Improve the health of the people of Wisconsin by supporting and strengthening physicians' ability to practice high-quality patient care in a changing environment.

The Autism ‘Epidemic’ in perspective

By Darold Treffert, MD

Headlines in the newspapers and on Internet postings on Nov. 13, 2015 read: “Autism cases double.” American Psychiatric Association headline reads: “Survey reveals big jump in the number of U.S. kids with Autism Spectrum Disorder.” Other reports similarly point to what they call an “autism epidemic.”

There is an actual rise in the prevalence of autism. But it is not at an epidemic level.

The beginning of wisdom is to call things by their right name. And in medicine that means “the first step in treatment is to make the correct diagnosis.” With respect to Autistic Spectrum Disorder (ASD), early identification produces better outcomes; the earlier the better. But that makes accurate diagnosis in early childhood all the more important to make appropriate intervention and education decisions.

My introduction to autism was 55 years ago, so long ago that I actually learned from Dr. Leo Kanner himself when he was a visiting professor at UW Medical School in Madison. Dr. Kanner was the first to describe “early infantile autism” in 1940. My first assignment after residency was to start a children’s unit for autistic children at Winnebago Mental Health Institute in 1962. That’s where I met my first savant. In 1970 I carried out an epidemiologic study of infantile autism in Wisconsin that is still cited today in similar studies.

So I am no newcomer to autism.

But articles on Nov. 13, 2015 highlight a just-released report from the National Health Interview Survey (NHIS) which says autism disorders rose from 1.25 percent of U.S. children between ages of 3 and 17 (one in 80) in 2013 to 2.24 percent (one in 45) in 2014. Now if that were an actual increase (doubling) in one year that would be an “epidemic.” But let’s look more closely at those how those figures were derived.

First, they depend on parent reports rather than any in-person, professional examination or even review of medical records. Parents were asked if any doctor or health professional had ever told them that their child had ASD (autism), ID (intellectual disability) or any other DD (developmental disability). That’s it. That’s the diagnostic criteria for the study. Not that parents are unreliable reporters, it is rather the reliability of the reports they are reporting based basically on memory.

Second, the rules for reporting changed between 2013 and 2014. The question regarding ASD was revised (expanded) and put in different order from other DD conditions than previous questionnaires. A Los Angeles Times story points out the doubling “reflects changes in surveying techniques designed to prod parents’ memories for past diagnoses their children might have received” and it now “appears parents who in the past would have reported an intellectual disability or other neurocognitive disorder in a child are now more likely to report autism spectrum disorder.”

Unfortunately, the newspaper headlines and evening news reports on the “big jump in the number of kids with autism” will not reflect the important expansion of definitions of “autism” embodied in the reports between 2013 and 2014. And similar oversights in methodology and ever-changing and expanding definitions of “autism” plague all the more recent reports on the prevalence of autism in the United States.

So when it comes to an “epidemic” of autism, buyer beware. It may be false advertising or sloppy reporting. It certainly is confusing to the public.

However, having said that, and having observed “autism” for these many years I am convinced there is an actual increase in the disorder, but not at an epidemic level. Part of the increase is due to ever-expanding diagnostic criteria for what started out as early infantile autism, narrowly defined. Now early infantile autism has lost its specificity. Part of the increase is earlier diagnosis which is commendable. And in my view, part of the increase is actually due to some environmental factors (pollutants of various sorts that may contribute as well to a rise in other congenital abnormalities and premature births).

Autistic Spectrum Disorder is serious disorder and needs to be taken seriously. But if we are to know its true prevalence, measure the true size of any increase, implement effective treatment and educational interventions and, ultimately, design preventive strategies for autism, we will need standardized, stringent diagnostic criteria that are not ever expanding.

And that brings me back to calling things by their right names. I know some parents and some professionals don’t like diagnostic “labels” for children and are confused by them. Each child does need to be approached individually to be sure. But if we are to have the most accurate prevalence figures and best intervention and prevention tools, we need to be more careful with proper definitions and classifications of this complex condition. That cannot be done by telephone or questionnaire. It will require in-person examinations and review of medical records of many, many children and adults by skilled multidisciplinary teams using standardized tools and criteria. Then such reliable data can be analyzed in research for etiology, treatment and prevention. The NHIS survey, in its limitations section, acknowledges that when it indicates “parents’ reports were not substantiated through clinical evaluation or educational records.”

The Centers for Disease Control and Prevention (CDC) has monitored autism prevalence for 20 years using three methods: National Health Interview Survey (NHIS) based on in-person interview with parents, National Survey of Children’s Health (NSCH) based on telephone interviews with parents, and Autism and Developmental Disabilities Monitoring Network (ADDM), which follows 8-year-old children in 14 communities in which clinicians do review medical and education records but in which there are no in-person examinations. Within its own three methods, prevalence for Autism Spectrum Disorder ranges from 14.7 cases per 1000 children in the 8-year-old group to 22.4 cases per 1000 children in the 3- to 17-year age group.

So if the CDC cannot get consistent figures within its own three studies, how can those figures be seen as reliable overall? In fairness, each of the three studies did show an increase in prevalence over time, but that figure also is an inconsistent one.

In short, there does seem to be an increase in prevalence of autism over time. But it overstates the case to call it an epidemic. At some point, hopefully the CDC or other public or private organizations or agencies will carry out standardized, in-person, multidisciplinary evaluations of individual children and adults to measure true prevalence of autistic spectrum disorder, to trace any change in prevalence and, even more importantly, to use such uniform categories of patients to explore the causes, treatment and prevention of this serious disorder.