Evidence against an “autism epidemic”

One of the key arguments by advocates claiming a link between mercury in childhood vaccines is that there is an “epidemic” of autism. They’ll claim that autism was unknown before the 1930’s, when thimerosal was first introduced into vaccines. (Never mind that there are plenty of descriptions of autism-like conditions dating from as far back as the 18th century.) They’ll then claim that there is an “epidemic” that accelerated in the 1990’s, when additional vaccines were added to the recommended childhood schedule, and that it was the additional mercury from those vaccines that was responsible. It has been argued that a lot of that apparent increase was actually due to a widening of the diagnostic criteria for autism in the early 1990’s.

Now there’s more evidence to support that view. Just published in the April issue of the journal Pediatrics is a study by Paul Shattuck at the University of Wisconsin, entitled, The Contribution of Diagnostic Substitution to the Growing Administrative Prevalence of Autism in US Special Education. (Hat tip to the “source” who informed me of this article.)

In the paper, Shattuck analyzes special education figures that are being used to bolster claims of an autism “epidemic” and finds them wanting. In essence, diagnostic substitution can explain nearly all of the apparent increase of autism as recorded by the number of children receiving special education services. What that means is that children that would have been diagnosed with something else 15-20 years ago are now being diagnosed with autism. Dr. Shattuck starts with an example from a different condition, mental retardation as one of his reasons for suspecting diagnostic substitution as a cause of the perceived “epidemic”:

Second, prior research has established a precedent of diagnostic substitution in special education enrollment. From 1976 to 1992 the number of children in the mental retardation (MR) category decreased by 41%, whereas the number in the learning disabilities (LD) category increased 198%. There is considerable evidence that suggests this was because of a growing likelihood that schools would use the LD label for children with mild MR, presumably because a label of LD was increasingly seen as carrying less stigma than MR. Finally, a recent epidemiological study depicted a downward deflection in the incidence trend of other developmental disorders just as the trend for autism made a sharp upturn in the early 1990s, again suggesting the possibility of diagnostic substitution.

I was somewhat surprised to learn that autism was first listed as a primary diagnostic category for special education in the early 1990’s, as a result of being mandated by part of the Individuals with Disabilities Education Act in 1990 as part of Public Law 101-476.28. Before Public Law 101-476, there was no separate autism category. What that meant is that and children with autism enrolled in special education were lumped together for administrative purposes in the legal definition of the “other health impairments” (OHI) service category. Shattuck states: “However, no reliable data exist that would indicate how the enrollment of children with autism was actually distributed among other enrollment categories before the 1990’s.” States were required to begin using the new category in 1993, and 1994 was the first year that all states used it. Using a variety of statistical analyses, Shattuck looked at the rates of autism and a variety of other diagnostic criteria reported as special education recipients over a 20 year time period. He also used a method to correct for the possibility that diagnostic substitution was taking place simultaneously among several diagnostic categories. The findings were:

The average administrative prevalence of autism among children increased from 0.6 to 3.1 per 1000 from 1994 to 2003. By 2003, only 17 states had a special education prevalence of autism that was within the range of recent epidemiological estimates. During the same period, the prevalence of mental retardation and learning disabilities declined by 2.8 and 8.3 per 1000, respectively. Higher autism prevalence was significantly associated with corresponding declines in the prevalence of mental retardation and learning disabilities. The declining prevalence of mental retardation and learning disabilities from 1994 to 2003 represented a significant downward deflection in their preexisting trajectories of prevalence from 1984 to 1993. California was one of a handful of states that did not clearly follow this pattern.

So, in nearly all states, as the reporting of autism cases for special education administrative purposes increased, the reporting of mental retardation and learning disabilities declined correspondingly, suggesting that children who once would have been classified as mentally retarded or learning disabled were now more appropriately being classified as having autism or ASDs. He further comments:

The mean administrative prevalence of autism in US special education among children ages 6 to 11 in 1994 was only 0.6 per 1000, less than one-fifth of the lowest CDC estimate from Atlanta (based on surveillance data from 1996). Therefore, special education counts of children with autism in the early 1990s were dramatic underestimates of population prevalence and really had nowhere to go but up. This finding highlights the inappropriateness of using special education trends to make declarations about an epidemic of autism, as has been common in recent media and advocacy reports.

Indeed. Using such estimates is fraught with peril even using good statistics. If you use bad statistics, as the Geiers do, you can make such numbers seem to show almost anything you want.

Shattuck then issues a stinging rebuke to those who would use these numbers to support the concept of an “epidemic”:

Steep growth in administrative prevalence after introducing a new category is a common pattern that was also seen in the other 2 reporting categories newly introduced in the 1990s (TBI and DD). As with autism, in the first few years these categories were used it was not uncommon for states to
report very few children with a primary diagnosis of TBI or DD. The prevalence for these categories also had nowhere to go but up. Suggestions that special education trends substantiate the existence of an autism epidemic would logically also have to either claim an epidemic of brain injury and DD or explain why the same pattern of growth in these 2 categories does not represent an epidemic as it does for the autism category.

(Emphasis mine; note TBI= traumatic brain injury and DD=developmental delay.)

While pointing out the number of studies looking for a link between mercury and autism that failed to find a link, Shattuck does emphasize that the results of his do not disprove a connection between environmental factors and autism. Rather, they simply show that you can’t legitimately use numbers of children assigned to administrative categories for special education as a reliable estimate of changes in the true prevalence of autism, because you’re comparing apples and oranges. The categories didn’t even include autism before 1993. Consequently, for purposes of special education, autism did not exist as a major diagnostic category, and in the vast majority of states children diagnosed with autism had to be assigned a different category. Of course, autism rates as listed for administrative purposes in special education children skyrocketed after that.

As Dr. Shattuck stated in an interview (hat tip to Autism Diva):

In the case of special education counts for children with autism, the administrative prevalence is simply the number of students with a primary classification of autism divided by the total number of students in that given region, whether it’s a state, district or county.

This means, among other things, that data collected in this manner often underestimate “the true population prevalence because, for instance, schools do not go out into the community and actively seek out and evaluate all kids for autism,” Shattuck said.

For instance, consider data collected in Wisconsin: In 1992, 18 children were counted in special education programs as being autistic. By 2002, that number had jumped to 2,739.

“The conclusion is that the prevalence of autism has grown by 15,117 percent. This is ridiculous,” Shattuck said. “No credible clinician or scientist in the field would ever suggest there were actually only 18 children with autism in all of Wisconsin in 1992.”

The key word phrase here is “no credible clinician or scientist.”

Clinicians have long suspected that the the apparent autism “epidemic” as detected using numbers of of children receiving special education in different categories was a sham. Now they have some hard data to support that view. Of course, this study is not the final word and is not without some problems. A commentary by Craig Newschaffer accompanying the article this issue points out some potential problems with the model that Shattuck used but in the end seemed to come to similar conclusions. The point of difference was that the rate of diagnoses of mental retardation may not be falling as fast as the rate of increase of diagnoses of autism. Dr. Shattuck includes a point-by-point response.

What both end up agreeing on is that epidemiologists may never be able to figure out if there truly has been an increase in the prevalence of autism over the last couple of decades. Near the end of his commentary, Dr. Newschaffer laments:

I also believe that the time has come to accept that, given the behavioral basis of the autism diagnosis, the lack of knowledge about autism’s underlying etiology, and the limitations of retrospective analyses, we are not likely to develop a conclusive body of evidence to either fully support or fully refute the notion that there has been some real increase in autism risk over the past 2 decades.

Dr. Shattuck more or less agrees but points out that it is still important to study the epidemiology and gives reasons:

Second, Newschaffer concludes with a plea to turn from the intractable, and perhaps unanswerable, question of whether autism’s true prevalence has increased over the past 2 decades toward a greater prospective focus on conducting epidemiologic and genetic research related to autism. I agree that we may never be able to fully understand the dynamics behind historical changes in autism prevalence and of course agree that we need more research into etiology. However, autism diagnosis is based on observing behavior rather than clearly identifiable biological markers, diagnostic agreement among clinicians is not perfect, black children tend to be identified later than white children, and clinicians have reported a willingness to give a diagnosis of autism for children with ambiguous symptoms if they know it will result in the child obtaining more services. Therefore, I also believe that we need to continue complementing research into etiology with careful examination of the sociopolitical context wherein processes of identification, referral, diagnosis, and intervention occur for children with autism and their families.

The bottom line is that, even if Newschaffer is correct and there has been an increase in the prevalance of autism, it is almost certainly not of the magnitude suggested by advocates claiming that mercury in childhood vaccines cause autism, who sometimes dismiss suggestions that the apparent increase during the 1990’s was due to improved diagnosis and broadening of the diagnostic classification with ridiculous statements such as, “No child with autism ever went undiagnosed.” The “autism epidemic” (or, as it is sometimes more offensively dubbed, “the autism tsunami“) is nothing of the sort.

How long do you think it will be before the mercury crowd starts attacking Dr. Shattuck as biased or a pharma shill?