31 verified experts answered this question on independent fact-checking platform Metafact.io. 17 answered 'likely' or higher, this is one of them.

The prevalence of autism – the number of cases diagnosed at any age, has increased hugely.

But prevalence is not the same as incidence - the number of cases born with autism. There is no reason to believe that the incidence of autism has increased.

An increase in the prevalence of cases was inevitable given the recognition that the classic definition of autism was too narrow and there was a whole spectrum of autism (Wing).

The historic factors in the increase in diagnosed cases (prevalence) was the widening of the criteria, to fit them also to adults, and to be able to apply them to cases previously diagnosed merely with learning disability, where this meant less access to special support.

For instance, a study found that the average administrative prevalence of autism among children in the US had increased from 0.6 to 3.1 per 1000 from 1994 to 2003. But this increase was accounted for by a decline in the prevalence of mental retardation and learning disabilities.

This was not the only factor that drove up prevalence figures: Cases with only very mild symptoms previously not clinically diagnosed at all, were now also included in the autism category. It turns out that it is these cases which are the most likely cause of the increase in prevalence since 2000.

So what are the numbers at present? An ongoing study across 14 European countries, estimates the prevalence currently to between 0.6 and 1 percent. In the US this value has varied between 0.6 to 1.69 percent in 2014, according to data from the Autism and Developmental Disabilities Monitoring Network of the US centres for disease control and prevention.

These figures were reported at the International Society for Autism Research 2018.

At this meeting it was also reported that the bulk of the recent increase in the US can be attributed to the increase in diagnosis of children at the mild end of the spectrum.

This conclusion was reached after measuring the extent of social impairments, a defining feature of autism, in 12,700 children using the Vineland Adaptive Behaviour Scales: The number of children with severe limitations remained the same in 2012 as in 2000, while the number of children with mild or no impairments rose.

This conclusion echoes that of a Swedish population study. Here autism prevalence had remained the same for cases with moderate or severe symptoms, but increased for those with milder symptoms, typically diagnosed after preschool age.

Thus, it seems safe to conclude that the observed increase in prevalence – that is, cases diagnosed as autistic, – is a consequence of changes in public awareness and changes in the interpretation of diagnostic criteria, especially as they apply at milder levels.

Is present day diagnostic practice over-inclusive? Possibly. Cases now labelled autistic vary so much from each other, that it is likely that a number of different neuro-cognitive phenotypes are mixed up together. These different phenotypes have yet to be identified.

Uta Frith is an expert on autism spectrum disorder and childhood development from University College London. 

This expert response was published in partnership with independent fact-checking platform Metafact.io. Read all responses to the question here.