This chapter presents the evolution of autism diagnosis and diagnostic criteria, by looking at the two main diagnostic manuals used internationally and how their definition of autism has changed over time. The chapter also looks at how diagnostic practices have evolved, through the introduction of standardised diagnostic instruments, guidelines and regulations, and lower age at diagnosis for children. The chapter also presents growing rates of autism diagnoses across countries, as well as literature exploring the reasons behind this growth. Different diagnosis rates across population groups, such as for girls and women, groups with different socio‑economic status, and ethnic and racial minorities are also discussed.
Policy Responses to Rising Autism Diagnoses in Childhood
Across the Spectrum
Abstract
Diagnosing autism: Trends and challenges
Autism diagnosis has undergone a profound transformation. Once considered a rare condition and classified under childhood schizophrenia, autism is now recognised as a common neurodevelopmental disorder with highly diverse presentations. The shift from fragmented subtypes to a single spectrum in the two main diagnostic manuals, DSM and ICD, reflects this broader understanding and has shaped diagnostic practices worldwide.
Diagnosis today is more structured and multi-disciplinary, combining clinical observation with caregiver input and standardised tools. These changes have helped identify autism earlier, though many children are still diagnosed late, often after entering school. Early diagnosis remains critical, as timely intervention can significantly improve developmental outcomes.
Autism diagnoses have risen sharply across OECD countries. This increase does not necessarily mean autism is becoming more common; rather, it reflects evolving criteria, greater awareness, and reduced stigma. Expanding the spectrum has brought into view individuals who might previously have gone unnoticed, including girls and women, whose symptoms often differ from those of boys and are harder to detect. Socio‑economic and cultural factors also shape access to diagnosis, with disadvantaged families and minority groups facing persistent barriers.
National guidelines play a central role in promoting consistency and quality. Most OECD countries recommend multi-disciplinary assessments and the use of DSM‑5 or ICD‑11 criteria, with the NICE guidelines developed in the United Kingdom widely regarded as best practice.
Autism is a neurodevelopmental disorder (NDD), a condition with onset in the early developmental period, described by impairments in social communication and characterised by restricted, repetitive behaviours. Autism was first recognised by Kenner (1943[1]), who used the term “early infantile autism” to describe the disorder and highlight the fact that early symptoms were already evident in infancy. The two internationally recognised diagnostic manuals, the American Psychiatric Association’s “Diagnostic and Statistical Manual” (DSM) and the World Health Organization’s “International Classification of Diseases” (ICD) first set out criteria for autism in 1977 (ICD‑9) and 1980 (DSM-III), respectively (APA, 1980[2]; WHO, 1977[3]).
Autism-like behaviours were long considered to be a type of childhood schizophrenia by both the DSM (DSM-I, 1952 and DSM-II, 1968) and the ICD (ICD‑9, 1977) (see Table A A.1). In 1980, the third edition of the DSM (DSM-III, 1980) classified “infantile autism” as separate from schizophrenia for the first time, listing “infantile autism” as a subtype of pervasive developmental disorder (PDD) (APA, 1980[2]). However, the ICD continued to consider autism as a type of schizophrenia for over a decade, until the publication of the 10th revision (ICD‑10) in 1992 (WHO, 1992[4]).
For about three decades, from the 1980s to the 2010s, autism was classified under pervasive developmental disorder, i.e. a type of severe, early developmental disorder characterised by delays and distortions in the development of social skills, cognition and communication. The DSM-III (1980) defined four subtypes of PDD (see Table A A.2) including infantile autism and childhood-onset PDD, with onset before and after 30 months of age respectively, as well as atypical PDD (an autism-like condition) and residual infantile autism (APA, 1980[2]). The ICD‑10 (1992) included a total of eight subtypes (see Table A A.2) including childhood autism and atypical autism, similar to the DSM-III but also other forms of autism, including Asperger syndrome, other and unspecified PDD, as well as other disorders such as Rett syndrome, other childhood disintegrative disorder, etc. (WHO, 1992[4]).
The fourth edition of the DSM (DSM-IV, 1994) included five subtypes (see Table A A.2) – excluding overactive disorder associated with mental retardation and “stereotyped movements” (repetitive, non-functional motor behaviours that are commonly observed in youth with autism, e.g. arm flapping, hand flapping, rocking back and forth), and combining other and unspecified PDD under PDD-not otherwise specified (PDD-NOS), which included atypical autism (APA, 1994[5]).
These subtypes were almost the same as those included in the ICD‑10, a sign of efforts to streamline disorders between the two diagnostic manuals. Consistency between the two diagnostic manuals contributed to the development of standardised assessment methods and facilitating research, which in turn lead to a dramatic increase in the number of scientific publications on autism. In addition, the focus of DSM-IV and ICD‑10 on consistent application of diagnostic criteria across functional levels contributed to increased awareness of severe social and communicational impairments in cognitively more able individuals, increasing access to services for this population group (Volkmar and McPartland, 2014[6]).
Progressively, from the publication of the DSM-IV (1994) onwards, clinicians started referring to three of the subtypes of PDD in the DSM-IV: autistic disorder, Asperger’s syndrome and PDD-NOS; as autism spectrum disorder (ASD) (see Table A A.3). Instead of referring to autism under three different clinical disorders, autism started to be considered as more of a spectrum with all three disorders part of the same clinical condition, but with differing levels of severity. Asperger’s and PDD-NOS were considered milder versions of autism, while autistic disorder was seen as a more severe version (Tsai, 2014[7]).
In 2013, autism spectrum disorder was included in the fifth edition of the DSM (DSM‑5, 2013). The DSM‑5 also dropped the subtypes (Asperger’s, PDD-NOS, and others) that had appeared in previous editions (APA, 2013[8]). Almost a decade later the ICD followed with the 11th revision (ICD‑11, 2019), which also replaced the previous subtypes by introducing ASD (WHO, 2019[9]).
Diagnostic criteria between the DSM-IV and the DSM‑5 differ quite significantly, suggesting an important shift in practice when it comes to diagnosing autism (see Table 2.1). This change was similar between the ICD‑10 and the ICD‑11. Other than the shift from PDD to ASD, the diagnostic domains were also slightly modified, grouping impairments in social interaction and impairment in communication together, reducing the diagnostic domains from three to two. Another important change between the DSM-IV/ICD‑10 and the DSM‑5/ICD‑11 was the required age of onset for diagnosing the disorder. While the DSM-IV and the ICD‑10 required that the age of onset be no later than 3 years of age, the DSM‑5 and the ICD‑11 were more permissive, allowing for diagnosing even if symptoms do not become apparent until later, when “social demands exceed limited capacities” (APA, 2013[8]; WHO, 2019[9]) (see Table 2.1).
ADHD = Attention deficit/hyperactivity disorder
Note: 1. The DSM-IV-TR (2000) classified Rett’s disorder as a separate genetic disorder, distinct from PDD (APA, 2000[10]). 2. Stereotypical movements refer to repetitive, non-functional motor behaviours that are commonly observed in youth with autism e.g. arm flapping, hand flapping, rocking back and forth (Olson, Bishop and Thurm, 2024[11]).
Source: American Psychiatric Association (1994[5]), Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), https://doi.org/10.1176/appi.books.9780890420614.dsm-iv; American Psychiatric Association (2013[8]), Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM‑5), https://psychiatryonline.org/doi/book/10.1176/appi.books.9780890425596.
In 2022, the DSM published a text-revision to the DSM‑5 (DSM‑5‑TR). This edition did not change the core diagnostic criteria for ASD but put more focus on different presentations of symptoms depending on cultural context and sex. The DSM‑5‑TR also included a more explicit recognition of ASD and co‑occurring conditions such as attention-deficit/hyperactivity disorder (ADHD), anxiety, and intellectual disabilities (APA, 2022[12]).
Co-diagnosis of ASD with Stereotypic Movement Disorder (ADHD in the DSM-IV-TR and onwards, 2000) was not possible under the DSM-IV but it became possible to co-diagnose ADHD under the DSM‑5. ADHD is now recognised as one of the most common co‑occurring conditions with ASD, affecting on average 28% of children with autism and even higher among older children and adolescents (Lai et al., 2019[13]).
Understanding of autism and diagnostic criteria for the disorder has changed considerably in the past 70 years. Until around 20 years ago, autism was considered to be a rare developmental disorder, generally accompanied by some level of intellectual disability (Constantino and Charman, 2016[14]). Now, autism is understood as a very common neurodevelopmental disorder, with very heterogenous presentations, hence the introduction of the term “spectrum” (Mottron and Bzdok, 2020[15]; Zeidan et al., 2022[16]).
Autism diagnosis in children usually relies on two main sources of information: description by parents or caregivers of the child’s development, and observation of the child’s behaviour by a professional or professionals (CDC, 2024[17]), as recommended by the DSM‑5 (APA, 2013[8]). A typical diagnostic process is described more in detail in Box 2.1.
Autism, like other neurodevelopmental conditions, is diagnosed through a series of structured observations and interactions to determine whether an individual meets the criteria for diagnosis. Screening and diagnostic practices do vary between countries (see “Diagnostic guidelines and regulations”), but there are some clear similarities in terms of overall approach.
1) Screening for autism in primary care
Screening for autism often begins in primary care settings, often during routine check-ups with a general practitioner or paediatrician. This may be done systematically for all children – as recommended by the American Academy of Paediatrics – or prompted by concerns from parents, healthcare, or childcare/educational providers when developmental milestones are not met.
Several screening tools are available to help identify children who may need further assessment. Commonly used tools include:
Ages and Stages Questionnaires SE‑2 (ASQ-SE2);
Communication and Symbolic Behaviour Scales (CSBS);
Modified Checklist for Autism in Toddlers (M-CHAT);
Childhood Autism Spectrum Test (CAST).
Among these, the M-CHAT and CAST are the most widely used. These relatively brief questionnaires (23 and 38 items, respectively) are completed by the child’s caregiver(s) or the paediatrician and help determine whether further diagnostic evaluation is needed.
2) Multi-disciplinary assessment in specialist care
If screening indicates a likelihood of autism, the child may be referred to a specialist or a specialist multi-disciplinary team for a comprehensive diagnostic assessment. The evaluation focusses primarily on two areas:
Difficulties with social communication and interaction
Repetitive behaviours, restricted interests, or activities
To ensure a thorough understanding of the child’s functioning, observations should ideally take place across multiple settings, such as at home, at school, or at kindergarten.
The diagnostic assessment may include:
Autism-specific tools that support behavioural observation, such as the Autism Diagnostic Observation Schedule – Second Edition (ADOS‑2), or the Childhood Autism Rating Scale, Second edition (CARS‑2); and diagnostic tools that are based on parent and/or caregiver interview, such as the Autism Diagnostic Interview – Revised (ADI-R) or the Developmental, Dimensional and Diagnostic Interview (3di).
Cognitive and adaptive functioning assessments to evaluate general cognitive abilities and adaptive behaviours, and to identify possible co‑occurring intellectual disability.
Communication and language evaluation to assess language development and functional communication skills, and screen for co‑occurring language disorders.
Physical and medical examination including e.g. vision and hearing checks, growth measurements (height, weight, head circumference), and screening for conditions such as epilepsy or sleep, feeding, and gastrointestinal problems, etc.
Behavioural and mental health assessment to identify co‑occurring psychiatric or behavioural conditions such as anxiety, mood disorders, ADHD, OCD, tic disorders, or conduct problems.
Developmental history gathered through interviews with parents and input from teachers or other caregivers.
The diagnostic team may include:
A clinician trained in autism diagnosis, such as a child psychiatrist, a developmental psychologist, a developmental paediatrician, or a paediatric neurologist.
A speech and language therapist, to assess social communication skills.
An occupational therapist and/or a physiotherapist, to evaluate motor skills and functional abilities.
Other professionals, such as a social worker, may assess family dynamics and support needs to ensure an appropriate environment for the child.
ADHD = Attention-deficit/hyperactivity disorder; OCD = Obsessive‑compulsive disorder.
Note: The exact diagnostic process may vary according to country and clinical practice.
Source: American Academy of Pediatrics Committee on Children with Disabilities Autism Subcommittee (2025[18]), How Is Autism Diagnosed?, https://www.healthychildren.org/English/health-issues/conditions/Autism/Pages/Diagnosing-Autism.aspx; Autism SA (n.d.[19]), The diagnostic process, https://autismsa.org.au/autism-diagnosis/autism-diagnosis-process/the-diagnostic-process/; eMentalHealth.ca (n.d.[20]), Screening Tool: Autism spectrum disorder, https://www.ementalhealth.ca/index.php?m=survey&ID=61; Barthélémy et al. (2019[21]), People with Autism Spectrum Disorder: Identification, Understanding, Intervention – Third edition, https://www.autismeurope.org/wp-content/uploads/2019/09/People-with-Autism-Spectrum-Disorder.-Identification-Understanding-Intervention_compressed.pdf.pdf; Mayo Clinic (2025[22]), Autism spectrum disorder, https://www.mayoclinic.org/diseases-conditions/autism-spectrum-disorder/diagnosis-treatment/drc-20352934.
Age at diagnosis has been going down
Numerous studies have demonstrated that intensive, early intervention programmes can help improve the cognitive and language abilities, as well as adaptive behaviour in children with ASD, possibly due to higher neuroplasticity (the capacity of the nervous system to modify itself, both functionally and structurally, in response to experience, e.g. injury) earlier in life (Daniels and Mandell, 2013[23]). Evidence suggesting that early intervention can contribute to improved outcomes has likely contributed to diagnoses of ASD at younger ages (Constantino and Charman, 2016[14]; Hus and Segal, 2021[24]).
A study by Daniels and Mandell (2013[23]) reviewing literature on autism diagnoses between 1990 and 2012 found that the mean age at diagnosis for ASD ranged from 38 to 120 months globally, with a decrease in age at diagnosis over time, although most people with autism tend to get diagnosed once they enter the education system. According to a more recent study from the United Kingdom, this is especially the case if emotional and behavioural difficulties overlap with autism, with the age of diagnosis at 8‑14 years of age (Mandy et al., 2022[25]). The same paper found that only a third of the study participants were diagnosed with autism by age 7, as reported by parents (Mandy et al., 2022[25]).
Similar findings were observed in Western Australia, where a population-based cohort-study by Nassar et al. (2009[26]) found that for children born in the early 1990s, the prevalence of autism was highest for children aged 4‑5, whereas for the cohort born in the mid‑1990s, prevalence was highest for 2‑3 year‑olds.
A follow-up to the study by Daniels and Mandell (2013[23]) reviewed studies published between 2012‑2019 and found that the reported mean age at diagnosis is 60.5 months, ranging between 31 and 235 months (van ’t Hof et al., 2020[27]). A subgroup analysis focussing on studies including children aged 10 years or younger found that the mean age of diagnosis was at 43 months on average, ranging between 31 and 75 months (van ’t Hof et al., 2020[27]). The results of this meta‑analysis show that continuous efforts are being made in various countries to lower the age at diagnosis, but many children are still diagnosed quite late, well after they reach school-age (van ’t Hof et al., 2020[27]).
Diagnosis in children as young as two years of age is relatively stable, i.e. it is unlikely to change over time, especially if diagnosis is done by a multi-disciplinary team of experienced clinicians (Constantino and Charman, 2016[14]). According to research, an autism diagnosis can be reliably set as of 24 months of age (Daniels and Mandell, 2013[23]). Although diagnosis is possible in children as young as 18 months old, these diagnoses tend to be less stable, as there can be overlap with other disorders and conditions (especially other NDDs), as well as difficulty in assessing the extent of impairment in social functioning (Constantino and Charman, 2016[14]).
Standardised diagnostic tools are increasingly being used for assessments
Diagnosis of autism in the last two decades has benefited from the development of standardised measures in the form of screening and diagnostic instruments to measure the symptoms of ASD, developed at the end of the 20th and at beginning of the 21st century.
These diagnostic tools can be in the form of a checklist questionnaire to screen and rapidly assess symptom severity, the most known and widely validated tools being the Autism Diagnostic Interview, Revised (ADI-R) and the Developmental, Dimensional and Diagnostic Interview (3di). Standardised instruments can also be in the form of observational measures, such as the Autism Diagnostic Observation Schedule, second edition (ADOS‑2), which together with the ADI-R is the best studied diagnostic instrument, and is often seen as the “gold standard” for diagnosing autism (Constantino and Charman, 2016[14]; Zeidan et al., 2022[16]; Zwaigenbaum and Penner, 2018[28]; Wolff and Piven, 2021[29]).
It is important to note that the utility of these tools is contingent on the expertise and level of training of the professionals administering them (Zeidan et al., 2022[16]). There has been a growing recognition of the limits of the clinician standard and tools for diagnosing ASD (Constantino and Charman, 2016[14]). A study by Roman-Urrestarazu et al. (2021[30]) found that standardised procedures, protocols, and diagnostic tools and instruments used by clinical teams are inconsistent in diagnosing autism in children, especially when it comes to racial and ethnic minorities.
Standardised diagnostic tools are an important part of the diagnostic process but should be used in addition to information gathered from other sources and observations by a clinician trained specifically in autism. International NGO and advocacy group Autism Europe suggests that ideally all members of the multi-disciplinary assessment team should have autism-specific training but even if that is not possible, diagnosis should not rely on a single standardised diagnostic tool alone: “A valid diagnosis depends on expert clinical judgement based on information gathered from all relevant sources” (Barthélémy et al., 2019[21]).
New technology-enabled procedures for ASD diagnosis are currently being tested. Some AI applications, especially large language models (LLMs), have shown promise in supporting diagnostic processes. While still in its infancy, research suggests that AI has the potential to support autism diagnosis with a high degree of accuracy, although ethics and equity concerns will need to be considered to progress the introduction of AI diagnosis in public services. One study, for instance, used video-based AI to detect “stereotypical” motor movements in children with autism with over 90% accuracy (Barami et al., 2024[31]). Other studies have explored the use of AI in early diagnosis before the age of 3 (Bussu et al., 2018[32]), distinguishing between autism and ADHD (Duda et al., 2016[33]), and identifying atypical facial expressions and motor patterns (Liu, Li and Yi, 2016[34]; Li et al., 2017[35]; Anzulewicz, Sobota and Delafield-Butt, 2016[36]). More recent work has focussed on leveraging large language and deep learning models to reduce clinician bias and improve identification across the autism spectrum, though these models require large, diverse datasets and integration with electronic health records to reach their full potential (Stanley et al., 2025[37]; Sheik Abdullah et al., 2025[38]; Ibadi and Lakizadeh, 2025[39]). An Israeli longitudinal study also found that automatic prediction models can achieve high accuracy in early prediction of ASD (Amit et al., 2024[40]).
Pressure on diagnostic services is increasing
Growing awareness of autism has contributed to an increase in the number of referrals from primary care doctors, and growing demands for accessing diagnostic services (Russell et al., 2021[41]; Monteiro et al., 2015[42]). However, half of those referred do not end up having ASD, according to Monteiro et al. (2015[42]).
Frequent co-morbidity of ASD with other NDDs makes diagnosing ASD in children a complicated task. According to Monteiro et al. (2015[42]), 40‑55% of those diagnosed with ASD also have intellectual disability, and 29% of those referred to an ASD diagnosis ended up receiving an ADHD diagnosis instead. A more recent study – based on data from the CDC’s autism and Developmental Disabilities Monitoring Network (ADDM) – found that out of those diagnosed with ASD, 32% had co‑occurring intellectual disability, and 59% did not (Shenouda et al., 2023[43]). In Germany, an analysis done by Handelskrankenkasse – a health insurance fund providing statutory health insurance – found that more than half (53.6%) of all individuals with autism spectrum disorder had at least on other co‑occurring psychiatric disorder, such as ADHD (33.1%) or anxiety disorder (24.6%) (Nymbach, 2023[44]), complicating early identification and diagnosing,
Better developmental-behavioural training is necessary for primary healthcare providers to improve access to diagnostic and early intervention services for those with ASD (Monteiro et al., 2015[42]). Specialised autism training programmes for primary care physicians have been associated with positive changes in the knowledge and self-efficacy of physician’s care of patients with autism (Clarke and Fung, 2022[45]).
The WHO and UNICEF recommend well-care visits for children and adolescents in the form of regular check-up by healthcare providers to monitor health growth, development and well-being of children notably in the first two decades of their lives. Most countries already recommend universal or routine healthcare contacts, for antenatal, childbirth and postnatal care, as well as vaccinations. Well-care visits provide a critical platform for early identification of conditions such as autism, and developmental surveillance, and have the potential to reduce pressure on diagnostic services (WHO/UNICEF, 2023[46]). In line with these recommendations, Israel has a wide network of primary healthcare centres called Tipat Halav (טיפת חלב, meaning “drop of milk”) that provide postnatal and early childhood care. The majority of children in Israel go to these centres for regular check-ups, notably to make sure that they are meeting the appropriate developmental milestones. These centres and well-care visits are central to early screening for autism.
Global ASD prevalence is estimated at 1‑2% with relative consistency across international studies, making autism one of the most common neurodevelopmental disorders (Mandy et al., 2022[25]; Zwaigenbaum and Penner, 2018[28]). In the past 20 years, the number of people diagnosed with ASD has been increasing globally. Currently, it is unclear whether the increase in the number of people diagnosed with autism over the past 20‑30 years point to a “real” increase in the prevalence of autism, or rather an increase in the number of people diagnosed with the disorder. An increased rate of diagnosis will have been influenced by increased awareness of ASD, falling stigma around ASD, and changes to definition of autism as a disorder (see previous discussion).
Diagnosis rates and prevalence estimates are difficult to compare across countries
It is important to distinguish between autism prevalence estimates and diagnostic rates. Prevalence refers to the estimated proportion of people with a disease or condition in the overall population. Prevalence rates can be estimated through a population survey, or assumptions based on other indicators such as rate of diagnosis of a condition or contact with health services. True prevalence should seek to capture the actual proportion of persons with a condition in the population, regardless of whether they have been diagnosed or not. Diagnostic rate, by contrast, refers to the actual number of individuals diagnosed with a condition in a given population. Rate of diagnosis may be captured through clinical records, or surveys where respondents are asked whether they have received a diagnosis of autism.
A range of academic literature seeks to give an estimated prevalence for the rate of autism in the population. Most of the literature identified for this paper relies at least in part on rates of diagnosis (for example, recorded diagnoses or parent-reported diagnoses); such “prevalence” estimates can be expected to be influenced at least to a degree by trends in diagnostic practice for autism. Similarly, certain countries rely on diagnosis data to estimate “prevalence”. For instance, Canada has estimated “prevalence” rates for autism going back to 2000, gathered in the Canadian Chronic Disease Surveillance System. However, this database uses data from public hospital discharges and physician billing claims, making the “prevalence” susceptible to changes in diagnostic practices.
Several systematic and meta‑analyses have explored ASD prevalence in the population. These studies point to variations in estimations between countries, and over time, with a general trend towards an increase in measured autism prevalence globally. Zeidan et al. (2022[16]) reviewed prevalence estimates worldwide in the period up to 2021, and found an approximate rate of diagnosis of autism in children of 1%. A systematic review and meta‑analysis by Salari et al. (2022[47]) found an average global ASD prevalence rate of 0.6%, ranging from 1.7% in Australia, 1% in America and Africa, to 0.5% in Europe and 0.4% in Asia, with large variability in sample size and number of studies per continent. A mixed-effects meta‑analysis by Talantseva et al. (2023[48]) established a global prevalence estimate of 0.72%. Similar regional estimates were found in both the study by Talantseva et al. (2023[48]) and Salari et al. (2022[47]), as well as a clear trend in which higher income countries tended to report higher prevalence estimates, which was suggested as being associated with better detection of autism in wealthier countries (Talantseva et al., 2023[48]). This study also found that estimates were higher in prevalence studies that used records-review surveillance rather than other designs (i.e. studies that have a trained expert review a range of clinical and educational information obtained in routine practice, rather than only information from health insurance or administrative databases).
National-level prevalence estimates have been established through a range of different methods. For example, in the United States the Autism and Developmental Disabilities Monitoring (ADDM) Network is an active surveillance system that provides estimates of ASD amongst 8‑year‑olds in 11 sites (Baio et al., 2018[49]; Shaw et al., 2025[50]). This approach has two phases: review of evaluations by service providers in the community (record review) which includes a wide variety of data sources from general paediatric health clinics to specialised programmes, and a second phase of review of all extracted information by experienced clinicians to determine ASD case status. The latest wave of this study, for 2022, found ASD prevalence to be 31.2 per 1 000 children aged 8 years, ranging from 9.7 in Texas to 53.1 in California.
According to a study on autism diagnosis rates in patient health records and insurance claims in the United States between 2011 and 2022, the diagnosis rate increased almost threefold during the study period among 5‑8 year‑olds (from 2.3 per 1 000 in 2011 to 6.3 per 1 000 in 2022), and between four‑ and fivefold among 26‑34 year‑olds. The increase was also greater for women than for men. Among children, the relative increase was greater in ethnic minority groups compared to those identifying as White (Grosvenor et al., 2024[51]), suggesting that hitherto underdiagnosed populations are catching up.
Autism rates in selected OECD countries vary but show similar trends
While comparisons across countries are challenging with markedly different approaches to estimate autism trends, it is clear for all countries that the number of ASD diagnoses has been increasing (Figure 2.1). Countries take different approaches to estimating the rate of autism in their populations. For example:
Australia, Canada and the Netherlands undertake surveys of key populations to capture self-reported rates of ASD.
In Australia, respondents are asked whether they have a long-term health condition but not whether they have a diagnosis of the condition.
In Canada, the person most knowledgeable about the child or youth (usually a parent) reports on the child or youth’s diagnosed autism status.
In the Netherlands, the National Health Survey only started to record data on autism in 2022 making comparisons over time challenging. The data show important fluctuation in the 3‑year period for which data is available, probably related more to measurement bias than a real change in prevalence. Data also identify very high rates of autism, though influenced by both a self-reporting bias and the inclusion of suspected cases of autism.
Denmark and Sweden have a national patient registry that records diagnoses of autism established in the health system. Similarly, Estonia and Germany do not collect national prevalence data, but information is available on inpatient and/or outpatient hospital consultations – these measures can be used to give an indication of the rate of health system contacts for persons with an autism diagnosis. Data for Germany is excluded from Figure 2.1, as the only data available concerns main medical diagnosis for inpatient hospitalisations. Since autism by its nature does not require hospital care, this data is not considered representative enough to be included.
In Israel, data on children diagnosed with ASD who receive state‑funded services are collected by different agencies but are not consolidated into a single national registry. The most reliable data source is the National Insurance Institute’s registry of children eligible for the Disabled Child Allowance. This allowance is granted to all children diagnosed with ASD, regardless of their level of support needs, thereby strengthening the assumption that a very high share of those diagnosed are captured in this database.
The United Kingdom measure ASD prevalence among the adult population through their Adult Psychiatric Morbidity Survey. Since this survey data is quite different from what other survey data show, this data is not shown in Figure 2.1 (see Box 2.2 for more details).
In the United States, several different data sources exist that have been used to estimate national ASD prevalence. The most-commonly referred data is gathered by the ADDM Network, which reviews a range of records from the health and educational systems in a select number of states.
There are different patterns between countries, which may reflect the measure used to understand autism rates. For example, in Estonia (and Germany), where contacts with the health system are used to identify ASD cases, there was a clear inflection around 2020‑2021 which is likely attributable to the COVID‑19 crisis.
Per 100 000 in each category
ADDM: autism and developmental disabilities monitoring; ASD: autism spectrum disorder; NSCH: National Survey of Children’s Health; SECC: Special Education Child Count.
Note:
Australia: Respondents are asked whether they have a long-term health condition which has lasted for 6 months or more, autism being one of the conditions they can report.
Canada: Data cover children aged 1‑17 with autism. Parents are asked whether their child was diagnosed with autism/autism spectrum, autistic disorder, Asperger’s disorder or pervasive developmental disorder.
Denmark: OECD calculations based on data from the National Patient Registry (Landspatientregistret, LPR), which captures new and previously (up to 5 years) made diagnoses (Danish ICD‑10‑equivalent codes for F84, including F84.0, F84.1, F84.5, F848 and F84.9) for children and adolescents (0‑17).
Estonia: OECD calculations based on data from the National Institute for Health Development (Tervise Arengu Instituut), which captures new and prevalent outpatient case consultations by a psychiatrist by diagnosis (F84).
Israel: Any child between the age of 91 days and 18 years and 3 months with an autism diagnosis who benefits from the disabled child allowance.
Netherlands: OECD calculations based on data from the Dutch Central Bureau of Statistics (Centraal Bureau voor de Statistiek, CBS). Respondents are asked if they or their child has one or more chronic disorder (for 6 months or longer), autism spectrum disorder (until 2023, the question included Asperger’s syndrome and pervasive developmental disorder not otherwise specified) being one of the disorders they can report.
Sweden: Diagnoses of pervasive developmental disorders (F84) in inpatient and/or specialised open care as the number of patients per 100 000 habitants.
United States: ADDM Network data is collected from health and/or education records of 8‑year‑old children to estimate the number of 8‑year‑old children with ASD. The sites do not collect nationally representative data. Medicaid data shows administrative claims reported by states to the Centers for Medicare and Medicaid Services. Data represents children aged 3‑17 who receive Medicaid benefits. Data from 2014 and 2015 are excluded due to data quality issues for diagnosis codes. Data from the National Survey of Children’s Health (NSCH) is collected annually through a cross-sectional address-based survey that collects information on the health and well-being of children aged 0‑17 years. Special Education Child Count refers to administrative data collected by the U.S. Department of Education, reporting the number of children.
Source: Australian Bureau of Statistics (2024[52]), Autism in Australia, https://www.abs.gov.au/articles/autism-australia-2022#data-downloads; Statistics Canada (2025[53]), Health indicator statistics for children and youth aged 1 to 17 years, parent reported, https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1310094701; Social- og Boligstyrelsen (2024[54]), Demografi – børn og unge [Demographics – Children and young people], https://www.social.dk/data/databank/data-om-borgere/demografi-boern-og-unge; Tervise Arengu Instituut (2024[55]), PKH1: New and prevalent outpatient case consultations by psychiatrist by diagnosis, sex and age group, https://statistika.tai.ee/pxweb/en/Andmebaas/Andmebaas__02Haigestumus__05Psyyhikahaired/PKH1.px/; CBS (2025[56]), Health and healthcare; personal characteristics https://opendata.cbs.nl/#/CBS/en/dataset/85454ENG/table?searchKeywords=health; Socialstyrelsen (2025[57]), Statistikdatabas för diagnoser, https://sdb.socialstyrelsen.se/if_par/resultat.aspx; CDC (2025[58]), Autism Data Visualization Tool, https://www.cdc.gov/autism/data-research/autism-data-visualization-tool.html. Data for Israel was provided by national authorities.
Every 10 years or so, the United Kingdom’s National Health Service (NHS) conducts the Adult Psychiatric Morbidity Survey (APMS) among the adult population (aged 16 and over). These surveys provide a series of data on both treated and untreated psychiatric disorders in England:
Adult Psychiatric Morbidity in England – 2007, Results of a household survey
Adult Psychiatric Morbidity Survey: Survey of Mental Health and Well-being, England, 2014
Adult Psychiatric Morbidity Survey: Survey of Mental Health and Well-being, England, 2023/24
Autism prevalence is typically estimated through a detailed validation assessment. This means that, contrary to population-level surveys in other countries which measure self-reported autism (such as the Survey of Disability and Carers in Australia, the Canadian Health Survey on Children and Youth and the Dutch National Health Survey), England’s APMS measures the real prevalence. Prevalence in the APMS is estimated through the following steps:
In the phase‑one interview, autism is screened using an adapted version of the 20‑item Autism Spectrum Quotient (AQ‑20).
In the phase‑two interview, a full examination is carried out with a subset of participants by clinically trained interviewers using the Autism Diagnostic Observation Schedule (ADOS‑2).
ADOS‑2 results are weighted to generate a prevalence estimate for the general population.
According to APMS data, prevalence levels of autism have remained quite stable over time, at around 1% (Figure 2.2). Age‑specific data show a slight decline in those aged 16‑34 (from 1.7% in 2007 to 1.2% in 2023/24) and a slight increase in those aged 35‑54 (from 0.2% in 2007 to 0.5% in 2023/24).
The true prevalence of autism in the United Kingdom by broad age group, 2007, 2014 and 2023/24
ADOS: Autism Diagnostic Observation Schedule.
Note: Autism prevalence is estimated based on the profile of ADOS-examined autism amongst adults (16+) living in private households.
Source: NHS (2025[59]), Adult Psychiatric Morbidity Survey: Survey of Mental Health and Well-being, England, 2023/4, https://digital.nhs.uk/data-and-information/publications/statistical/adult-psychiatric-morbidity-survey/survey-of-mental-health-and-wellbeing-england-2023-24/autism-spectrum-disorder.
Table 2.2 provides an overview of ASD diagnoses per 100 000 population in ten OECD countries and the indicators on which these rates are based on. The nature of the indicator influences how sensitive the measure is to change and how large an increase is expected to be seen. For example, a comprehensive population survey in a population where ASD diagnoses are increasing significantly captures a bigger increase than data that captures inpatient hospitalisation which remains rare among people with autism.
People with ASD in the overall population, selected OECD countries, latest available year
ADDM: autism and developmental disabilities monitoring; ASD: autism spectrum disorder; LPR: Landspatientregistret (National Patient Registry).
Note: Calculated by the OECD. 1. ADDM Network data is collected from health and/or education records of 8‑year‑old children to estimate the number of 8‑year‑old children with ASD. The sites do not collect nationally representative data.
Source: Australian Bureau of Statistics (2024[52]), Autism in Australia, https://www.abs.gov.au/articles/autism-australia-2022#data-downloads; Statistics Canada (2025[53]), Health indicator statistics for children and youth aged 1 to 17 years, parent reported, https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1310094701; Social- og Boligstyrelsen (2024[60]), Data om borgere [Data on citizens], https://www.social.dk/data/databank; Tervise Arengu Instituut (2024[55]), PKH1: New and prevalent outpatient case consultations by psychiatrist by diagnosis, sex and age group, https://statistika.tai.ee/pxweb/en/Andmebaas/Andmebaas__02Haigestumus__05Psyyhikahaired/PKH1.px/; Statistisches Bundesamt (Destatis) (2025[61]), Diagnoses of hospital inpatients, https://www-genesis.destatis.de/datenbank/online; CBS (2025[56]), Health and healthcare; personal characteristics https://opendata.cbs.nl/#/CBS/en/dataset/85454ENG/table?searchKeywords=health; Socialstyrelsen (2025[57]), Statistikdatabas för diagnoser, https://sdb.socialstyrelsen.se/if_par/resultat.aspx; NHS (2025[59]), Adult Psychiatric Morbidity Survey: Survey of Mental Health and Well-being, England, 2023/4, https://digital.nhs.uk/data-and-information/publications/statistical/adult-psychiatric-morbidity-survey/survey-of-mental-health-and-wellbeing-england-2023-24/autism-spectrum-disorder; CDC (2025[58]), Autism Data Visualization Tool, https://www.cdc.gov/autism/data-research/autism-data-visualization-tool.html. Data for Israel was provided by national authorities.
Changes in autism diagnosis rates
The growth in rates of ASD might be explained by a range of factors, some of which have already been mentioned, such as:
changes in diagnostic criteria (from subtypes of PDD to understanding the disorder as a spectrum);
changes in diagnostic practices (e.g. a lower age at diagnosis);
more accurate diagnostic tools (development of standardised measures);
better understanding of the disorder (growing literature around genetic and environmental causes of autism, better diagnosis in girls and adults); and
growing awareness around autism and other neurodevelopmental disorders (NDDs), thanks to the neurodiversity movement.
A Swedish study by Lundstörm et al. (2015[62]) found that while the number of clinically diagnosed individuals with ASD increased substantially in a 10‑year period between 1993 and 2002, the measured level of autism symptoms actually remained stable. Similarly, a follow-up study by Arvidsson et al. (2018[63]) found that, while more autism diagnoses were made with time, this was due to a drop in the threshold of autism symptoms necessary for a diagnosis, rather than an increase in autistic symptoms in the population.
A study from the United Kingdom (Russell et al., 2021[41]) found that the increase in the rate of autism diagnosis was higher in adults and in females, coherent with the hypothesis that expanding the autism spectrum – as was done through the DSM‑5 and the ICD‑11 – has allowed for the inclusion of more cognitively able individuals as well as of differently presenting persons who might have gone unnoticed previously, such as women (see “Girls and women”). Accordingly, changes in identification and recording of ASD diagnoses may lead to changes in prevalence rates (Russell et al., 2021[41]).
Growing autism awareness might also be contributing to the growing rates of autism. Since the beginning of the millennium, there has been growing awareness of neurodevelopmental disorders among the population, particularly autism and ADHD. The global neurodiversity movement, advocating for people with brains different from what is considered “typical” developmentally, also called “neurodiversity”, has pushed towards destigmatisation of conditions such as autism and ADHD which may have led to increased diagnosis-seeking. For instance, there has been an increased awareness of ASD in females related to expanded representation in social media, in addition to improved diagnosing tools and training programmes (Grosvenor et al., 2024[51]).
A range of research into the possible causes of autism exists, and some seeks to understand whether the underlying prevalence of autism has been increasing alongside the growth in rates of autism diagnosis. Most studies fall into two categories: genetic studies and environmental studies (Box 2.3). Current research is also focussing on identifying critical developmental windows (particularly the pre‑ and perinatal periods) to better understand how biological and environmental factors shape neurodevelopmental trajectories in autism (Barthélémy et al., 2019[21]).
Studies seeking to understand the causes of autism tend to focus on genetic and epigenetic factors, potential environmental exposures, and the interactions between genes and the environment.
Genetic research
Research over the past few decades has established that genetic factors play a major role in the development of autism, with heritability estimates consistently high. While no single mutation accounts for most cases, some genetic variants have been found to contribute to autism susceptibility in up to 40% of cases – many of which are also associated with other neurodevelopmental or psychiatric conditions such as epilepsy, ADHD, schizophrenia, and intellectual disability (Constantino and Charman, 2016[14]).
Genome‑wide techniques have identified hundreds of potentially important genetic loci – i.e. a specific, physical area on a chromosome – with evidence suggesting that both inherited and new mutations may contribute to autism risk (Barthélémy et al., 2019[21]). Some cases of autism are associated with specific genetic syndromes, such as Fragile X, neurofibromatosis, and tuberous sclerosis, as well as other chromosomal rearrangements detectable by genetic testing (Barthélémy et al., 2019[21]). The clinical heterogeneity of autism suggests that multiple genes and mechanisms are involved, and many findings require cautious interpretation.
Environmental research
In addition to genetic susceptibility, environmental factors and gene‑environment interactions are believed to play a role in autism risk, although the exact causal pathways remain unclear. Potential environmental risk factors include advanced parental age (particularly paternal), prenatal exposure to toxins (such as mercury, lead, polychlorinated biphenyls, or valproic acid), nutritional deficiencies (e.g. vitamin D), birth complications, and prematurity (Sealey et al., 2016[64]; Grabrucker, 2013[65]; Wu et al., 2016[66]; Barthélémy et al., 2019[21]). However, while many associations have been observed, clear causal mechanisms have not yet been established, and this area of research remains in its early stages (Grabrucker, 2013[65]; Modabbernia, Velthorst and Reichenberg, 2017[67]).
Source: Constantino and Charman (2016[14]), Diagnosis of autism spectrum disorder: reconciling the syndrome, its diverse origins, and variation in expression, https://doi.org/10.1016/s1474-4422(15)00151-9; Barthélémy et al. (2019[21]), People with Autism Spectrum Disorder: Identification, Understanding, Intervention – Third edition, https://www.autismeurope.org/wp-content/uploads/2019/09/People-with-Autism-Spectrum-Disorder.-Identification-Understanding-Intervention_compressed.pdf.pdf; Sealey et al. (2016[64]), Environmental factors in the development of autism spectrum disorders, https://doi.org/10.1016/j.envint.2015.12.021; Grabrucker (2013[65]), Environmental factors in autism, https://doi.org/10.3389/fpsyt.2012.00118; Wu et al. (2016[66]), Advanced parental age and autism risk in children: a systematic review and meta‐analysis, https://doi.org/10.1111/acps.12666; Modabbernia, Velthorst and Reichenberg (2017[67]), Environmental risk factors for autism: an evidence‑based review of systematic reviews and meta‑analyses, https://doi.org/10.1186/s13229-017-0121-4.
Girls and women are diagnosed less with autism
Historically autism has been viewed primarily as a “male disorder.” Boys have tended to be diagnosed with ASD at rates four to six times higher than girls (a ratio of 4‑6:1). However, more recent studies suggest that the gender gap is smaller than previously thought, with updated estimates placing the male‑to-female diagnostic ratio closer to 2‑3:1 (Duvekot et al., 2016[68]; Hamdani et al., 2023[69]; Lockwood Estrin et al., 2020[70]). This shift reflects a growing awareness amongst medical professionals that autism has likely been underdiagnosed in girls and women. Similar findings have also been reported for ADHD (Duvekot et al., 2016[68]).
One likely reason for the historically lower rate of diagnosis of autism in females is that diagnostic criteria have traditionally been based on how autism presents in males, especially young boys. It is now better understood that autism can manifest differently in girls and women (Gould, 2017[71]; Hamdani et al., 2023[69]). Studies have shown that girls and women are more likely to internalise their difficulties, leading to symptoms associated with anxiety and/or depressive disorders, whereas boys are more likely to show externalised (and sometimes disruptive) symptoms, making autism more “noticeable” in boys (Duvekot et al., 2016[68]). This can result in autistic traits going unnoticed in girls, or being attributed to other mental health issues, or learning difficulties, especially at a younger age (Hamdani et al., 2023[69]). Lockwood Estrin et al. (2020[70]) suggest that in many cases, girls under age 21 must display more severe behavioural or cognitive difficulties than boys to be considered for diagnosis: “for females to be diagnosed using existing criteria, their observable characteristics must be exaggerated to score sufficiently to warrant a diagnosis”.
Girls are also more prone to “masking” or “camouflaging” their autistic traits, i.e. adapting their behaviour in social settings to conform with peers, which makes their challenges less visible to teachers, clinicians, and even family members. These subtler presentations are more likely to emerge or be recognised during adolescence and young adulthood, when social interactions become increasingly complex (Gould, 2017[71]; Hamdani et al., 2023[69]; Lockwood Estrin et al., 2020[70]).
Girls with autism are less likely to be diagnosed with autism than boys, and when they are, they tend to be diagnosed later, which can delay access to needed supports and interventions (Duvekot et al., 2016[68]; Hamdani et al., 2023[69]).
In recognition of these differences, characteristics of what is sometimes referred to as a female “phenotype” of autism have been included in the DSM‑5‑TR, including reference to “masking” and to special interests being more similar to peers than for boys (APA, 2022[12]). However, standardised instruments often referred to as “gold standards” such as the ADOS‑2 or the ADI-R may still not accurately capture autistic traits in girls, particularly in relation to social interests or repetitive behaviours. For example, a girl with autism with restricted interest in e.g. animals or dolls, may appear more typical and socially acceptable, whereas a boy’s interests with autism e.g. in train timetables or mechanical systems, may raise more clinical concern. Likewise, girls may be less likely to engage in “stereotyped” behaviours such as lining up toys, which are key diagnostic indicators in many tools (Gould, 2017[71]; Duvekot et al., 2016[68]). In response, new diagnostic approaches are being developed. These include the Camouflaging Autistic Traits Questionnaire (CAT-Q) developed by Hull et al. (2019[72]), and a checklist of compensation strategies proposed by Livingston et al. (2020[73]), both designed to better identify autism in girls and women (Lockwood Estrin et al., 2020[70]).
Socio‑economic status has an impact on diagnostic access and outcomes
The relationship between socio‑economic status (SES) and autism diagnosis is complex and varies across countries. In many contexts, access to healthcare, awareness of autism symptoms, and parental education can influence when and whether a child receives an autism diagnosis.
In the United States, children from higher-income families are more likely to be diagnosed with autism and more likely to receive a diagnosis at a younger age than children from middle‑ or low-income families. This disparity is likely due to differences in access to healthcare and associated diagnostic tools, as wealthier families often have better access to specialised services, earlier screening, and higher-quality care. Parental education may also play a role, as more educated parents often have higher levels of autism awareness and are better equipped to navigate the healthcare system and advocate for their child’s needs (Thomas et al., 2011[74]; Durkin et al., 2010[75]). Similarly, in the United Kingdom, studies found lower autism diagnosis rates among children of mothers with lower educational attainment – 0.7% compared to 1.5% for children whose mothers had at least a high school education – possibly due to reduced awareness of developmental signs (Kelly et al., 2017[76]).
However, more recent evidence suggests that the pattern may be shifting. A large‑scale UK study found that autism prevalence was higher among children from lower SES backgrounds, with an average national prevalence of 1.8%, albeit with important regional differences (Roman-Urrestarazu et al., 2021[30]). Similar trends have been observed in other European countries, suggesting that socio‑economic disadvantage may be associated with higher underlying risk, possibly due to increased exposure to environmental stressors – e.g. heavy metal and air-pollution exposure is more common in low-income neighbourhoods – or other social determinants of health (Lyall, Schmidt and Hertz-Picciotto, 2014[77]; Volk et al., 2011[78]; Volk et al., 2014[79]; Wu et al., 2024[80]).
Public health policies and health system structures play a role in shaping diagnostic patterns. In Sweden, where children undergo routine developmental screening and where healthcare tends to be highly accessible regardless of income level, children from lower-income families and those with parents in manual occupations had a slightly higher risk of getting an ASD diagnosis – 1.4% compared to 1% for children with parents from higher SES – a reversal of the pattern observed in the United States (Rai et al., 2012[81]). The availability of free and standardised diagnostic pathways likely facilitates earlier and more equitable identification of autism across socio‑economic groups.
Diagnostic patterns in France seem like those in Sweden. A study of eight‑year-olds in South-Western France (Haute‑Garonne) found that children from more socially deprived backgrounds, single‑parent households, and immigrant families had the highest autism prevalence, particularly when associated with intellectual disability (Delobel-Ayoub et al., 2015[82]). However, these results should be interpreted with caution, as France does not collect prevalence data systematically. Since 2018, France has made efforts to collect better data under their national autism and neurodevelopmental strategy (Box 4.2). A 2022 analysis was able to use data from the national health data system in France. The study found that the proportion of patients with autism aged 0‑17 in medical-administrative databases living in adverse social conditions, gradually rose from 29% in 2010 to 42.4% in 2022 (Ponnou et al., 2025[83]). Another study on the prevalence of ASD in France (Haute‑Garonne, Isère, Savoie and Haute‑Savoie territories), based on data from the child disability database, found that there has been a significant drop in the proportion of children with autism with co-morbid intellectual disability (Delobel-Ayoub et al., 2020[84]).
Race and ethnicity may also affect ASD diagnosis rates
The relationship between race and ethnicity, and autism diagnosis is shaped by a range of social, cultural, and systemic factors. In many countries, children from racial or ethnic minority groups appear to face delays in diagnosis, possibly due to cultural stigma and unequal access to healthcare. There is also evidence of diagnostic bias, where clinicians may interpret similar behaviours differently across racial or cultural groups.
In the United States, autism prevalence differs across racial and ethnic groups, with findings varying by state. In 2022, estimated prevalence was 32.2 per 1 000 children on average, ranging from 14.3 per 1 000 in Texas to 53.1 per 1 000 in California (CDC, 2025[58]). Data from the ADDM Network sites show that prevalence has increased across all racial and ethnic groups, with a notable shift in the mid‑2010s, likely related to raising awareness in some population groups. Until 2016, children identified as “White” had the highest estimated autism prevalence; by 2022, however, they had the lowest prevalence of all racial and ethnic groups (Figure 2.3). Despite this shift, disparities remain in diagnostic patterns. For instance, in the United States, Black children are more likely to be diagnosed with co-morbid intellectual disability than Hispanic or White children, which might be a reflection of diagnostic bias (Zeidan et al., 2022[16]).
ASD prevalence in the United States by race and ethnicity, 2002‑2022
ADDM: autism and developmental disabilities monitoring, ASD: autism spectrum disorder.
Note: Estimate is based on ADDM Network data. ADDM data is not representative for each state.
Source: CDC (2025[58]), Autism Data Visualization Tool, https://www.cdc.gov/autism/data-research/autism-data-visualization-tool.html.
In Europe, patterns differ. In the United Kingdom, recent national studies have found higher autism prevalence rates among children from ethnic minority and immigrant backgrounds. A 2021 study using school census data reported the highest autism prevalence among Black pupils, while also showing that children who spoke English as an additional language or faced socio‑economic disadvantage were more likely to be diagnosed as having autism (Roman-Urrestarazu et al., 2021[30]). A follow-up study in 2022 found that Asian, Black, and Chinese girls were significantly less likely to be diagnosed than White girls. The authors suggest there may be underdiagnosis among minority ethnic girls due to language barriers, cultural stigma, and limited access to services (Roman-Urrestarazu et al., 2022[85]).
In France, a population-based study found that autism with intellectual disability was more common in areas with higher concentrations of immigrants, as well as in areas marked by social deprivation, such as higher unemployment and lower educational attainment. No such pattern was found for autism without intellectual disability, suggesting that severe or visible cases may be more likely to be identified in disadvantaged and immigrant communities (Delobel-Ayoub et al., 2015[82]).
In Israel, autism diagnosis rates are notably lower among children from certain minority population groups, including Bedouins, Israeli Arabs and – to a lesser extent – Haredi Jews (a branch of Orthodox Judaism characterised by the strict interpretation of religious sources and traditions), who represent 5%, 16% and 20% of the child population, respectively. Recent national data on the participation in the special education system show that 73% of diagnosed children are from the “Secular Jewish” population, which makes up only 42% of all children. By contrast, only 10% of diagnosed children are Haredi Jew, 7% are Arab and 1% are Bedouin (Figure 2.4). However, special education data are not representative of the entire population, as families have many and different reasons for sending their child to a regular school or a special school. Due to language and cultural reasons, children issued from minority ethnic populations (Haredi Jews, Arabs, Bedouins) may be less likely to go into the special education system, even if they have an autism diagnosis, compared to children from the majority Jewish population. Indirect estimates derived from district data suggest large underdiagnosing in neighbourhoods with a large Arab population but not in neighbourhoods with a large Haredi population. A recent study based on district data from the National Insurance Institute found that Jewish children residing in predominantly Haredi neighbourhoods became eligible for the Disabled Child Allowance (an allowance given to all Israeli children with an ASD diagnosis) significantly later (on average at 4.7 years) than Jewish children from non-Haredi neighbourhoods (on average at 2.7 years) (Silverman, Amit and Sadaka, 2026[86]).
Distribution across ethnic and religious groups in the total child population and among children in the special education system diagnosed with ASD, 2024
ASD: autism spectrum disorder.
Source: Data provided by the Israeli national authorities.
Most OECD countries have a national guideline for health professionals on the diagnosis of autism, including recommendations around the diagnosing professional(s), the diagnostic criteria to be applied (e.g. DSM‑5 or ICD‑11), and the diagnostic tools to be used, such as the ADOS or the ADI-R (see Table A B.1). Guidelines may highlight that these diagnostic instruments should be used in addition to a clinician’s observation and do not replace a full autism assessment, such as the National Guideline for the assessment and diagnosis of autism in Australia (Goodall et al., 2023[87]), France’s Best Practice Guideline for autism diagnosis and assessment (HAS, 2018[88]), Germany’s S3 guideline (the most rigorous type of medical guideline in Germany) for ASD diagnosis (AWMF, 2016[89]), the Dutch Guideline for the diagnosis and treatment of ASD in children and adolescents (NVvP, 2009[90]), as well as the United Kingdom’s NICE guideline (NICE, 2011[91]) (see Box 2.4 for more details).
The National Institute for Health and Care Excellence (NICE) in the United Kingdom has issued comprehensive guidelines on the diagnosis and treatment of autism, as well as recommendations on the support for children and adults with autism. These guidelines are renowned internationally and have been used as a source for similar guidelines in many other OECD countries.
Development of NICE guidelines follows a rigorous process
NICE guidelines are developed following a rigorous pre‑defined process. Topics are chosen based on referral from a national organisation, e.g. the National Health Service. Scoping is done in consultation with relevant stakeholders. The guideline is developed by reviewing the latest evidence, considering costs on the provision of services based on the guidelines and developing recommendations. Finally, a committee of professionals, care providers and service users consider the evidence and produce recommendations. Published guidelines are reviewed and updated regularly if needed (NICE, 2025[92]).
Currently, NICE has three clinical guidelines on autism:
Autism spectrum disorder in under 19s: recognition, referral and diagnosis (CG128). First published in 2011, last updated in 2017 and last reviewed in 2021 this guideline covers the first steps that lead up to the diagnosis of ASD in children and young people (NICE, 2011[91]).
Autism spectrum disorder in under 19s: support and management (CG170). First published in 2013, last updated in 2021 and last reviewed in 2025, the guideline covers the different ways health and social care professionals can provide support, treatment and help for children and young people with autism and their families (NICE, 2013[93]).
Autism spectrum disorder in adults: diagnosis and management (CG142). First published in 2012, last updated in 2021 and last reviewed in 2025, this guideline covers diagnosis and management of suspected or confirmed ASD in adults. It aims to improve access to health and social services and improved experience of care for people with autism (NICE, 2012[94]).
International guidelines are often inspired by NICE
NICE guidelines have achieved broad international recognition. Many OECD countries have cited NICE guidelines in their own policies or developed guidelines aligning closely with NICE’s recommendations.
In 2016, Australia published a report on good practices for support to pre‑school children with ASD and their families and carers. The NICE 2013 guidelines on support and management of autism for under 19s served as a basis (CG170) (Roberts and Williams, 2016[95]).
In 2019, Canada’s Paediatric Society published a position paper on “Standards of diagnostic assessment for autism spectrum disorder”. The 2011 NICE guidelines (Clinical Guideline 128) on recognition, referral and diagnosis of ASD in children and adolescents were one of the main guidelines reviewed (Brian, Zwaigenbaum and Ip, 2019[96]).
France’s Haute Autorité de Santé (HAS) published a guideline on good practice on intervention and life trajectory of children and adolescents with ASD. The 2011 NICE guidelines (CG128) was one of the guidelines considered for the recommendations (HAS, 2023[97]).
Germany’s S3 autism guidelines cite NICE guidelines (CG128, CG142 and CG170) as “methodologically excellent” and as the main source for the S3 guideline (AWMF, 2016[89]).
For the Netherlands’ clinical guidelines on ASD in children and youth, the NICE 2011 guideline (CG128) is one of the two main sources (NVvP, 2025[98]).
Source: NICE (2025[92]), How we develop NICE guidelines, https://www.nice.org.uk/what-nice-does/our-guidance/about-nice-guidelines/how-we-develop-nice-guidelines; NICE (2011[91]), Autism spectrum disorder in under 19s: recognition, referral and diagnosis, www.nice.org.uk/guidance/cg128; NICE (2013[93]). Autism spectrum disorder in under 19s: support and management, https://www.nice.org.uk/guidance/cg170; NICE (2012[94]), Autism spectrum disorder in adults: diagnosis and management, https://www.nice.org.uk/guidance/cg142; Roberts and Williams (2016[95]), Autism spectrum disorder: Evidence‑based/evidence‑informed good practice for support provided to preschool children, their families and carers, https://www.ndis.gov.au/media/863/download; Brian, Zwaigenbaum and Ip (2019[96]), Standards of diagnostic assessment for autism spectrum disorder, https://doi.org/10.1093/pch/pxz117; HAS (2023[97]), Trouble du spectre de l’autisme (TSA): interventions et parcours de vie de l’enfant et de l’adolescent – Note de cadrage, https://www.has-sante.fr/jcms/p_3448980/fr/trouble-du-spectre-de-l-autisme-tsa-interventions-et-parcours-de-vie-de-l-enfant-et-de-l-adolescent-note-de-cadrage; AWMF (2016[89]), S3‑Leitlinie Autismus-Spektrum-Störungen im Kindes-, Jugend- und Erwachsenenalter, Teil 1: Diagnostik, [S3 Guideline for Autism Spectrum Disorders in Childhood, Adolescence, and Adulthood. Part 1: Diagnostics], https://register.awmf.org/assets/guidelines/028_D_G_f_Kinder-_und_Jugendpsychiatrie_und_-psychotherapie/028-018l_S3_Autism_spectrum_disorders_in_childhood__adolescence_and_adulthood_2021-09_abgelaufen.pdf.
The assessment process recommended by national guidelines usually includes two steps: an assessment of needs and functioning combined with a medical evaluation, and a diagnostic assessment (see Table A B.1). The first part usually includes information on developmental history of the child, a cognitive and/or intelligence assessment, assessment of adaptive skills, assessment of language and communication level, and assessment of psycho-motor capabilities. The medical evaluation usually includes screening for hearing and visual disorders, detailed family medical history, and a clinical paediatric exam including e.g. weight, height and head circumference measurement. The diagnostic assessment usually includes two kinds of measures: an observational measure using tools, such as the ADOS, and a questionnaire filled out by the parent or caregiver, such as the ADI-R.
Most countries also suggest that the assessment be administered by at least two, or ideally a team of multi-disciplinary professionals, in line with the recommendations in the DSM‑5 (APA, 2013[8]); this is the case for example for Australia, Canada, France, Germany, the Netherlands, Sweden and the United Kingdom. It may also be suggested that these professionals are trained specifically in administering a neurodevelopmental or an autism assessment (see Table A B.1). The growing demand for diagnostic assessments has made it difficult in many countries (e.g. France, the United Kingdom) to ensure that appropriately trained multi-disciplinary teams conduct the assessments. Although the Danish Health Authority’s clinical guidelines also mention the need for a specialist assessment, the guideline mainly focusses on treatment and different intervention options and has few details on the diagnostic and assessment process. Interestingly, the Danish guideline is also the only one that includes measuring parental well-being as part of the assessment process (Sundhedsstyrelsen, 2021[99]).
Israel is the only country reviewed in this report where autism assessment for children is not issued as a guideline or a set of recommendations by a public health institution. Instead, diagnosing professionals must follow the assessment process set out in Circular n°15 of 2013 of the Ministry of Health (Ministry of Health, 2013[100]). The Circular includes information about professionals that can diagnose autism, diagnostic criteria, what kind of assessments should be included in the diagnostic process, as well as recommended assessment tools (see Table A B.1). Other legal instruments exist in several OECD countries. In 2014, France issued an “Instruction on the national framework for detection, diagnosis and early interventions for children with autism or other pervasive developmental disorders provided for in the autism plan”, but this doesn’t go into detail about the diagnostic process and focusses on laying the ground for a national autism framework instead (Ministère des affaires sociales et de la santé, 2014[101]). Similarly, in 2023 Canada’s Federal Framework on Autism Spectrum Disorder Act received Royal Assent, i.e. formal approval from the governor to become law – requiring the development of a federal framework on autism spectrum disorder. However, this Act does not set out diagnostic guidelines (Minister of Justice, 2023[102]). The 2024 Framework for Autism in Canada includes a measure to advance opportunities to develop and update national guidelines for screening, diagnosis and services (Public Health Agency of Canada, 2024[103]).
Although these guidelines are key for diagnosing ASD, they become especially relevant for countries that require a valid ASD diagnosis to access disability benefits. Requirements for diagnosis to access a disability scheme may or may not be in line with national guidelines. For instance, in Canada, the Canadian Paediatric Society’s guidelines allow for three different approaches when it comes to diagnosis, but some provinces only accept diagnosis made by an interdisciplinary or multi-disciplinary team (Approach 3) to access specialised services and benefits (Brian, Zwaigenbaum and Ip, 2019[96]).
References
[18] American Academy of Pediatrics (2025), How Is Autism Diagnosed?, https://www.healthychildren.org/English/health-issues/conditions/Autism/Pages/Diagnosing-Autism.aspx (accessed on 17 June 2025).
[40] Amit, G. et al. (2024), “Early Prediction of Autistic Spectrum Disorder Using Developmental Surveillance Data”, JAMA Network Open, Vol. 7/1, p. e2351052, https://doi.org/10.1001/jamanetworkopen.2023.51052.
[36] Anzulewicz, A., K. Sobota and J. Delafield-Butt (2016), “Toward the Autism Motor Signature: Gesture patterns during smart tablet gameplay identify children with autism”, Scientific Reports, Vol. 6/1, https://doi.org/10.1038/srep31107.
[12] APA (2022), Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, Text Revision (DSM-5-TR), American Psychiatric Association, Arlington, VA, https://doi.org/10.1176/appi.books.9780890425787.
[8] APA (2013), Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), American Psychiatric Association, Arlington, VA, https://psychiatryonline.org/doi/book/10.1176/appi.books.9780890425596.
[10] APA (2000), Diagnostic and Statistical Manual of Mental Disorders, Fourth edition, Text Revision (DSM-IV-TR), American Psychiatric Association, Arlington, VA.
[5] APA (1994), Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), American Psychiatric Association, Arlington, VA, https://doi.org/10.1176/appi.books.9780890420614.dsm-iv.
[2] APA (1980), Diagnostic and Statistical Manual of Mental Disorders, Third edition (DSM-III), American Psychiatric Association, Arlington, VA.
[63] Arvidsson, O. et al. (2018), “Secular changes in the symptom level of clinically diagnosed autism”, Journal of Child Psychology and Psychiatry, Vol. 59/7, pp. 744-751, https://doi.org/10.1111/jcpp.12864.
[52] Australian Bureau of Statistics (2024), Autism in Australia, 2022, https://www.abs.gov.au/articles/autism-australia-2022#data-downloads (accessed on 20 May 2025).
[19] Autism SA (n.d.), The diagnostic process, https://autismsa.org.au/autism-diagnosis/autism-diagnosis-process/the-diagnostic-process/ (accessed on 17 June 2025).
[89] AWMF (2016), S3-Leitlinie Autismus-Spektrum-Störungen im Kindes-, Jugend- und Erwachsenenalter, Teil 1: Diagnostik [S3 Guideline for Autism Spectrum Disorders in Childhood, Adolescence, and Adulthood. Part 1: Diagnostics], https://register.awmf.org/assets/guidelines/028_D_G_f_Kinder-_und_Jugendpsychiatrie_und_-psychotherapie/028-018l_S3_Autism_spectrum_disorders_in_childhood__adolescence_and_adulthood_2021-09_abgelaufen.pdf (accessed on 24 June 2025).
[49] Baio, J. et al. (2018), “Prevalence of Autism Spectrum Disorder Among Children Aged 8 Years — Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2014”, MMWR. Surveillance Summaries, Vol. 67/6, pp. 1-23, https://doi.org/10.15585/mmwr.ss6706a1.
[31] Barami, T. et al. (2024), “Automated Analysis of Stereotypical Movements in Videos of Children With Autism Spectrum Disorder”, JAMA Network Open, Vol. 7/9, p. e2432851, https://doi.org/10.1001/jamanetworkopen.2024.32851.
[21] Barthélémy, C. et al. (2019), People with Autism Spectrum Disorder: Identification, Understanding, Intervention - Third edition, Autism-Europe, https://www.autismeurope.org/wp-content/uploads/2019/09/People-with-Autism-Spectrum-Disorder.-Identification-Understanding-Intervention_compressed.pdf.pdf (accessed on 25 January 2025).
[96] Brian, J., L. Zwaigenbaum and A. Ip (2019), “Standards of diagnostic assessment for autism spectrum disorder”, Paediatrics & Child Health, Vol. 24/7, pp. 444-451, https://doi.org/10.1093/pch/pxz117.
[32] Bussu, G. et al. (2018), “Prediction of Autism at 3 Years from Behavioural and Developmental Measures in High-Risk Infants: A Longitudinal Cross-Domain Classifier Analysis”, Journal of Autism and Developmental Disorders, Vol. 48/7, pp. 2418-2433, https://doi.org/10.1007/s10803-018-3509-x.
[56] CBS (2025), Health and healthcare; personal characteristics, https://opendata.cbs.nl/#/CBS/en/dataset/85454ENG/table?searchKeywords=health (accessed on 28 July 2025).
[58] CDC (2025), Autism Data Visualization Tool, https://www.cdc.gov/autism/data-research/autism-data-visualization-tool.html (accessed on 20 May 2025).
[17] CDC (2024), Clinical Testing and Diagnosis for Autism Spectrum Disorder, https://www.cdc.gov/autism/hcp/diagnosis/index.html (accessed on 4 June 2025).
[45] Clarke, L. and L. Fung (2022), “The impact of autism-related training programs on physician knowledge, self-efficacy, and practice behavior: A systematic review”, Autism, Vol. 26/7, pp. 1626-1640, https://doi.org/10.1177/13623613221102016.
[14] Constantino, J. and T. Charman (2016), “Diagnosis of autism spectrum disorder: reconciling the syndrome, its diverse origins, and variation in expression”, The Lancet Neurology, Vol. 15/3, pp. 279-291, https://doi.org/10.1016/s1474-4422(15)00151-9.
[23] Daniels, A. and D. Mandell (2013), “Explaining differences in age at autism spectrum disorder diagnosis: A critical review”, Autism, Vol. 18/5, pp. 583-597, https://doi.org/10.1177/1362361313480277.
[84] Delobel-Ayoub, M. et al. (2020), La prévalence des TSA continue de croître en France : données récentes des registres des, pp. 128-135, https://hal.science/hal-04886534v1 (accessed on 11 February 2026).
[33] Duda, M. et al. (2016), “Use of machine learning for behavioral distinction of autism and ADHD”, Translational Psychiatry, Vol. 6/2, pp. e732-e732, https://doi.org/10.1038/tp.2015.221.
[68] Duvekot, J. et al. (2016), “Factors influencing the probability of a diagnosis of autism spectrum disorder in girls versus boys”, Autism, Vol. 21/6, pp. 646-658, https://doi.org/10.1177/1362361316672178.
[20] eMentalHealth.ca (n.d.), Screening Tool: Autism spectrum disorder, https://www.ementalhealth.ca/index.php?m=survey&ID=61 (accessed on 17 June 2025).
[87] Goodall, E. et al. (2023), National Guideline for the assessment and diagnosis of autism in Australia, Autism CRC, Brisbane, https://www.autismcrc.com.au/best-practice/assessment-and-diagnosis (accessed on 23 June 2025).
[71] Gould, J. (2017), “Towards understanding the under-recognition of girls and women on the autism spectrum”, Autism, Vol. 21/6, pp. 703-705, https://doi.org/10.1177/1362361317706174.
[65] Grabrucker, A. (2013), “Environmental Factors in Autism”, Frontiers in Psychiatry, Vol. 3, https://doi.org/10.3389/fpsyt.2012.00118.
[51] Grosvenor, L. et al. (2024), “Autism Diagnosis Among US Children and Adults, 2011-2022”, JAMA Network Open, Vol. 7/10, p. e2442218, https://doi.org/10.1001/jamanetworkopen.2024.42218.
[69] Hamdani, Y. et al. (2023), “Roadblocks and detours on pathways to a clinical diagnosis of autism for girls and women: A qualitative secondary analysis”, Women’s Health, Vol. 19, https://doi.org/10.1177/17455057231163761.
[97] HAS (2023), Trouble du spectre de l’autisme (TSA) : interventions et parcours de vie de l’enfant et de l’adolescent - Note de cadrage, Haute Autorité de santé, https://www.has-sante.fr/jcms/p_3448980/fr/trouble-du-spectre-de-l-autisme-tsa-interventions-et-parcours-de-vie-de-l-enfant-et-de-l-adolescent-note-de-cadrage (accessed on 1 December 2025).
[88] HAS (2018), Trouble du spectre de l’autisme : Des signes d’alerte à la consultation dédiée en soins primaires, Haute Autorité de santé, Saint-Denis La Plaine, https://www.has-sante.fr/jcms/c_468812/fr/trouble-du-spectre-de-l-autisme-signes-d-alerte-reperage-diagnostic-et-evaluation-chez-l-enfant-et-l-adolescent (accessed on 24 June 2025).
[72] Hull, L. et al. (2019), “Development and Validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q)”, Journal of Autism and Developmental Disorders, Vol. 49/3, pp. 819-833, https://doi.org/10.1007/s10803-018-3792-6.
[24] Hus, Y. and O. Segal (2021), “Challenges Surrounding the Diagnosis of Autism in Children”, Neuropsychiatric Disease and Treatment, Vol. Volume 17, pp. 3509-3529, https://doi.org/10.2147/ndt.s282569.
[39] Ibadi, H. and A. Lakizadeh (2025), “ASDvit: Enhancing autism spectrum disorder classification using vision transformer models based on static features of facial images”, Intelligence-Based Medicine, Vol. 11, p. 100226, https://doi.org/10.1016/j.ibmed.2025.100226.
[76] Kelly, B. et al. (2017), “The association between socioeconomic status and autism diagnosis in the United Kingdom for children aged 5–8 years of age: Findings from the Born in Bradford cohort”, Autism, Vol. 23/1, pp. 131-140, https://doi.org/10.1177/1362361317733182.
[1] Kenner, L. (1943), “Autistic disturbances of affective contact”, Nervous child, Vol. 2/3, pp. 217-250.
[13] Lai, M. et al. (2019), “Prevalence of co-occurring mental health diagnoses in the autism population: a systematic review and meta-analysis”, The Lancet Psychiatry, Vol. 6/10, pp. 819-829, https://doi.org/10.1016/s2215-0366(19)30289-5.
[34] Liu, W., M. Li and L. Yi (2016), “Identifying children with autism spectrum disorder based on their face processing abnormality: A machine learning framework”, Autism Research, Vol. 9/8, pp. 888-898, https://doi.org/10.1002/aur.1615.
[73] Livingston, L. et al. (2020), “Quantifying compensatory strategies in adults with and without diagnosed autism”, Molecular Autism, Vol. 11/1, https://doi.org/10.1186/s13229-019-0308-y.
[70] Lockwood Estrin, G. et al. (2020), “Barriers to Autism Spectrum Disorder Diagnosis for Young Women and Girls: a Systematic Review”, Review Journal of Autism and Developmental Disorders, Vol. 8/4, pp. 454-470, https://doi.org/10.1007/s40489-020-00225-8.
[62] Lundstrom, S. et al. (2015), “Autism phenotype versus registered diagnosis in Swedish children: prevalence trends over 10 years in general population samples”, BMJ, Vol. 350/apr28 2, pp. h1961-h1961, https://doi.org/10.1136/bmj.h1961.
[77] Lyall, K., R. Schmidt and I. Hertz-Picciotto (2014), “Maternal lifestyle and environmental risk factors for autism spectrum disorders”, International Journal of Epidemiology, Vol. 43/2, pp. 443-464, https://doi.org/10.1093/ije/dyt282.
[25] Mandy, W. et al. (2022), “Mental health and social difficulties of late‐diagnosed autistic children, across childhood and adolescence”, Journal of Child Psychology and Psychiatry, Vol. 63/11, pp. 1405-1414, https://doi.org/10.1111/jcpp.13587.
[22] Mayo Clinic (2025), Autism spectrum disorder, https://www.mayoclinic.org/diseases-conditions/autism-spectrum-disorder/diagnosis-treatment/drc-20352934 (accessed on 17 June 2025).
[102] Minister of Justice (2023), Federal Framework on Autism Spectrum Disorder Act, S.C. 2023, c. 2, https://laws-lois.justice.gc.ca/eng/acts/F-7.33/page-1.html (accessed on 24 June 2025).
[101] Ministère des affaires sociales et de la santé (2014), Instruction No. DGCS/SD3B/DGOS/SDR4/CNSA/2014/221 of 17 July 2014 on the national framework for detection, diagnosis and early interventions for children with autism or other pervasive developmental disorders provided for in the autism plan (2013-2017), https://www.legifrance.gouv.fr/download/pdf/circ?id=38551 (accessed on 24 June 2025).
[100] Ministry of Health (2013), Circular No. 15/13 of November 10, 2013 of the Director General of the Ministry of Health on diagnosing children on the autism spectrum, https://www.gov.il/he/pages/mk15-2013 (accessed on 24 June 2025).
[67] Modabbernia, A., E. Velthorst and A. Reichenberg (2017), “Environmental risk factors for autism: an evidence-based review of systematic reviews and meta-analyses”, Molecular Autism, Vol. 8/1, https://doi.org/10.1186/s13229-017-0121-4.
[42] Monteiro, S. et al. (2015), “Prevalence of Autism Spectrum Disorder in Children Referred for Diagnostic Autism Evaluation”, Clinical Pediatrics, Vol. 54/14, pp. 1322-1327, https://doi.org/10.1177/0009922815592607.
[15] Mottron, L. and D. Bzdok (2020), “Autism spectrum heterogeneity: fact or artifact?”, Molecular Psychiatry, Vol. 25/12, pp. 3178-3185, https://doi.org/10.1038/s41380-020-0748-y.
[75] Myer, L. (ed.) (2010), “Socioeconomic Inequality in the Prevalence of Autism Spectrum Disorder: Evidence from a U.S. Cross-Sectional Study”, PLoS ONE, Vol. 5/7, p. e11551, https://doi.org/10.1371/journal.pone.0011551.
[26] Nassar, N. et al. (2009), “Autism spectrum disorders in young children: effect of changes in diagnostic practices”, International Journal of Epidemiology, Vol. 38/5, pp. 1245-1254, https://doi.org/10.1093/ije/dyp260.
[59] NHS (2025), Adult Psychiatric Morbidity Survey: Survey of Mental Health and Wellbeing, England, 2023/4, NHS England, https://digital.nhs.uk/data-and-information/publications/statistical/adult-psychiatric-morbidity-survey/survey-of-mental-health-and-wellbeing-england-2023-24/autism-spectrum-disorder (accessed on 18 December 2025).
[92] NICE (2025), How we develop NICE guidelines, https://www.nice.org.uk/what-nice-does/our-guidance/about-nice-guidelines/how-we-develop-nice-guidelines (accessed on 1 December 2025).
[93] NICE (2013), Autism spectrum disorder in under 19s: support and management, https://www.nice.org.uk/guidance/cg170 (accessed on 1 December 2025).
[94] NICE (2012), Autism spectrum disorder in adults: diagnosis and management, https://www.nice.org.uk/guidance/cg142 (accessed on 1 December 2025).
[91] NICE (2011), Autism spectrum disorder in under 19s: recognition, referral and diagnosis, http://www.nice.org.uk/guidance/cg128 (accessed on 25 June 2025).
[98] NVvP (2025), Autismespectrumstoornis bij kinderen/jeugd [Autism spectrum disorder in children/young people], Federatie van Medisch Specialisten, https://richtlijnendatabase.nl/richtlijn/autismespectrumstoornis_bij_kinderen_jeugd/startpagina_-_autismespectrumstoornissen_kinderen_en_jeugd.html (accessed on 27 November 2025).
[90] NVvP (2009), Richtlijn diagnostiek en behandeling autismespectrumstoornissen bij kinderen en jeugdigen [Guideline for the diagnosis and treatment of autism spectrum disorders in children and adolescents], Nederlandse Vereniging voor Psychiatrie, Utrecht, https://www.vkjp.nl/media/files/Richtlijn%20autisme%20bij%20kinderen%20en%20jeugdigen%20(NVvP%2C%202009).pdf (accessed on 24 June 2025).
[44] Nymbach, A. (2023), Autismusfälle bei Kindern und Jugendlichen gestiegen [Cases of autism among children and adolescents have increased], Verband der Ersatzkassen, Berlin, pp. 46-47, https://www.vdek.com/magazin/ausgaben/2023-05/hkk-autismusfaelle-kinder-jugendliche.html (accessed on 11 February 2026).
[11] Olson, L., S. Bishop and A. Thurm (2024), “Differential Diagnosis of Autism and Other Neurodevelopmental Disorders”, Pediatric Clinics of North America, Vol. 71/2, pp. 157-177, https://doi.org/10.1016/j.pcl.2023.12.004.
[83] Ponnou, S. et al. (2025), “Prévalence des diagnostics d’autisme via le système national de données de santé. Analyse rétrospective de cohorte sur la période 2010–2022”, Neuropsychiatrie de l’Enfance et de l’Adolescence, Vol. 73/5, pp. 276-288, https://doi.org/10.1016/j.neurenf.2025.01.006.
[103] Public Health Agency of Canada (2024), Framework for Autism in Canada, https://www.canada.ca/en/public-health/services/publications/diseases-conditions/framework-autism-canada.html#a3.1 (accessed on 11 February 2026).
[81] Rai, D. et al. (2012), “Parental Socioeconomic Status and Risk of Offspring Autism Spectrum Disorders in a Swedish Population-Based Study”, Journal of the American Academy of Child & Adolescent Psychiatry, Vol. 51/5, pp. 467-476.e6, https://doi.org/10.1016/j.jaac.2012.02.012.
[95] Roberts, J. and K. Williams (2016), Autism spectrum disorder: Evidence-based/evidence-informed good practice for support provided to preschool children, their families and carers, National Disability Insurance Agency, https://www.ndis.gov.au/media/863/download (accessed on 1 December 2025).
[30] Roman-Urrestarazu, A. et al. (2021), “Association of Race/Ethnicity and Social Disadvantage With Autism Prevalence in 7 Million School Children in England”, JAMA Pediatrics, Vol. 175/6, p. e210054, https://doi.org/10.1001/jamapediatrics.2021.0054.
[85] Roman-Urrestarazu, A. et al. (2022), “Autism incidence and spatial analysis in more than 7 million pupils in English schools: a retrospective, longitudinal, school registry study”, The Lancet Child & Adolescent Health, Vol. 6/12, pp. 857-868, https://doi.org/10.1016/s2352-4642(22)00247-4.
[41] Russell, G. et al. (2021), “Time trends in autism diagnosis over 20 years: a UK population‐based cohort study”, Journal of Child Psychology and Psychiatry, Vol. 63/6, pp. 674-682, https://doi.org/10.1111/jcpp.13505.
[35] Sakakibara, M. (ed.) (2017), “Applying machine learning to identify autistic adults using imitation: An exploratory study”, PLOS ONE, Vol. 12/8, p. e0182652, https://doi.org/10.1371/journal.pone.0182652.
[47] Salari, N. et al. (2022), “The global prevalence of autism spectrum disorder: a comprehensive systematic review and meta-analysis”, Italian Journal of Pediatrics, Vol. 48/1, https://doi.org/10.1186/s13052-022-01310-w.
[64] Sealey, L. et al. (2016), “Environmental factors in the development of autism spectrum disorders”, Environment International, Vol. 88, pp. 288-298, https://doi.org/10.1016/j.envint.2015.12.021.
[50] Shaw, K. et al. (2025), “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, Vol. 74/2, pp. 1-22, https://doi.org/10.15585/mmwr.ss7402a1.
[38] Sheik Abdullah, A. et al. (2025), “Leveraging deep learning for enhanced diagnosis of autism spectrum disorder using resting-state functional magnetic resonance imaging and clinical data”, Results in Engineering, Vol. 25, p. 104444, https://doi.org/10.1016/j.rineng.2025.104444.
[43] Shenouda, J. et al. (2023), “Prevalence and Disparities in the Detection of Autism Without Intellectual Disability”, Pediatrics, Vol. 151/2, https://doi.org/10.1542/peds.2022-056594.
[86] Silverman, S., G. Amit and Y. Sadaka (2026), Timing of Autism Recognition Among Israeli Children: Community and Socioeconomic Factors, Taub Center for Social Policy Studies in Israel and KI Institute, Jerusalem, https://doi.org/10.5281/zenodo.18266391 (accessed on 11 February 2026).
[60] Social- og Boligstyrelsen (2024), Data om borgere [Data on citizens], https://www.social.dk/data/databank (accessed on 13 February 2026).
[54] Social- og Boligstyrelsen (2024), Demografi - børn og unge [Demographics - children and young people], https://www.social.dk/data/databank/data-om-borgere/demografi-boern-og-unge (accessed on 13 February 2026).
[57] Socialstyrelsen (2025), Statistikdatabas för diagnoser [Statistical database for diagnoses], https://sdb.socialstyrelsen.se/if_par/resultat.aspx (accessed on 10 December 2025).
[37] Stanley, J. et al. (2025), “Large language models deconstruct the clinical intuition behind diagnosing autism”, Cell, Vol. 188/8, pp. 2235-2248.e10, https://doi.org/10.1016/j.cell.2025.02.025.
[53] Statistics Canada (2025), Health indicator statistics for children and youth aged 1 to 17 years, parent reported, https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1310094701 (accessed on 10 February 2026).
[61] Statistisches Bundesamt (Destatis) (2025), Diagnoses of hospital inpatients, https://www-genesis.destatis.de/datenbank/online (accessed on 20 May 2025).
[99] Sundhedsstyrelsen (2021), Behandling af autismespektrumforstyrrelser hos børn og unge [National clinical guideline for the treatment of autism spectrum disorders in children and adolescents], https://www.sst.dk/media/fuvlzofd/nkr-for-behandling-af-autismespektrumforstyrrelser-hos-boern-og-unge.pdf (accessed on 14 February 2026).
[48] Talantseva, O. et al. (2023), “The global prevalence of autism spectrum disorder: A three-level meta-analysis”, Frontiers in Psychiatry, Vol. 14, https://doi.org/10.3389/fpsyt.2023.1071181.
[55] Tervise Arengu Instituut (2024), PKH1: New and prevalent outpatient case consultations by psychiatrist by diagnosis, sex and age group, https://statistika.tai.ee/pxweb/en/Andmebaas/Andmebaas__02Haigestumus__05Psyyhikahaired/PKH1.px/ (accessed on 20 May 2025).
[74] Thomas, P. et al. (2011), “The association of autism diagnosis with socioeconomic status”, Autism, Vol. 16/2, pp. 201-213, https://doi.org/10.1177/1362361311413397.
[82] Triche, E. (ed.) (2015), “Socioeconomic Disparities and Prevalence of Autism Spectrum Disorders and Intellectual Disability”, PLOS ONE, Vol. 10/11, p. e0141964, https://doi.org/10.1371/journal.pone.0141964.
[7] Tsai, L. (2014), “Impact of DSM-5 on epidemiology of Autism Spectrum Disorder”, Research in Autism Spectrum Disorders, Vol. 8/11, pp. 1454-1470, https://doi.org/10.1016/j.rasd.2014.07.016.
[27] van ’t Hof, M. et al. (2020), “Age at autism spectrum disorder diagnosis: A systematic review and meta-analysis from 2012 to 2019”, Autism, Vol. 25/4, pp. 862-873, https://doi.org/10.1177/1362361320971107.
[78] Volk, H. et al. (2011), “Residential Proximity to Freeways and Autism in the CHARGE Study”, Environmental Health Perspectives, Vol. 119/6, pp. 873-877, https://doi.org/10.1289/ehp.1002835.
[79] Volk, H. et al. (2014), “Autism Spectrum Disorder: interaction of air pollution with the MET receptor tyrosine kinase gene”, Epidemiology, Vol. 25/1, pp. 44-47, https://doi.org/10.1097/ede.0000000000000030.
[6] Volkmar, F. and J. McPartland (2014), “From Kanner to DSM-5: Autism as an Evolving Diagnostic Concept”, Annual Review of Clinical Psychology, Vol. 10/1, pp. 193-212, https://doi.org/10.1146/annurev-clinpsy-032813-153710.
[9] WHO (2019), International Classification of Diseases, 11th edition (ICD-11), World Health Organization, Geneva, https://icd.who.int/browse/2025-01/mms/en (accessed on 6 February 2025).
[4] WHO (1992), International Classification of Diseases, 10th edition (ICD-10), World Health Organization, Geneva.
[3] WHO (1977), International Classification of Diseases, 9th edition (ICD-9), World Health Organization, Geneva.
[46] WHO/UNICEF (2023), Improving the health and wellbeing of children and adolescents: guidance on scheduled child, WHO and UNICEF, Geneva, https://www.who.int/publications/i/item/9789240085336 (accessed on 11 February 2026).
[29] Wolff, J. and J. Piven (2021), “Predicting Autism in Infancy”, Journal of the American Academy of Child & Adolescent Psychiatry, Vol. 60/8, pp. 958-967, https://doi.org/10.1016/j.jaac.2020.07.910.
[80] Wu, J. et al. (2024), “Association between Autism Spectrum Disorder and Environmental Quality in the United States”, ISPRS International Journal of Geo-Information, Vol. 13/9, p. 308, https://doi.org/10.3390/ijgi13090308.
[66] Wu, S. et al. (2016), “Advanced parental age and autism risk in children: a systematic review and meta‐analysis”, Acta Psychiatrica Scandinavica, Vol. 135/1, pp. 29-41, https://doi.org/10.1111/acps.12666.
[16] Zeidan, J. et al. (2022), “Global prevalence of autism: A systematic review update”, Autism Research, Vol. 15/5, pp. 778-790, https://doi.org/10.1002/aur.2696.
[28] Zwaigenbaum, L. and M. Penner (2018), “Autism spectrum disorder: advances in diagnosis and evaluation”, BMJ, p. k1674, https://doi.org/10.1136/bmj.k1674.