090821-percentage-550×458.jpg

The autism blogosphere is aflutter over a new survey showing that one percent of kids in the United States have autism. Because that’s markedly higher than the Center for Disease Control and Prevention’s current estimate of 1 in 150, or 0.67 percent, some argue that the new survey bolsters the specious idea of an “epidemic” of autism.

The data in the spotlight come from the National Survey of Children’s Health, a telephone survey primarily funded by the Maternal and Child Health Bureau.

During one period in 2003 and another in 2007, researchers called nearly 200,000 U.S. households with at least one child and asked a variety of questions about the family’s demographics and the child’s physical and mental health.

The 2007 data, released earlier this month, found that about one percent of parent respondents answered ‘yes’ when asked if a health professional has ever told them their child has autism. That was up from 0.5 percent in the 2003 survey.

The numbers seem striking, but I’m leery of the survey’s methodology. First, the question about autism changed slightly — from asking in 2003 whether the child had been labeled with “autism” to asking in 2007 if the child had “autism, Asperger’s Disorder, pervasive developmental disorder, or other autism spectrum disorder.” Casting a wider net of conditions is likely to up the ‘yes’ response rate.

More troubling is that the survey didn’t ask whether the child had received an official autism diagnosis. If a child has a developmental delay, a “medical professional” — which the parent might interpret to mean doctor, nurse, therapist or social worker — could suggest autism without giving the child a rigorous diagnostic test.

Finally, even if the results are sound, the idea of one percent general autism prevalence is not new. Since 1999, in fact, rigorous epidemiological studies from Sweden, England and Canada have estimated the prevalence to be one percent or more. So the new survey needn’t spur worries over a dramatic rise.