How many people have autism?" is one of the most frequent questions asked and unfortunately it is also one of the most difficult to answer. There is no central register of everyone who has autism- which means that any information about the possible number of people with autism in the community must be based on epidemiological surveys (i.e. studies of distinct and identifiable populations).

It is more than 50 years since Leo Kanner first described his classic autistic syndrome. Since then, the results of research and clinical work have led to the broadening of the concept of autistic disorders. In consequence, estimates of prevalence have increased considerably. This process has occurred in stages, the start of each of which can be linked to particular studies. The history is summarised and the most up to date figures are given below and in Tables 1 and 2 overleaf.

1943

The specific pattern of abnormal behaviour first described by Leo Kanner is also known as ‘early infantile autism’. Kanner made no estimate of the possible numbers of people with this condition but he thought that it was rare (Kanner 1943).

Over twenty years later, Victor Lotter published the first results of an epidemiological study of children with the behaviour pattern described by Kanner in the former county of Middlesex, which gave an overall prevalence rate of 4.5 per 10,000 children (Lotter 1966).

The Triad of Impairments in Children with Learning Disabilities

In 1979 Lorna Wing and Judith Gould examined the prevalence of autism, as defined by Leo Kanner, among children known to have special needs in the former London Borough of Camberwell. They found a prevalence in those with IQ under 70 of nearly 5 per 10,000 for this syndrome, closely similar to the rate found by Lotter. However, as well as looking at children with Kanner autism, Wing & Gould also identified a larger group of children (about 15 per 10,000) who had impairments of social interaction, communication and imagination (which they referred to as the ‘triad’ of impairments), together with a repetitive stereotyped pattern of activities. Although these children did not fit into the full picture of early childhood autism (or typical autism) as described by Kanner they were identified as being within the broader ‘autistic spectrum’. Thus, the total prevalence rate for the spectrum in all children with special needs in the Camberwell study was found to be approximately 20 in every 10,000 children (Wing & Gould 1979). Gillberg et al (1986) in Gothenburg, Sweden, found very similar rates in children with learning disabilities.

There have been a number of other epidemiological studies in different countries examining the prevalence of autism (but not the whole spectrum). These results range from 3.3 to 16.0 per 10,000, possibly due to differences in definitions or case-finding methods (Wing, 1993).

The studies described above identified autistic disorders in children the great majority of whom had learning disabilities and special educational needs. However, in 1944, Hans Asperger in Vienna had published an account of children with many similarities to Kanner autism but who had abilities, including grammatical language, in the average or superior range. There are continuing arguments concerning the exact relationship between Asperger and Kanner syndromes but it is beyond dispute that they have in common the triad of impairments of social interaction, communication and imagination and a narrow, repetitive pattern of activities (Wing, 1981; 1991).

In 1993, Stephan Ehlers and Christopher Gillberg published the results of a further study carried out in Gothenburg in which they examined children in mainstream schools in order to find the prevalence of Asperger syndrome and other autistic spectrum disorders in children with IQ of 70 or above. From the numbers of children they identified they calculated a rate of 36 per 10,000 for those who definitely had Asperger syndrome and another 35 per 10,000 for those with social impairments. Some of the latter may have fitted Asperger description if more information had been available, but they certainly had disorders within the autistic spectrum. The children who were identified were known by their teachers to be having social and/or educational problems but the nature of their difficulties had not been recognised prior to the study.

For over 30 years, Sula Wolff, in Edinburgh, has studied children of average or high ability who are impaired in their social interaction but who do not have the full picture of the triad of impairments. In her book giving results of her studies (Wolff, 1995), she emphasises that the clinical picture overlaps with Asperger syndrome to a large extent. However, these children represent the most subtle and most able end of the autistic spectrum. The majority become independent as adults, many marry and some display exceptional gifts, though retaining the unusual quality of their social interactions.

Why include them in the autistic spectrum? As Sula Wolff points out, they often have a difficult time at school and they need recognition, understanding and acceptance from their parents and teachers. The approach that suits them best is the same as that which is recommended for children with Asperger syndrome and high functioning autism.

In her discussion of prevalence, Sula Wolff quotes Ehlers and Gillberg’s study. She considers that their total figure of 71 per 10,000, includes the children she describes.

Table 1

It is possible that there are real differences in prevalence of autistic spectrum disorders in different parts of the world, even in different parts of the same country, and at different times. An epidemic of encephalitis, for example, could increase the number of affected children. However, it is very likely that some, even most, of the variation is due to differences of definitions and the difficulty of defining the borderlines of sub-groups within the whole autistic spectrum (Wing, 1996). There are no sharp boundaries separating ‘typical’ autism from other autistic disorders, including Asperger syndrome. [For a complete overview of the problems of establishing prevalence rates and the difficulties of diagnosing and defining autism, see Wing 1 993].

The best estimates of the total prevalence of autistic spectrum disorders are those based on the Camberwell and Gothenburg studies, because these focused on the whole spectrum and not just specific sub-groups:

Table 2

Ref: National Autistic society