Data provides misleading picture of autism

Terry Devitt, University Communications

If statistics compiled by the U.S. Department of Education are to be believed, in 1992 the state of Illinois had only 322 diagnosed cases of autism among school children. In 2003, according to the same statistical source, Illinois had more than 6,000 children diagnosed as autistic.

National special education statistics, which showed a 657 percent increase in autism over the decade from 1993 to 2003, are routinely used to suggest the country is experiencing an epidemic of autism, a developmental disorder of children characterized by impaired social and communication skills as well as repetitive behaviors and obsessive interests.

But inconsistencies in how the condition is diagnosed throughout the nation’s schools, and the fact that the increasing trend for autism coincides with a corresponding slump in the reporting of mental retardation and learning disabilities, challenges the use of special education data to portray a national epidemic of autism, according to a new study published in the current issue of a leading medical journal (April 3, 2006).

Paul Shattuck, a researcher at the University of Wisconsin-Madison’s Waisman Center, writing in Pediatrics, says special education data cannot be used to claim there is an autism epidemic because the figures are “hopelessly confounded” by changing and uneven identification and reporting practices among schools and states.

At issue, says Shattuck, is the practice of “diagnostic substitution,” where educators, over time, have increasingly applied the autism label to children who, in the past, would have been labeled differently.

“My research indicates that the increase in the number of kids with an autism label in special education is strongly associated with a declining usage of the mental retardation and learning disabilities labels in special education during the same period,” Shattuck says. “Many of the children now being counted in the autism category would probably have been counted in the mental retardation or learning disabilities categories if they were being labeled 10 years ago instead of today.”

The point, says Shattuck, is that identification and diagnostic practices change over time and can lead to a misperception that a condition is more prevalent than it has been in the past.

“Each year since 1994, the probability of using the autism label has increased while there has been a corresponding decrease in the likelihood of educators using the mental retardation and learning disabilities categories.”

In contrast to the dominant pattern, California was found to be one of only a handful of states where there was no decrease in the number of children labeled mentally retarded corresponding with an upward trend in identification of autism. This undermines the use of data coming out of California as a representative indicator of what is happening in the rest of the country, as has been suggested in recent press accounts and official reports, Shattuck argues.

In the U.S., federal law mandates the provision of special education in the nation’s schools. Schools are required to provide specialized services to children identified as having special needs and they must classify individuals according to 13 specified categories for the sake of annual enrollment reports. Autism was only added to the list in the early 1990s, and the federal statute’s definition of the condition, amounting to just a few sentences, is vague, Shattuck says.

What’s more, there are no uniform diagnostic practices or guidelines among states or school districts, meaning that how children are diagnosed and sorted in special education settings can vary dramatically from school to school.

“Schools nationwide don’t adhere to any common diagnostic guidelines when they’re sorting kids into these categories,” Shattuck explains. “States and individual schools are left to devise criteria. Everyone is using a different yardstick to measure the same thing.”

Shattuck notes that the diagnostic methods employed by schools in special education settings are distinct from medical and psychological diagnoses of autism, which entail a more precise definition of the condition and uniform diagnostic methodologies.

Catherine Lord, an expert on autism at the University of Michigan, noted that the new study has important implications for understanding the link between the everyday lives of children with autism and the ways the condition is identified and labeled in schools. “Shattuck’s work has both theoretical and practical importance. It highlights the need to consider the immediate implications for children’s lives of the lag between scientific findings regarding the diagnosis and prevalence of autism, and state and school system policies,” Lord says.

Shattuck emphasized that his study does not show, one way or another, whether there is an epidemic of autism. His study, he says, only shows that trends from special education data are insufficient to make such a determination, as is commonly done.

“The upshot is that it is not valid to conclude there is an epidemic of autism by looking to special education data, because the special education data is plagued with this diagnostic substitution phenomenon and the lack of nationally uniform diagnostic procedures.”

Shattuck says his study reveals, among other things, a national inability to accurately measure the scope of developmental disability among Americans.

“In fact, we simply do not have an adequate infrastructure of public health data in this country to say one way or another whether the true prevalence of autism has changed in the general population since the early 1990s. And this study emphasizes why it is so important to continue funding the longitudinal study of prevalence in several states, funded by the Centers for Disease Control, that began recently,” he says.

“It is in the absence of good quality information that people understandably make the mistake of looking to the special education data to draw conclusions. And I think parents and advocates have every right to be angry at our collective inability to answer such consequential questions as, ‘Do more of our children have autism than in the past? If yes, then why?’ ”

Shattuck’s research was supported by a grant from the National Institute of Child Health and Human Development.

© 2005 Board of Regents of the University of Wisconsin System