04/05/2012 03:42 pm ET Updated Jun 05, 2012

The New Autism Numbers

The Centers for Disease Control isn't usually known for its sense of dramatic timing. But by releasing its new report about autism prevalence just before Autism Awareness Month began, the CDC sparked a renewed discussion about what the nation should do about autism. The report shows that 1 in 88 children has an autism spectrum disorder, a 23 percent increase since the CDC's last report in 2009 and a 78 percent increase since the previous one in 2007.

Those numbers have made for good headlines and for new claims that autism is an epidemic. But it's worth looking beyond the headlines to understand how the CDC arrives at the prevalence numbers and what those numbers do -- and don't -- mean.

The CDC makes its prevalence estimate based on data it collects from the Autism and Developmental Disabilities Monitoring Network. The latest 1 in 88 estimate comes from ADDM Network data related to children who were 8 years old in 2008 and lived in 14 areas of the United States. The 2009 estimate of 1 in 110 was based on 8-year-old children in 2006 in 11 communities, and the 2007 estimate of 1 in 150 was based on 8-year-olds in 2002 in 14 communities.

There are a couple of striking details to notice here. First, although the autism estimate is reported as a national number, it's not. In fact, the 14 communities that make up the ADDM Network included about 8 percent of the U.S. population of 8-year-olds in 2008. And in those individual communities, the prevalence of autism varied widely -- from 1 in 210 in Alabama to 1 in 47 in Utah. We tend to pay a lot more attention to the "1 in 88" than to the word "estimate," but it turns out to be a pretty important word.

Second, while it's true that the prevalence of autism has increased, it's not an increase in the way we usually think of that word. What "increase" means in this case is that the prevalence of autism was greater among 8-year-olds in 2008 than it was among eight year olds in 2006 or in 2002. The increase is not retroactive -- that is, it doesn't necessarily mean that the prevalence of autism in that group of children from 2002 is now 1 in 88 instead of 1 in 150. And it's not predictive -- it doesn't necessarily tell us what the prevalence will be among children who are 8 this year. It is, simply, a snapshot of a particular population at a particular time.

As one person comments in my documentary, comparing two different prevalence studies is sort of like comparing "apples and automobiles." Nevertheless, groups such as Safe Minds reacted to the CDC numbers with alarm. In its document, "The Autism Crisis," Safe Minds asks, "Can anyone with a conscience claim that this isn't an epidemic?"

To understand why my conscience is clear, you just have to think about another controversy that rocked the autism community a couple of months ago. The New York Times broke the news of a study that suggested that a proposed change to the definition of autism would greatly reduce the number of people who could get a diagnosis. The new definition, set to appear in the next edition of the Diagnostic and Statistical Manual of Mental Disorders, collapses the various subcategories of autism into a single condition called Autism Spectrum Disorder.

As with all things in the autism community, there was disagreement about the study's conclusions. Some researchers suggested that the new definition would not cause a significant decrease in diagnoses of autism. But the prevailing reaction among parents of autistic children was panic that their children could lose the autism diagnosis and the services that go with it.

So here is where we are, in the space of only two months: The increase in autism prevalence under the current definition is viewed by most people as bad news, but the possible decrease of autism prevalence under the new definition is also viewed as bad news.

You can't ask for better evidence that autism is socially constructed and fluid in meaning. After all, we don't spend a lot of time debating whether decreases in the prevalence of medical conditions like diabetes or stroke or cancer are good or bad. We understand that decreases of those things are good because we have a much clearer idea of what they really are.

It's not the same with autism. Even with all of the publicity and with another Autism Awareness Month upon us, we're still not sure exactly how to define autism. We all write our own meanings onto autism and react to news about the condition based on those meanings.

Using the word "epidemic" is one way that people do this. It's a political term meant to suggest one possible definition of autism. But given autism's elusive nature, "epidemic" is not a precise scientific term the way it is when we're talking about the flu. If you believe in an autism epidemic, you believe in a certain definition of autism that is not the same as other people's autism.

So I would look skeptically at anyone who tries to tell you definitively what the new autism numbers mean. If the answers were that easy, we wouldn't still be asking the questions.

For more by Todd Drezner, click here.

For more on autism, click here.