guess by John Elder Robison. He thought only 3 to 10 out of 100 autistic adults are diagnosed. That is a guess, not a measured number.c5 Judith Gainsborough works with schools. She notes that many researchers think wider rules alone do not explain the rise. Experts also got better at spotting autism.c6 Nick Chown asks if the gap between boys and girls partly comes from tests that miss girls.c7 Carrie Cariello is a parent. She says people wonder why the numbers rose, but she has no answer.c8

The research above is more mixed. Changes in who gets diagnosed explain a large part of the rise. Most studies find little real change, but one UK study found more autistic traits over time.

Suggest a correction
Community

How autistic writers, a non-autistic journalist, clinicians, an educator and a parent describe the rise in diagnoses and the "autism epidemic". Not a research finding.

See also
Community

This is how autistic writers, and some doctors, teachers and parents, describe it. It is not a research finding.

See also

Sources

Research

  1. 1Talantseva OI, Romanova RS, Shurdova EM, et al. The global prevalence of autism spectrum disorder: A three-level meta-analysis. Frontiers in Psychiatry, 2023. https://doi.org/10.3389/fpsyt.2023.1071181
  2. 2Zeidan J, Fombonne E, Scorah J, et al. Global prevalence of autism: A systematic review update. Autism Research, 2022. https://doi.org/10.1002/aur.2696
  3. 3Shaw KA, Williams S, Patrick ME, et al. Prevalence and Early Identification of Autism Spectrum Disorder Among Children Aged 4 and 8 Years: Autism and Developmental Disabilities Monitoring Network, 16 Sites, United States, 2022. MMWR Surveillance Summaries, 2025. https://doi.org/10.15585/mmwr.ss7402a1
  4. 4Yan X, Li Y, Li Q, et al. Prevalence of Autism Spectrum Disorder Among Children and Adolescents in the United States from 2021 to 2022. Journal of Autism and Developmental Disorders, 2026. https://doi.org/10.1007/s10803-024-06390-7
  5. 5Kim YS, Leventhal BL, Koh YJ, et al. Prevalence of Autism Spectrum Disorders in a Total Population Sample. American Journal of Psychiatry, 2011. https://doi.org/10.1176/appi.ajp.2011.10101532
  6. 6Autism and Developmental Disabilities Monitoring Network Surveillance Year 2002 Principal Investigators; CDC. Prevalence of autism spectrum disorders: Autism and Developmental Disabilities Monitoring Network, 14 sites, United States, 2002. MMWR Surveillance Summaries, 2007. https://pubmed.ncbi.nlm.nih.gov/17287715/
  7. 7Maenner MJ, Warren Z, Williams AR, et al. Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 8 Years: Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2020. MMWR Surveillance Summaries, 2023. https://doi.org/10.15585/mmwr.ss7202a1
  8. 8Russell G, Stapley S, Newlove-Delgado T, et al. Time trends in autism diagnosis over 20 years: a UK population-based cohort study. Journal of Child Psychology and Psychiatry, 2022. https://doi.org/10.1111/jcpp.13505
  9. 9Salkić A, Tideman M, Martini MI, et al. The proportion of intellectual disability in autism spectrum disorder over two decades. Psychiatry Research, 2026. https://doi.org/10.1016/j.psychres.2026.117246
  10. 10Fyfe C, Winell H, Dougherty J, et al. Time trends in the male to female ratio for autism incidence: population based, prospectively collected, birth cohort study. BMJ, 2026. https://doi.org/10.1136/bmj-2025-084164
  11. 11Hansen SN, Schendel DE, Parner ET. Explaining the Increase in the Prevalence of Autism Spectrum Disorders. JAMA Pediatrics, 2015. https://doi.org/10.1001/jamapediatrics.2014.1893
  12. 12King M, Bearman P. Diagnostic change and the increased prevalence of autism. International Journal of Epidemiology, 2009. https://doi.org/10.1093/ije/dyp261
  13. 13Maenner MJ, Rice CE, Arneson CL, et al. Potential Impact of DSM-5 Criteria on Autism Spectrum Disorder Prevalence Estimates. JAMA Psychiatry, 2014. https://doi.org/10.1001/jamapsychiatry.2013.3893
  14. 14Idring S, Lundberg M, Sturm H, et al. Changes in Prevalence of Autism Spectrum Disorders in 2001–2011: Findings from the Stockholm Youth Cohort. Journal of Autism and Developmental Disorders, 2015. https://doi.org/10.1007/s10803-014-2336-y
  15. 15Autism and Developmental Disabilities Monitoring Network Surveillance Year 2000 Principal Investigators; CDC. Prevalence of autism spectrum disorders: Autism and Developmental Disabilities Monitoring Network, six sites, United States, 2000. MMWR Surveillance Summaries, 2007. https://pubmed.ncbi.nlm.nih.gov/17287714/
  16. 16Frith U. Autism spectrum disorder: has it lost its meaning and is it leading to misdiagnosis?. Psychological Medicine, 2026. https://doi.org/10.1017/s0033291726105376
  17. 17Shattuck PT. The Contribution of Diagnostic Substitution to the Growing Administrative Prevalence of Autism in US Special Education. Pediatrics, 2006. https://doi.org/10.1542/peds.2005-1516
  18. 18Coo H, Ouellette-Kuntz H, Lloyd JEV, et al. Trends in Autism Prevalence: Diagnostic Substitution Revisited. Journal of Autism and Developmental Disorders, 2008. https://doi.org/10.1007/s10803-007-0478-x
  19. 19Ranjan M, Breunig R. Individualized disability support schemes and their impact on autism diagnoses. Journal of Health Economics, 2026. https://doi.org/10.1016/j.jhealeco.2025.103100
  20. 20Brugha TS, McManus S, Bankart J, et al. Epidemiology of Autism Spectrum Disorders in Adults in the Community in England. Archives of General Psychiatry, 2011. https://doi.org/10.1001/archgenpsychiatry.2011.38
  21. 21Brugha TS, Spiers N, Bankart J, et al. Epidemiology of autism in adults across age groups and ability levels. British Journal of Psychiatry, 2016. https://doi.org/10.1192/bjp.bp.115.174649
  22. 22Lundström S, Reichenberg A, Anckarsäter H, et al. Autism phenotype versus registered diagnosis in Swedish children: prevalence trends over 10 years in general population samples. BMJ, 2015. https://doi.org/10.1136/bmj.h1961
  23. 23Kim YS, Liu X, Chang J, et al. Cumulative Incidence and Prevalence of Autism Spectrum Disorder. JAMA Pediatrics, 2026. https://doi.org/10.1001/jamapediatrics.2026.4066
  24. 24Santomauro DF, Erskine HE, Mantilla Herrera AM, et al. The global epidemiology and health burden of the autism spectrum: findings from the Global Burden of Disease Study 2021. The Lancet Psychiatry, 2025. https://doi.org/10.1016/S2215-0366(24)00363-8
  25. 25Baxter AJ, Brugha TS, Erskine HE, et al. The epidemiology and global burden of autism spectrum disorders. Psychological Medicine, 2015. https://doi.org/10.1017/S003329171400172X
  26. 26Santomauro DF, Miller PA, Shadid J, et al. Updated trends in the global prevalence and burden of mental disorders, 1990–2023: a systematic analysis for the Global Burden of Disease Study 2023. The Lancet, 2026. https://doi.org/10.1016/S0140-6736(26)00519-2
  27. 27Taylor LE, Swerdfeger AL, Eslick GD. Vaccines are not associated with autism: An evidence-based meta-analysis of case-control and cohort studies. Vaccine, 2014. https://doi.org/10.1016/j.vaccine.2014.04.085
  28. 28Hviid A, Hansen JV, Frisch M, et al. Measles, Mumps, Rubella Vaccination and Autism. Annals of Internal Medicine, 2019. https://doi.org/10.7326/M18-2101
  29. 29Loomes R, Hull L, Mandy WPL. What Is the Male-to-Female Ratio in Autism Spectrum Disorder? A Systematic Review and Meta-Analysis. Journal of the American Academy of Child & Adolescent Psychiatry, 2017. https://doi.org/10.1016/j.jaac.2017.03.013
  30. 30Russell G, Collishaw S, Golding J, et al. Changes in diagnosis rates and behavioural traits of autism spectrum disorder over time. BJPsych Open, 2015. https://doi.org/10.1192/bjpo.bp.115.000976
  31. 31Arvidsson O, Gillberg C, Lichtenstein P, et al. Secular changes in the symptom level of clinically diagnosed autism. Journal of Child Psychology and Psychiatry, 2018. https://doi.org/10.1111/jcpp.12864
  32. 32Rødgaard EM, Jensen K, Vergnes JN, et al. Temporal Changes in Effect Sizes of Studies Comparing Individuals With and Without Autism. JAMA Psychiatry, 2019. https://doi.org/10.1001/jamapsychiatry.2019.1956
  33. 33Wu S, Wu F, Ding Y, et al. Advanced parental age and autism risk in children: a systematic review and meta-analysis. Acta Psychiatrica Scandinavica, 2017. https://doi.org/10.1111/acps.12666
  34. 34Wigdor EM, Weiner DJ, Grove J, et al. The female protective effect against autism spectrum disorder. Cell Genomics, 2022. https://doi.org/10.1016/j.xgen.2022.100134
  35. 35Andersson NW, Bech Svalgaard I, Hoffmann SS, Hviid A. Aluminum-Adsorbed Vaccines and Chronic Diseases in Childhood: A Nationwide Cohort Study. Annals of Internal Medicine, 2025. https://doi.org/10.7326/ANNALS-25-00997
  36. 36Burstain TL, Jacques D, Burstain JM. Association Between First MMR Vaccination Before Age 2 Years and Childhood Autism in a U.S. EHR Cohort of 2.5 Million Children. The Pediatric Infectious Disease Journal, 2026. https://doi.org/10.1097/INF.0000000000005349
  37. 37Neeman S, Hemo M, Meiri G, Shmueli D, Menashe I. Maternal influenza vaccination during pregnancy and risk of autism spectrum disorder in the offspring. JCPP Advances, 2026. https://doi.org/10.1002/jcv2.70085
  38. 38Doyon-Plourde P, Chong J, Abrams EM, et al. Aluminium adjuvants in vaccines and potential health effects: systematic review. BMJ, 2026. https://doi.org/10.1136/bmj-2025-088921
  39. 39CDC. Autism and Vaccines. Centers for Disease Control and Prevention, 2026. https://www.cdc.gov/vaccine-safety/about/autism.html
  40. 40Interagency Autism Coordinating Committee. IACC Strategic Plan 2026-2028 (working draft). US Department of Health and Human Services, 2026. https://iacc.hhs.gov/meetings/iacc-meetings/2026/full-committee-meeting/august/
  41. 41Department of Health and Social Care. Independent review into mental health conditions, ADHD and autism: final report. GOV.UK, 2026. https://www.gov.uk/government/publications/independent-review-into-mental-health-conditions-adhd-and-autism-final-report
  42. 42Tromans SJ, Morgan ZL, Brewer B, Williams R, Brugha TS, McManus S. Trends in Self-Reported Autism Among Adults in England: Analysis of a Repeated Cross-Sectional Patient Survey Series of 5,999,433 Adults. Journal of Autism and Developmental Disorders, 2026. https://doi.org/10.1007/s10803-026-07421-1
  43. 43Beaglehole B, Lu G, Frampton C. Autism spectrum disorder trends for children and adults in New Zealand: A longitudinal examination of three national databases. Australian & New Zealand Journal of Psychiatry, 2026. https://doi.org/10.1177/00048674261472427
  44. 44Sadik A, Lundberg M, Khandaker GM, et al. Changes in the profile of adults diagnosed as autistic since 2010: population based studies in the United Kingdom and Sweden. medRxiv (preprint), 2026. https://doi.org/10.64898/2026.05.20.26353486

Community sources

  1. c1Garcia E. We're Not Broken: Changing the Autism Conversation. Houghton Mifflin Harcourt, 2021. Chapter "What Comes Next". Eric Garcia's site
  2. c2Silberman S. NeuroTribes: The Legacy of Autism and the Future of Neurodiversity. Avery, 2015. Chapter 10, "Pandora's Box". Steve Silberman's site
  3. c3Smith S. The Autistic's Guide to Self-Discovery: Flourishing as a Neurodivergent Adult. New World Library, 2025. Chapter 5, "Your Mask", section "There's Got to Be Something Else Going On". Sol Smith's site
  4. c4Price D. Unmasking Autism: Discovering the New Faces of Neurodiversity. Harmony Books, 2022. Chapter 1, "What Is Autism, Really?", section "Suspect You're Autistic?". Devon Price's site
  5. c5Henderson D, Wayland S, White J. Is This Autism? A Guide for Clinicians and Everyone Else. Routledge, 2023. Introduction, section "Key points".
  6. c6Gainsborough J. Effective Support and Education of Autistic Young People: Educational Psychology Perspectives. Jessica Kingsley Publishers, 2022. Chapter 1, "Autism and Educational Psychology: Two Developing Stories".
  7. c7Chown N. Understanding and Evaluating Autism Theory. Jessica Kingsley Publishers, 2016. Chapter 9, "What Has 70 Years Taught Us about Autism?".
  8. c8Swenson K, Wood A, Cariello C. Autism Out Loud: Life with a Child on the Spectrum, from Diagnosis to Young Adulthood. Park Row Books, 2025. Chapter 2, "Diagnosis Day", Carrie.