Changing characteristicsDr. Eric Fombonne has said that comparing current autism prevalence to a prevalence from studies 30 years ago is like comparing oranges to sheep. Another way to express this is that the diagnoses are not equivalent. But even with (presumably) the same subjective diagnostic criteria it is possible for diagnosed groups of autistics to be inequivalent as time goes by. This hypothesis (let's call it the expanding criteria hypothesis) is verified by looking at client characteristics in the California DDS data. See Table 1.
|Quarter||Epilepsy||Profound MR||Severe Behaviors||Lack of MR|
Every quarter, as caseload rises above what would be expected from changes in the population of the state of California, the proportion of certain characteristics of CDDS autistic clients simultaneously drop. As a group, autistics in California today are quite inequivalent to autistics in California in 1992. Effectively, autism one quarter in California is not equivalent to autism the next quarter. Therefore, it is not possible to say that a real increase in autism prevalence has occurred.
The following figure [courtesy of CDDS] also illustrates the point. Notice that the autism curve and the autism without MR curve run almost parallel to one another.
Evidence from regional differencesThe bulk of the "epidemic" in the state of California occurred in the Los Angeles area, as illustrated by the following figure [courtesy of CDDS]:
Current differences in administrative prevalence between regional centers are substantial. Table 2 shows caseloads of various CDDS categories in the Westside RC and the Central Valley RC, expressed as ratios to the epilepsy caseload.
|Regional Center||Autism Ratio||PMR+SMR Ratio||PMR Ratio||Epilepsy Ratio||Severe Behavior Ratio|
It is telling that despite the 500% difference in administrative prevalence of autism between Westside and Central Valley, there is apparently no difference in the prevalence of mental retardation between these two regional centers. If regional prevalence differences are not real, we can extrapolate and conclude that state-wide prevalence changes over time could also not be real.
The hypothesis that regional prevalence differences may be due to environmental factors, such as pollution, appears to be improbable, as I recently argued.
Caseload growth patterns are telling in that regard. Note that the highest annual caseload growth (in percentage terms) in California occurred in the 2002-2003 timeframe. It was about 20%. Currently, the Central Valley regional center, which has the lowest prevalence of autism in the state, has an annual caseload growth of 24%. Autism in Central Valley is currently undergoing what might be described as staggering growth. In contrast, the Westside regional center, with the highest prevalence in the state, has an annual caseload growth of just over 8%. It is not far fetched to suppose that, eventually, Central Valley will catch up to Westside, and that it was simply behind in its recognition of autism.
Evidence of diagnostic substitutionShattuck (2006) found that as the prevalence of autism has risen in the United States, the prevalence of idiopathic mental retardation and learning disability has undergone a corresponding decline.
California was one of a handful of states that does not clearly follow this pattern, according to Shattuck. But as noted, expanding criteria is what apparently drives most of the autism caseload growth in California. We also know that the recognition of autism in the population with mental retardation in California was around 3.5% in 1992, whereas it is about 7% today. This recognition varies considerably from one regional center to the next. Typically, but not always, regional centers with lower autism prevalence have a lower proportion of clients with autism in the mental retardation category. Note that de Bildt et al (2005) states that the most reliable and well-founded estimate for the prevalence of PDD in children and adolescents with mental retardation is the DSM-IV-TR prevalence of 16.7%. So the California recognition of autism in this population still has considerable room for additional growth, and it is not far fetched to conclude that improving recognition (i.e. diagnostic substitution) has occurred in California over time, even though this is not the most significant factor driving caseload growth in that state.
High prevalence in adultsStahlberg et al (2004) found that 30% of consecutively referred adult patients with ADHD had "comorbid" ASD. If we consider that the prevalence of ADHD in adults is estimated at about 4.2% as of 2006, it would appear that the prevalence of ASD in adults can be at least 126 in 10,000. Granted, Stahlberg's ADHD patients could be more "severe" than usual, but his findings are notable considering the limited screening of a specific population.
Nylander & Gillberg (2001) found that 89.5% of adult psychiatric outpatients with "definite ASD" had previously been missed. They had received other psychiatric diagnoses, such as schizophrenia.
Baron-Cohen et al (2001) found that 2% of randomly selected adult controls scored 32 or higher in the AQ test. Further, 7 of 11 interviewed students who were high-scorers met threshold criteria for a DSM-IV diagnosis. It would appear then that the prevalence of ASD in adults can be as high as 127 in 10,000 if screened thoroughly. This does not even consider those autistics who are missed by an AQ score threshold of 32.
Mark Blaxill's "hidden horde" appears to not be hiding very well, because it has been found repeatedly.
Evidence from prevalence studiesMike Stanton has posted a detailed analysis of autism prevalence studies as part of a rebuttal of Richard Lathe's book "Autism, Brain and Environment".
When we compare apples with apples, the huge increases of autism prevalence often claimed to have occurred seem to vanish. Williams et al. (2005) provides a systematic review of prevalence studies. The researchers conclude that 61% of the variation among prevalence studies may be explained by a model that includes diagnostic criteria used, age of children screened, and study location. Note that this model does not even consider awareness or cultural factors.
What went on in the pastLorna Wing was there and explains it well:
One of us (LW) was involved in the planning of the study by Vic Lotter  of 78,000 children aged 8, 9, and 10 years living in the former English county of Middlesex. This was the study in which the 4-5 in 10,000 prevalence rate was first found. I (LW) know what sorts of children were included as classically autistic because I was one of the small group (Neil O'Connor, John Wing, Vic Lotter and myself) who decided on the criteria. In those days we were interested only in really classic Kanner's syndrome and Vic was determined to keep the criteria as narrow as possible. Later, in the Camberwell study described above, Vic was shown case histories of the children Judy Gould and I thought fitted Kanner's descriptions - to our surprise, Vic said we ought to exclude some because they were not classic enough! I think it is fair to say that, when Vic specified narrow criteria, they were NARROW.
When Judy Gould and I started the Camberwell study, we still thought that Kanner's autism could easily be differentiated from other developmental disorders.. By the end of the study our ideas had been turned upside down. We had learnt from direct experience that the psychological dysfunctions underlying autism were manifested in many different ways, far beyond the boundaries of Kanner's syndrome. We developed the hypothesis of an autistic spectrum based on the triad of impairments of social interaction, communication and imagination. Because we concentrated on the children with learning disabilities (IQ under 70) we saw very few with the pattern described by Asperger. We had to wait for the study by Christopher Gillberg in Gothenberg to find out how many children with IQ of 70 and above were also in the autistic spectrum. As described above, combining the results of these two studies gave an overall prevalence rate for the whole autistic spectrum, including those with the most subtle manifestations, of 91 per 10,000 - nearly 1% of the general population.
The epilepsy argumentAutism has been linked to a seizure liability, and the CDDS data itself shows that the prevalence of epilepsy among autistics is considerably higher to that of the general population. It follows that an environmental trigger capable of producing an epidemic of autism might also result in an epidemic of epilepsy. Surprisingly, we find in the data that the epilepsy caseload grows at about the same pace as the population in the state of California. Additionally, the prevalence of epilepsy does not appear to depend on degree of urbanization.
The mental retardation argumentAn epidemic-causing environmental trigger that results in brain injury should be expected to increase the probability that an individual will have mental retardation, and thus result in an epidemic of all levels of mental retardation. But again, there is no evidence of an increase in the prevalence of mental retardation. In California, the mental retardation caseload increases at about the rate that should be expected from population growth.
Furthermore, it is known that the average IQ score has actually risen over time and IQ tests have to be re-normalized periodically (see Flyyn effect). While this does not necessarily say anything about autism, it should at least put to rest fears that an epidemic of neurological disorders will "destroy the United States."
The institutionalization argumentProponents of an autism epidemic generally not only refer to autism, but to an increase in the prevalence of all sorts of neurological disorders. They also often refer to the increased fiscal burden that will presumably result from this epidemic. While increased awareness can result in increased service expenses, I contend that there are certain types of services that are not likely to be significantly affected by increased awareness. Let's look at institutionalization of developmentally disabled individuals in the state of California.
In Q2 1992, the total number of institutionalized individuals registered with CDDS was 32,943. In Q2 2005, the number was 36,869. Adjusting for population growth, we get 10.6 per 10,000 persons institutionalized in 1992 vs. 9.97 per 10,000 in 2005. There appears to be a decrease in the prevalence of institutionalized individuals with developmental disabilities, which is a positive trend.
Granted, very young children would not tend to be institutionalized, but even in the younger age cohorts the CDDS data shows a decreasing trend in number of institutionalized clients. At the very least, this should lay to rest fears that a generation from now institutions will be overflowing with adult autistics.