Evid Based Mental Health 10:28 doi:10.1136/ebmh.10.1.28
  • Prevalence

About 1% of children in the South Thames region have an autistic spectrum disorder

 Q How common are autism and autism-related spectrum disorders in children in the South Thames region of the UK?



Prospective cohort study.


Twelve districts, South Thames, UK; time period of assessment not stated.


56 946 children born between 1 July 1990 and 31 December 1991 in South Thames (mean age at assessment 12 years, range 9.8–14.4 years).


Children with an autistic spectrum disorder (ASD) or with social and communicative impairment were identified from the child health and speech and language therapy services registers. Children with a possible undiagnosed ASD (defined as those having a statement of special educational needs) were identified using the special needs register. All 1770 children identified in this survey received the social communication questionnaire (SCQ). In-depth clinical assessment using the autism diagnostic interview-revised (ADI-R), autism diagnostic observation schedule-generic (ADOS-G) plus IQ, language and adaptive behaviour assessment was carried out on a random sample of 255 children (stratified by previous diagnosis and SCQ score) of screened children. Information from health records and from teachers about social, communication and other behaviour was also used to contribute to the diagnosis. Consensus clinical diagnoses of ASD were made based on all information collected. A more stringent definition of autism was also used, which required a consensus clinical diagnosis of autism plus meeting ADI-R criteria for autism at 4–5 years of age and currently meeting ADOS-G criteria for autism. Results from the in-depth clinical assessment sample were extrapolated to the entire sample of children with a previous diagnosis of ASD or possible ASD. These results were then extrapolated to give population prevalences of autism and autism-related spectrum disorders.


Diagnosis of autism or autism-related spectrum disorder.


In the South Thames region, the estimated prevalence of childhood autism based on clinical consensus diagnosis alone was 39 cases per 10 000 (95% CI 30 to 48 per 10 000). When more stringent diagnostic criteria (clinical consensus plus ADI-R and ADOS-G criteria) were used the prevalence was 25 cases of childhood autism per 10 000 (95% CI 18 to 32 per 10 000). The prevalence of other childhood autism-related spectrum disorders was 77 cases per 10 000 (95% CI 52 to 102 per 10 000).


The prevalence of childhood autistic spectrum disorders is estimated to be about 116 cases per 10 000 in the South Thames region of the UK.


Authors note that although prevalence estimates were calculated for the whole population, the whole population was not screened in this study. Therefore estimates represent minimum figures as some cases of childhood autism spectrum disorders will have been missed.


  1. Simon Baron-Cohen
  1. Autism Research Centre, Department of Psychiatry, Cambridge University, UK

      This study reports a rate of (classic) childhood autism of 4 per 1000 and a rate of other autism spectrum conditions (such as Asperger’s syndrome, pervasive developmental disorder—not otherwise specified, and atypical autism) of 8 per 1000, making a total prevalence estimate of autism spectrum conditions of just over 1% of children. Since the rate of classic autism in 1978 was given as 4 per 10 000,1 this represents a tenfold increase in this subgroup, and a 30-fold increase of autism when considering the whole spectrum, in less than 30 years. This is the highest prevalence reported to date, although it is not wildly different to other recent studies.2

      This study is important in demonstrating the importance of multiple case-ascertainment methods and standardised diagnostic assessment in prevalence studies. It also highlights how narrow or broad definitions of autism can affect prevalence estimates. One practical implication is that it will enable service providers (diagnostic clinics, special education, therapists, among others) to plan provision for children, and across the lifespan into adulthood (sheltered employment, sheltered accommodation, and related social services). The results of this study are likely to be replicated in other countries but such research is very labour intensive and takes years to complete.

      Clinical research will have to explain what has driven up prevalence rates. Much of this may be attributable to better recognition, more services, better training of clinicians, and to broadening the diagnostic criteria (for example, including Asperger’s syndrome). Whether genetic factors (such as more frequent assortative mating of strongly systemising parents3) and/or fetal hormone exposure factors4 play a role in making autism more common remains to be thoroughly tested. Fears that the rise is related to measles-mumps-rubella (MMR) vaccination damage or mercury toxicity now appear unlikely.5



      • For correspondence: Professor Gillian Baird, Newcomen Centre, Guy’s Hospital, London, UK; gillian.baird{at}

      • Sources of funding: Wellcome Trust and Department of Health, UK.