And maybe you are too.

OK, that’s an exaggeration. For both of us, I’m guessing. But a widely publicized study in the American Journal of Psychiatry indicates that the official prevalence rates for autism are much lower than they ought to be. They’re already pretty high–one in 110 kids, one in 70 boys, as the banner at the top of the Autism Speaks website proclaims. But the new numbers are really scary: one in 38, nearly three times higher.

This is a vexing problem. And skeptical as I am of psychiatric diagnostics, I don’t think the increasing rates of autism are due solely to a clever manufacturing and marketing campaign, which is more than I can say about some other diagnoses. For one thing, there’s no drug to drive the diagnostic juggernaut. For another, there isn’t much subtle about what the DSM-IV calls autistic disorder. Once you’ve seen one of these kids, shrieking and rocking, unable to talk, hard-pressed to respond to love, and and their shell-shocked, despairing parents, you’d have to be some kind of raving ideologue, or maybe just a Scientologist, to claim that the psychiatrists made up this one to line their pockets and feed their egos.

I suspect that the increase in autistic disorder is real and due to some kind of environmental hazard that is so diffuse that it’s hard to see, let alone to track. Free estrogens, heavy metals, too much  flushed prozac (excreted and discarded) in the water supply, electromagnetic pollution from cell ophones and wi-fi, bad tv, who knows? But something is triggering this malfunction, or, as we used to say, a stress is turning a diathesis into a really bad disease.

So we can;’t blame this one on psychiatry. Or can we?

If you look at those CDC numbers (and I suggest you do; they’re surprisingly accessible, and the CDC provides simple explanations of their statistical charts and tables), you discover that they were derived by examining the records of kids already identified by doctors or teachers as having developmental problems. Most of them were already diagnosed with a developmental disorder; the CDC reviewed the notes, confirmed the autism diagnoses and added in the kids who had the symptoms but not the diagnoses and concluded that one in 110 of them qualified for a diagnosis along the autism spectrum.

Now, here’s the problem: the “autism spectrum.” At some point in the post-DSM-IV era, the decision was made to create an autism spectrum disorder (ASD) out of diagnoses ranging from Asperger Disorder to Autistic Disorder and taking in some other, mostly rare, DSM-IV diagnoses along the way, and to call it the autism spectrum. This move, which will be made official in the DSM-5, has the effect of declaring all these problems as variations on a theme, with the strong implication that they share some biochemical basis and differ primarily in severity and presentation rather than in essence. Scientifically, politically, and philosophically, this is a rat’s nest, which I won’t go into here, except to say that by creating the spectrum, psychiatrists (and neurologists) are claiming much more than they have proved. Which means that they set the agenda with what they think (and maybe hope) is the case, and subsequent research that uses the spectrum can only confirm that it exists.

Among the implications of the spectrum, however, is the fact that epidemiological research using it, like the new AJP study, will count kids with Asperger Syndrome as “autistic.” There is no doubt that this will increase prevalence rates, but it is still an open question by how much. Previous studies, like the CDC’s, just don’t provide an answer.

But the new research is clear on this question. Rather than looking only at kids already identified, this study looked at 36,800 students in a single Korean city in a single school year. Three hundred of the kids were already identified as special needs, and thus as having a high probability for a diagnosis; the rest were just your average kids on the street minding their own business and un-sick enough not to have come to the attention of the authorities. But after the numbers were crunched, the researchers concluded that many of them were evidently harboring a serious mental illness without knowing it: 2.64 percent of the school age population of Goyang City, Korea, is “on the spectrum,” or as Autism Speak(a cosponsor of the study) is sure to proclaim soon, one in 38 kids is autistic.

Now that is a stunning number, especially when you consider that only one in 133 (three-quarters of a percentage point) of the population had already been identified as autistic. In other words, if you go out looking for trouble, you’ll find it in one in 38 kids. But if you wait for trouble to come to you, only one in 133 will turn out to have the kind of trouble you’re interested in. Now that could be because of reluctance among parents (the researchers note among “cultural considerations” the fact that “autism impugns the child’s lineage on both sides and threatens the marriage prospects of unaffected relatives”) or ignorance or something. But it’s also possible (I would say likely) that autistic disorder (as opposed to the other disorders on the spectrum), the kind of suffering that is bound to come to notice, is relatively rare.

Some other numbers bear this out. The overall rate of those “other ASDs” was roughly twice that of the autistic disorders. One in 106 of the overall population had autistic disorder; one in 58 had another ASD, and presumably many, if not most, of these were Asperger cases. To put it another way, out of any thousand of these children, 9 had autistic disorder and seventeen had another ASD. Now this could simply reflect the fact that Asperger’s is more common than autistic disorder, which is undoubtedly true. But that’s the point. Maybe it isn’t such a good idea to fold a muchlarger category into a smaller one, and then to use the name of the smaller (and more severe) disorder for both of them.

Not that it’s a good thing that one out of every 58 kids qualifies for an Asperger diagnosis and one in 106 has autistic disorder. But it’s at least a little less frightening than the idea that one in 38 is autistic, especially when the commonly held meaning of “autism,” accurate or not, is the head-banging, shit-throwing, totally regressed nightmare child, while the head of Facebook is widely thought of as an “Aspie.” And this is exactly the kind of problem that jiggering diagnoses can lead to: increased prevalence that can’t be easily dimissed as a statistical artifact, that indeed seems plausible until you engage in some deconstruction of the terms of the research. A study like these needs a hermeneutician as much as a statistician.

And one other thing: What are we to make of the fact that the prevalence of autistic disorder in the Korean  study–one in 106–is so close to the CDC’s prevalence figure for autism (1 in 110)? Could it be that the CDC numbers, by relying on cases thast have already come to the attention of doctors or teachers, inadvertently filter out most of those “other ASD” cases? Is it possible that what the Korean study really shows is what happens when a diagnosis is actually applied to a population, and that the answer says more about the diagnosis than the population?