In the ever-more-visible field of autism epidemiology, a new study grabbed headlines today. It suggests that the prevalence of the poorly-understood mental and social disorder is nearly triple the most recent, 0.9 percent, estimate from the US Centers for Disease Control and Prevention (CDC).

The study, led by Young Shin Kim of the Child Study Center at the Yale University School of Medicine with senior author, Roy Richard Grinker (pictured), an anthropologist at George Washington University, among others, appears online today in the American Journal of Psychiatry. It estimates the prevalence of autism spectrum disorder (ASD) at 2.6%, or 1 in 38 children, in a population of roughly 55,000 schoolchildren, aged seven to twelve, in the Ilsan district of Goyang City, South Korea. It is the first comprehensive autism prevalence study to survey an entire population, rather than to rely on case reports, as do studies by the CDC and other epidemiological groups; the case report approach misses undiagnosed cases which are not in school or health facility records.

Autism Speaks, an advocacy group that is the lead funder of the six-year study, said in a press release today that the study “does not suggest that Koreans have more autism than any other population in the world. What it does suggest is that autism is more common than previously thought and that, if researchers look carefully, especially in previously understudied, non-clinical populations, they may find more children with ASD.”

Arthur Caplan, who runs a program on mental health ethics as the director of the Center for Bioethics at the University of Pennsylvania, called the findings “significant” and “important’, and added that the new number might help research advocates’ cause on Capitol Hill.

The study also belies claims that a rise in autism is linked to vaccination programs, Caplan says. "You wouldn’t see a sudden explosion in cases unless there was a sudden explosion in vaccinations. And there was nothing like that in South Korea. It puts another dagger in the heart of ‘vaccines are the cause’.”

Today, Nature caught up briefly with Grinker as he rode to a television interview. Our question and answer exchange with him appears below:

What is your take home message from this study?

The take home message is that autism is more common than previously thought and that if we look hard enough we will find many cases of children and adults who are undiagnosed and untreated. We only looked at kids. The scientific take home message is that method makes all the difference. The CDC look at records. They are not looking at actual human beings. So they are only counting kids who have a record, or a diagnosis….What we did is we looked at an entire population, meaning the mainstream schools. Not the kids everyone expects would have autism. And this is true for every medical and psychiatric condition, if you look in an entire population.

Let’s say you were to pick something like a speech impediment. Yes you could count the cases in clinic records or school records. But wouldn’t you imagine there are people who have diagnosable speech impediments for which there is no record….When you use a total population study like we did, you will always find a higher prevalence.

So were we surprised to find a higher prevalence? No. Were we surprised to find prevalence as high as we did? Yes. We didn’t expect it to be as high as it was. So we’re really still thinking about how it is possible that so many kids are in these classrooms, some getting by okay, some not so well, undiagnosed and untreated. Two-thirds of the kids that were eventually diagnosed with autism in our study are untreated, undiagnosed and in the mainstream schools.

Are you getting pushback? Are people saying the definition of autism is so broad that now it’s catching more people than it should?

We didn’t broaden the definition. We are using the definition that is considered to be the gold standard right now. We are using what the scientific community believes to be that. And we are confident about the validity. I’m an anthropologist. I didn’t do the diagnosis. But the child psychiatrists are confident about the validity of the diagnosis. Because we didn’t just apply some loose set of criteria. Each child who was diagnosed went through a very comprehensive, multi-hour diagnostic confirmation.

The pushback is that this is an overestimate, that somehow it’s too high. And it may be an overestimate. Our study has all the imprecision of a good epidemiologic study. (The last comment meant humorously.) Every epidemiological study involves estimation and statistical weighting. And it’s complicated. And you need other studies to replicate it. What we’re hoping we’ve done is to establish standards for what constitutes thorough epidemiological investigation. It’s not quick.

How long did it take?

Almost six years. And the CDC will tell you they can’t do this, because Americans want statistics right away and they need to do ongoing surveillance. They can’t do this type of very time consuming, lengthy study. But what the CDC do, they do extremely well. Our study is not a criticism of the CDC study.

It seems that by definition, if you were largely in schools that are not for special needs or intellectually impaired kids, that you must have been discovering milder cases on the autism spectrum. Wouldn’t it be hard for a profoundly affected child to pass in a mainstream school?

In the US we are so sensitized to picking up special needs and providing services. But not every country in the world does that. Depending on the state, 10-15% of American kids are getting some special education services. That number is less than 1% in South Korea. So of course you are going to find those kids in mainstream school environments. Sixteen percent of the kids that were in the mainstream schools that we diagnosed had some degree of mental retardation. Also there were certainly children that I saw in schools that had significant impairments. But South Korea has a pretty strong mandate for inclusion, legally. They have laws in place for inclusion. Unfortunately that inclusion does not come along with a lot of services. Some kids can get by and adapt to the situation.

I remember a third grade girl [in a different city] Pundang, that I asked a teacher about. And the teacher said: “But she’s not a problem at all.” And this was a girl who clearly had a lot of issues [probably some chromosomal anomaly if not autism]. But the teacher doesn’t necessarily flag a child who doesn’t misbehave. So in these highly structured, often very large classrooms, sometimes 40 kids in a class, a quiet, calm kid who really has a lot of issues and really needs help still may end up not getting flagged.