In part 1 of this series I examined the desperate attempt of National Health Service in the UK to locate a hidden horde of adults replicating the numbers in the child population in its survey of adult psychiatric morbidity in England 2007-9. In the end using non-standard and diluted psychiatric testing tools they managed to find only 19 cases, less than a third of the one per cent figure they were looking for and postulated that they had mislaid most of the cases (which notwithstanding were held to be there and accounted for). This is only to give a brief summary of this extraordinarily ramshackle project which was carried out to be incorporated in National Statistics (which in turn became a reference point for US National Institute of Mental Health director Thomas Insel). (Part 2 is here.)
Where the autism problem was beginning to show was in the childhood surveys conducted for National Statistics (there seem to be two such documents, one for 1999 and one for 2004 but not subsequently). In the survey conducted in 1999 between January and May from approx. 10,500 interviews they discovered a rate of PDD of 0.2% (i.e. 1 in 500) among 11-15 year-olds and 0.4% (i.e. 1 in 250) among 5-10 year-olds (Table 4.1). It should be noted that with these children, who would have been monitored by the same services, you would expect if the rate was stable for the older group to have a higher rate of diagnosis (the older they were the more likely they would be to get one) but in fact it was lower.
The same pattern is evident in the survey carried out between March and June 2004 from approx. 8,000 children. In the older group now 11-16 year-olds and classified as ASD the rate is 0.8% (1 in 125), also Table 4.1 in this document. This would be roughly the same birth cohort as the younger group in the earlier survey and the fact that rate has now doubled may be partly explained by the fact that they are older and will therefore be more likely to have a diagnosis, but it also has to be allowed that the study group is different and that may be a variable. But again the younger group of 5 to 10 year-olds have a higher rate of 1% despite not having had as long to get a diagnosis (notable also that remarkably few girls in this cohort have a diagnosis). In both cases autism disorder whether PDD or ASD is listed as “a less common disorder”. However, we note a five-fold rise between those born between 1983-88 and those born between 1994-1999 despite the fact that former group had had longer to get a diagnosis - so the position was undoubtedly worse than immediately visible.
At this point we may turn to Baron-Cohen’s long suppressed study of 5-9 year-olds in Cambridgeshire studied in 2003-4, a very similar group perhaps to the younger cohort in 2004 study. This study, which provoked a vicious and unprincipled establishment backlash when a story was published about it in 2007 in The Observer (just before the corrupt General Medical Council proceedings against Wakefield and colleagues), showed that the figure of 1% in the National Statistic cohort was more or less replicated but there would be a further two cases for every three already identified. In 2007 The Observer reported a figure of 1 in 58: when the study was finally published two years later a figure was given of 157 in 10,000 or 1 in 64, still a lot higher than National Statistics figure for children (and suggesting a nearly eight-fold rise from the older group in the first set of National Statistics). Here it is from the abstract:
The prevalence estimates generated from the SEN register and diagnosis survey were 94 per 10 000 and 99 per 10 000 respectively. A total of 11 children received a research diagnosis of an autism-spectrum condition following screening and assessment. The ratio of known:unknown cases is about 3:2 (following statistical weighting procedures). Taken together, we estimate the prevalence to be 157 per 10 000, including previously undiagnosed cases.
It follows that if the authors of the Adult Psychiatric Morbidity Survey of 2007 were looking to replicate the 1% figure of 2004 children’s survey, not only had that figure moved on, but they had little expectation of matching it by any conventional means. The older children from 1999 children’s study, now adults, had only registered an incidence of 1 in 500 and being older they would be relatively well diagnosed. In the end even pioneering new tests which might have a propensity to throw up more diagnoses by relaxing the criteria did not get them very far, and they had to postulate that they had missed most of the cases (while insisting that they were really there). It is also troubling that there was no follow up to the 2004 children’s mental health survey: given what that might have revealed about a former “less common disorder” – no survey for 2009, no survey for 2014, or any other year.
There are two striking features to this story. The first is that until we come to the Adult Psychiatric Morbidity Survey of 2007 and its autism off-shoot of 2009 the data seems to be relatively consistent, though frightening. Secondly, the point that young people are more likely to acquire a diagnosis as they get older is a firm underlying theme of this data. But also if we look at the children surveys of 1999 and 2004 we must consider that the older children were being monitored by the same services as the younger and there was no particular reason why cases would have been missed - and they get more difficult to miss as they get older. There was no identifiable discontinuity, just a strong upward trend: in the 1983-88 birth cohort it is 0.2%, in 1989-94 cohort it is 0.7/8% and in the post 1994 cohort about 1.6%. When they saw this they stopped counting (there was a major official tantrum over media coverage for the Cambridgeshire survey, which had almost given the game away) and the phantom adult horde was moved into position: the adult prevalence was announced as 1%, and the child prevalence was frozen at 1% (2004 figures) before it inevitably went up again as the unpublished Baron-Cohen's paper had predicted it would (and just as the numbers had gone up for the 1989-94 cohort between the 1999 and 2004 publications).
Our officials, protecting their own backs and their friends in the pharmaceutical industry, have a lot to answer for. They could toss all the blame on to Andrew Wakefield but no one knew better what was really happening to our children than they did. In 2007 Baron-Cohen denied the story, The Observer was forced to retract the report, and the editor lost his job. In 2009 the paper - substantially corroborating what had been reported by The Observer in 2007 - was published amid zero mainstream publicity.
PS These were some comments I made to a friend about this off site:
In the mid 90s our local authority was forced to begin developing a new kind of provision, for a largely new group of children - it was a practical issue. While they were closing down special provision for children with various difficulties and putting them into mainstream they were by no means similar to the new group. Even the old severe learning difficulty schools were not geared to dealing with the new intake.
From the late 90s the local authority was coming under increasing political/ideological pressure from central government to keep a lid on special educational needs. They were under pressure from the other direction because the cases were so badly in need of support. It was certainly not because they were recognising odd little quirks which had been passed over in ealier generations.
John Stone is UK Editor for Age of Autism.