The Diagnosis Gap: Why Autistic Girls and Women Are Identified Later

Autistic girls and women are diagnosed later than autistic boys and men, and for decades less often than their numbers would predict. From the US to Sweden, Denmark, England and Australia, the long-quoted figure of four autistic boys for every girl turns out to depend partly on who gets noticed: screen whole populations and it falls to about three to one.

Behind those ratios is a pattern many autistic women will recognise. It is the girl who holds herself together all day at school and comes apart at home, while her teachers see a quiet, well-behaved pupil. It is the teenager treated for anxiety, depression or an eating disorder, with nobody asking what sits underneath. It is the woman who, after years of being told she was too sensitive, recognises herself in a description of autism in her thirties or forties. Each delay is time without the understanding and support that recognition can bring.

The 4:1 ratio partly measures who gets noticed

For decades, autism was described as about four times more common in boys. The best-known test of that figure is a 2017 meta-analysis of 54 prevalence studies covering 13.8 million people, 53,712 of them autistic^[1]. Pooled, the studies gave a male-to-female ratio of 4.20, with very substantial variation between studies, and the answer depended on method. Where researchers screened a whole population, finding autistic children whether or not they had a diagnosis, the ratio was 3.25; where they counted only existing diagnoses, it was 4.56. High-quality studies came out at 3.32. The authors concluded that among children who meet criteria the true ratio is "closer to 3:1", and that there "appears to be a diagnostic gender bias", leaving girls who meet criteria "at disproportionate risk of not receiving a clinical diagnosis"^[1].

Many prevalence estimates still sit close to the traditional ratio. A 2022 global review of 71 prevalence studies published since 2012 found a median male-to-female ratio of 4.2^[4]. In the US, the CDC's Autism and Developmental Disabilities Monitoring (ADDM) Network found autism identified in 49.2 per 1,000 boys and 14.3 per 1,000 girls aged 8 across 16 sites in 2022, a ratio of 3.4, down from 4.2 in 2018 and 3.8 in 2020^[5]. The CDC cautions that this could be deceiving if read as better identification of girls: over the same years the absolute difference between boys and girls widened, from 27.7 to 34.9 per 1,000^[5].

The gap narrows as girls grow up

Whole-population registers give some of the clearest evidence that many girls are recognised late rather than not autistic at all. In Sweden, a study of all 2.76 million children born between 1985 and 2020 found that 2.8% had been diagnosed autistic by 2022^[3]. Up to age 10 the male-to-female ratio of new diagnoses was about 3; after that it fell steadily. Among people diagnosed in 2020–2022, the ratio was no longer above 1 for those older than 15: girls and young women were being newly diagnosed at least as often as boys and young men. By 2022 the ratio of diagnoses made by age 20 was 1.2, and the authors projected parity by 2024 if trends continued. They concluded that in Sweden the ratio "may no longer be distinguishable by adulthood"^[3].

Denmark shows the same late surge. Across 2.8 million children followed from 2000 to 2024, new diagnoses in boys peaked at age 5 and stayed high until 15; in girls they stayed low until age 8, then climbed to a peak at 14^[6]. Rates rose for both sexes, more steeply among girls, but by age 18, 6.5% of boys and 3.9% of girls had been diagnosed.

Other national data, though not directly comparable, point the same way. In Australia's national disability survey, the number of females reported as autistic rose 95.5% between 2018 and 2022, from 46,500 to 90,900, against 24.2% for males^[7]. In English GP records, the rate of new diagnoses rose 8.6-fold for females between 2000 and 2018, against 5.5-fold for males^[2]. A UK primary-care study found the rate of newly recorded diagnoses rose 787% from 1998 to 2018, more for females than males; its authors judged increased reporting and application of the diagnosis, rather than a real rise in prevalence, the more likely explanation^[8].

Girls are diagnosed later, often by years

Later identification of girls shows up wherever it has been measured. The measures differ; the direction does not.

For many, later has meant not yet. Researchers estimated from GP records that as of 2018 between 435,700 and 1.2 million people in England may be autistic but undiagnosed, 59–72% of all autistic people, with the undiagnosed share highest among older adults^[2]. In those records, 1.22% of males but only 0.35% of females had a diagnosis, and the authors cautioned that "there could be even more undiagnosed autistic females than we have estimated". Our articles on diagnosis in adulthood and age at diagnosis around the world look at these gaps across the whole population.

Masking, and a template drawn from boys

Researchers point to two overlapping reasons why girls are missed. The first is camouflaging, or masking: hiding autistic characteristics and compensating for social differences. In a qualitative study of 92 autistic adults, people described masking to fit in and connect with others, at a cost that included exhaustion and "threats to self-perception"^[11]. A 2025 meta-analysis of ten studies using a self-report camouflaging questionnaire found that autistic females reported more masking and compensation than autistic males, though the differences were small, and the few non-binary participants did not differ significantly from either group^[12]. Camouflaging is also linked to when people are recognised: in online samples of 242 autistic men and 570 autistic women, the women had been diagnosed later on average, and the association between camouflaging and later diagnosis was stronger for women^[13].

The second is a clinical template built mainly on boys. In the same paper, a separate review of 67 studies comparing autistic females and males found that on standard clinical measures autistic males scored slightly higher on core autistic characteristics, while females showed more cognitive and behavioural difficulties; the authors read this as support for "a bias in clinical procedures towards males"^[12]. A Dutch study of 231 children referred for assessment found that repetitive and restricted behaviours predicted a diagnosis less strongly in girls, while emotional and behavioural problems raised the chance of diagnosis more in girls than in boys^[14], a finding since summarised as girls needing "additional behavioural challenges" to be diagnosed^[2]. The Dutch study included only 24 autistic girls. A systematic review of barriers for girls and young women under 21 named compensatory behaviours, parental concerns, other people's perceptions, lack of information and clinician bias^[15].

Masking is not unique to women, not every autistic girl masks, and the research relies mostly on self-report from adults without an intellectual disability. But the pattern is consistent enough to matter to anyone designing an assessment.

Diagnosed with something else first

Before autism is recognised, many autistic girls and women are given other psychiatric diagnoses. In Sweden, researchers followed everyone born between 1990 and 2015 who went on to receive a clinical autism diagnosis: 72,331 people, a third of them female^[9]. Before that diagnosis, 54.2% of the females and 40.9% of the males already had at least one psychiatric diagnosis on record. Allowing for birth year and age at autism diagnosis, females had higher odds than males of most earlier diagnoses, except psychotic disorders and ADHD, and females with an earlier diagnosis were diagnosed autistic 1.27 years later, on average, than males with one.

An earlier diagnosis is not necessarily a wrong one: anxiety and depression are common among autistic people, and many of those diagnoses will have been accurate and needed. For most conditions, fewer than half of people with an earlier diagnosis were still receiving specialist care or medication for it in the five years after their autism diagnosis. The authors call for "disentangling persistent support needs from overlapping diagnostic presentations, particularly in autistic females"^[9].

Autistic adults' own accounts point the same way. In a Dutch register of 1,211 autistic adults, 31.7% of women and 16.7% of men said they had received an earlier psychiatric diagnosis that they considered a misdiagnosis, most often a personality disorder, followed by anxiety and mood disorders^[16]. Among people with an earlier diagnosis, women and men were equally likely to call it a misdiagnosis; the gap arose because women had more often been given one. These are personal judgements, not clinical re-assessments, in a relatively highly educated sample.

Two areas come up again and again. Borderline personality disorder shares features with autism, including social and emotion-regulation difficulties, but a 2021 meta-analysis found the research, mostly small clinical samples at high risk of bias, too limited to say whether one is diagnosed in place of the other^[17]. A 2025 meta-analysis of 22 studies found that 29% of children and adults with anorexia nervosa scored above the autism cut-off on a standard observational assessment^[18], though a screening score is not a diagnosis. Our article on autism and eating disorders covers this in more depth.

What being missed can cost

A diagnosis does not fix anything by itself, but going without one can mean years without an explanation, adjustments or support that fits. Research on the cost is mostly cross-sectional, so it shows associations, not proof that late diagnosis causes harm.

In the US SPARK study of 4,657 autistic adults, those diagnosed at 21 or older had 2.71 times the odds of reporting a psychiatric condition compared with those diagnosed in childhood, and female participants had 1.68 times the odds of male participants (the study recorded sex at birth)^[19]. Camouflaging, linked to later diagnosis in women, is itself associated with more anxiety and depression in autistic adults, modestly once autistic traits and age are allowed for, and similarly across genders^[20].

Sex differences also appear in the most serious outcomes, though this study did not examine timing of diagnosis. In a Danish national register study of 6.56 million people aged 10 and over, followed from 1995 to 2016, the rate of suicide attempts among autistic females was 4.41 times that among autistic males; among non-autistic people, the equivalent female-to-male ratio was 1.41^[21]. More than nine in ten autistic people who attempted suicide also had at least one other diagnosed condition, pointing to health needs that can be recognised and treated. Suicide is not an inevitable part of being autistic: timely recognition, adapted mental-health care and support that reduces isolation are all things services can provide. Our article on autism and suicide research sets out what prevention research recommends.

What the evidence asks of us

Several of the steps the evidence points to already appear in official guidance.

• Train professionals to recognise autism in girls and women. England's national autism strategy commits to training that covers "the presentation of autism in women and girls", and accepts that current approaches "identify a higher proportion of boys than girls as autistic"^[22]. Researchers call for mental-health practitioners to be trained in adult presentations^[16], and for "greater clinician and key stakeholder awareness" of camouflaging^[13].
• Treat mental-health contact as a chance to recognise autism. NICE guidance already lists current or past contact with mental health services, and a history of mental disorder, among the circumstances in which clinicians should consider an autism assessment for adults with persistent social or behavioural differences^[23]. The Swedish and Dutch data suggest this is where many women could be recognised sooner.
• Fund adult assessment, and design it to reach women. The researchers who estimated how many autistic adults in England are undiagnosed describe "an urgent need to improve access to adult autism diagnostic services"^[2]. NICE asks local autism pathways to promote access for women specifically^[23].
• Report by sex, and read ratios carefully. Services and statistics bodies should publish diagnosis and waiting data by sex and age. As the CDC's figures show, a narrowing ratio can sit alongside a widening gap^[5].
• Close the research gaps. Research funders should back studies that include autistic women with intellectual disability, trans and non-binary autistic people, and countries beyond Northern Europe, North America and Australia.

Language: this piece uses identity-first language ("autistic people"), the preference of most autistic-led organisations, and uses clinical terms such as "autism spectrum disorder" only when naming a study or diagnostic category. Most studies cited record sex as male or female, and the camouflaging and misdiagnosis surveys asked about gender; we use "girls and women", "boys and men", "females" and "males" to reflect the categories each study used.

Why we're publishing this

Health Insurance UK is a commercial health-insurance resource, not a charity and not a campaign. We compiled this evidence because autistic women and girls, their families and the autistic-led organisations working alongside them deserve the numbers clean, current and in one place, ready to put in front of a commissioner, an MP or a journalist. The case is theirs. Our part is to say plainly what the research shows: autistic girls and women have been diagnosed later and less often, the gap is closing in some countries, and the years unrecognised have had real costs. This article is part of our Autism & Health research library.

How to read this data

The ratios here are measured in different ways: the meta-analysis figure is a male-to-female odds ratio that allows for the sex balance of the non-autistic population; the CDC figure compares prevalence among 8-year-olds; the Swedish and Danish figures count new diagnoses in national registers; the Australian figures come from a household survey; the English figures come from GP records. They are not directly comparable and should not be read as one trend. Ages at diagnosis are reported as medians, means or incidence peaks, depending on the study.

Registries record sex rather than gender, and trans and non-binary autistic people are rarely counted separately. The Dutch misdiagnosis figures are participants' own judgements. Studies linking late diagnosis with mental health are mostly cross-sectional. Very little is known about sex differences in diagnosis in low- and middle-income countries. Where only an abstract was available we used only what it reports. This is a synthesis of published research, not new research.

Use this data

Free to cite with attribution to Health Insurance UK. Charities, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Important: This article summarises published research for general information. It is not medical advice and cannot be used to diagnose, treat or predict anything about an individual. If you have concerns about your own or someone else's health, please speak to a qualified health professional.

Sources

1. Loomes 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. pubmed.ncbi.nlm.nih.gov/28545751
2. O'Nions E, Petersen I, Buckman JEJ, et al. Autism in England: assessing underdiagnosis in a population-based cohort study of prospectively collected primary care data. The Lancet Regional Health – Europe, 2023. pmc.ncbi.nlm.nih.gov/articles/PMC10114511
3. Fyfe 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. europepmc.org/article/MED/41638711
4. Zeidan J, Fombonne E, Scorah J, et al. Global prevalence of autism: a systematic review update. Autism Research, 2022. pubmed.ncbi.nlm.nih.gov/35238171
5. Shaw 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 74(SS-2), Centers for Disease Control and Prevention, 2025. cdc.gov/mmwr/volumes/74/ss/ss7402a1.htm
6. Bliddal M, Gram EB, Sonne H, et al. Trends and incidence rates of neurodevelopmental disorders in Danish children and adolescents 2000–2024. European Child & Adolescent Psychiatry, 2026. europepmc.org/article/MED/41665685
7. Australian Bureau of Statistics. Autism in Australia, 2022 (Survey of Disability, Ageing and Carers). ABS, 2024. abs.gov.au/articles/autism-australia-2022
8. Russell 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. pubmed.ncbi.nlm.nih.gov/34414570
9. Martini MI, Kuja-Halkola R, Butwicka A, et al. Sex differences in psychiatric diagnoses preceding autism diagnosis and their stability post autism diagnosis. Journal of Child Psychology and Psychiatry, 2025. europepmc.org/article/MED/40017336
10. Begeer S, Mandell D, Wijnker-Holmes B, et al. Sex differences in the timing of identification among children and adults with autism spectrum disorders. Journal of Autism and Developmental Disorders, 2013. pubmed.ncbi.nlm.nih.gov/23001766
11. Hull L, Petrides KV, Allison C, et al. “Putting on my best normal”: social camouflaging in adults with autism spectrum conditions. Journal of Autism and Developmental Disorders, 2017. pubmed.ncbi.nlm.nih.gov/28527095
12. Cruz S, Zubizarreta SC, Costa AD, et al. Is there a bias towards males in the diagnosis of autism? A systematic review and meta-analysis. Neuropsychology Review, 2025. pubmed.ncbi.nlm.nih.gov/38285291
13. Milner V, Colvert E, Hull L, et al. Does camouflaging predict age at autism diagnosis? A comparison of autistic men and women. Autism Research, 2024. pubmed.ncbi.nlm.nih.gov/38031639
14. Duvekot J, van der Ende J, Verhulst FC, et al. Factors influencing the probability of a diagnosis of autism spectrum disorder in girls versus boys. Autism, 2017. pubmed.ncbi.nlm.nih.gov/27940569
15. Lockwood Estrin G, Milner V, Spain D, Happé F, Colvert E. Barriers to autism spectrum disorder diagnosis for young women and girls: a systematic review. Review Journal of Autism and Developmental Disorders, 2021. pubmed.ncbi.nlm.nih.gov/34868805
16. Kentrou V, Livingston LA, Grove R, Hoekstra RA, Begeer S. Perceived misdiagnosis of psychiatric conditions in autistic adults. eClinicalMedicine, 2024. pubmed.ncbi.nlm.nih.gov/38596613
17. May T, Pilkington PD, Younan R, Williams K. Overlap of autism spectrum disorder and borderline personality disorder: a systematic review and meta-analysis. Autism Research, 2021. pubmed.ncbi.nlm.nih.gov/34608760
18. Inal-Kaleli I, Dogan N, Kose S, Bora E. Investigating the presence of autistic traits and prevalence of autism spectrum disorder symptoms in anorexia nervosa: a systematic review and meta-analysis. International Journal of Eating Disorders, 2025. pubmed.ncbi.nlm.nih.gov/39530423
19. Jadav N, Bal VH. Associations between co-occurring conditions and age of autism diagnosis: implications for mental health training and adult autism research. Autism Research, 2022. pubmed.ncbi.nlm.nih.gov/36054777
20. Hull L, Levy L, Lai MC, et al. Is social camouflaging associated with anxiety and depression in autistic adults? Molecular Autism, 2021. pubmed.ncbi.nlm.nih.gov/33593423
21. Kõlves K, Fitzgerald C, Nordentoft M, Wood SJ, Erlangsen A. Assessment of suicidal behaviors among individuals with autism spectrum disorder in Denmark. JAMA Network Open, 2021. pubmed.ncbi.nlm.nih.gov/33433599
22. Department of Health and Social Care. The national strategy for autistic children, young people and adults: 2021 to 2026. GOV.UK, 2021. gov.uk/government/publications/national-strategy-for-autistic-children-young-people-and-adults-2021-to-2026
23. National Institute for Health and Care Excellence. Autism spectrum disorder in adults: diagnosis and management (CG142), recommendations 1.2.2 and 1.8.3. NICE, 2012, updated 2021. nice.org.uk/guidance/cg142/chapter/Recommendations