You are out for a drive, when a policeman waves you over. He looks at your dashboard, and declares "I see that you have a full tank of gas. I'm giving you a ticket for speeding!" Sounds ridiculous, doesn't it? Yet, lots of otherwise sensible people make a similar error when they equate the increase in the percent of children with an ASD diagnosis (prevalence) with an increase the rate at which new cases of ASD are occurring (incidence). You have both a gas gauge and a speedometer on your dashboard, but you cannot equate one with the other; the same is true for prevalence and incidence of ASD.
Prevalence is a percent: either the percent of your tank that contains gas, or the percent of children with ASD. The prevalence of an ASD diagnosis among children has certainly risen in the past decade (although even the claim of an "explosion" of persons with ASD - as opposed to persons with a diagnosis of ASD - turns out to be bogus - we'll get to that a few posts down the road).
Incidence is a rate: either your speed in miles per hour, or, in the case of ASD, the rate at which new cases of ASD are occurring. An epidemic is defined as a sudden increase in incidence. There is simply no evidence to support the claim that there has been a change in incidence. And until someone figures out how to stand in the delivery room and count autistic-to-be babies as they are born, we are not likely to have such evidence.
Why should you care? Because most of the fears about the allegedly deleterious effects of mercury, gluten, yeast, immunizations, etc., and many of the claims for dubious "therapies" such as chelation, hyperbaric oxygen, and dietary supplements, are based on the unfounded assertion that we are in the midst of an autism epidemic. Take away that touchstone, and a lot of those fears and claims collapse.
So, why has the prevalence of ASD apparently skyrocketed in the past 20 years? Let me count the ways......
First, prevalence depends how we define "a case." Suppose we want to know the prevalence of tall stature. If we define "tall" as being over 7 feet, the prevalence of tall stature among adults may be 1 in 1000. If we lower the cutoff for "tallness" to 6 feet 6 inches, perhaps 1 in 100 will meet the definition; a tenfold increase in prevalence, but no one is an inch taller than before, and tall babies aren't being born at a faster rate than before. All we did was change the cutoff. The same thing has happened in ASD.
The Bible of diagnosis for ASD is the Diagnostic and Statistical Manual of the American Psychiatric Association - the "DSM." In the figure, we see our old friend, the XY graph of IQ versus atypicality. Only, this time I've mapped the children who meet DSM criteria for an autistic spectrum disorder.
"Infantile Autism" makes its appearance for the first time in the DSM-III (published in 1980), as a severe disorder with continuous, bizarre behavior. "Autism, Residual State" also appears in the DSM-III, and is limited to persons who once met criteria for Infantile Autism, but no longer do. In 1987 the DSM-III-Revised comes along. Autism, Residual State is replaced by Pervasive Developmental Disorder, Not Otherwise Specified. In contrast to "Autism, Residual State," PDD-NOS includes children with milder forms of atypicality, who never passed through the stage of severe, bizarre behavior of fully-expressed autism. In 1994 the DSM-IV comes along. This time, the diagnosis of autism itself is broadened, to include a menu of items, some quite mild. Asperger Syndrome also makes its appearance for the first time in the DSM-IV. Each edition of the DSM has cast a broader net, pulling in more children.
As we speak, the DSM-V is under development. Will the DSM-V recognize Nonverbal Learning Disability (NLD) and the Broad Autism Phenotype (BAP) - which we discussed in our previous blog post - as part of the autistic spectrum? If so, we can expect another big jump the prevalence of ASD - with no change in incidence.
Next time, we'll talk about several other factors that have also worked in concert to bring about the spike in prevalence of ASD. (Hint: None of these factors have anything to do with incidence). And see my book for citations to the scientific literature on all that we've discussed here.