Clinicians who treat children with pervasive developmental disorders are likely to be asked by parents to explain the rising prevalence of autism and the possible environmental causes. Let me offer a brief background and a few facts that may help to address these questions and show that autism is not an epidemic.
Let's start with the numbers. Historically, autism was considered a rare disorder with a reported prevalence of 2-5 cases per 10,000. Today, prevalence estimates are 20 per 10,000 for autism and about 60 per 10,000 for pervasive developmental disorders. This is not a trivial difference, but I do not believe that the change represents a rise in new cases--a requirement for an epidemic.
The likely reasons for the increase in autism prevalence are the broadening of the diagnostic rules for the disorder, better sampling of cases in the community, and better diagnostic precision.
Let's look first at classification. In 1994, DSM-IV broadened the diagnostic definition for autism and the potentially related disorders of Asperger's disorder and pervasive disorder not otherwise specified. Any time you broaden the definition of a disorder, prevalence will go up.
Next there's the issue of sampling. In the past, researchers estimated prevalence by counting all the known cases identified by clinicians in the field. In psychiatric epidemiology, this approach always resulted in an underestimate of the prevalence. There always was a gap between identified cases and all cases. This was shown for depression, attention-deficit/hyperactivity disorder, obsessive-compulsive disorder--to name a few. Even with serious disorders like autism, cases went unidentified. Newer studies surveyed the community to include previously unidentified cases in the estimate of prevalence, thus correcting the systematic undercounting of cases.
Finally, there's the issue of diagnostic precision. In the past, the instruments were not very good at differentiating between children with mental retardation alone and those with both autism and mental retardation. Similarly, children in the average IQ range rarely were diagnosed with autism and rarely contributed to the numerator in the expression of prevalence. Today the diagnostic tools are far from perfect, but they are more precise and capable of making the diagnosis in both lower- and higher-functioning children.
Data from California consistently show that lower- and higher-functioning children are contributing to the rise in newly identified cases. …