As journalists and celebrities alike have been telling us, increasing numbers of American children have been receiving diagnoses of autism or other forms of Pervasive Developmental Disorders. For many readers, this fact has raised the question: What's causing this increase in the occurrence of autism? Before we rush to find the answer, though, let's examine this question under a strong light.
Does an increase in diagnosed cases of autism indicate that there are more autistic children than there were before? This might be the case, but there are other possible explanations of the increase. For example, it may be that parents who are more aware of autism are more likely to bring their children for evaluation. People who never heard of autism might consider autistic behaviors to be a matter of mental retardation, poor speech and hearing, defiance, or the results of ineffective parenting, and might regard these causes as impossible to treat, or curable by popular home interventions like frequent spanking. States with higher adult educational levels are also found to have a higher rate of diagnosed autism, so unless we want to assume that educating parents makes their children autistic (this was thought about mothers at one time), we might want to consider the possibility that increasing parent awareness has led to increased diagnoses, although not necessarily an actual increase in the frequency of autism.
A related issue has to do with realistic concerns about early diagnosis to allow early intervention. As a society, we are committed to the idea that interventions in the first three years have a much better chance of being effective than later treatments do, and there is certainly evidence that this is the case for problems of speech development, or difficulties related to maternal depression or other parenting problems. This belief, which is to some extent supported by evidence, provides a powerful motive for early diagnosis of serious problems like autism. But efforts toward early diagnosis may not always give accurate information.
Certainly, there is evidence that many autistic children are "different" from infancy, and are different in ways that could be observed and diagnosed. For example, such infants may not show "joint attention" by looking back and forth from an interesting object to a parent, inviting the parent to join in looking and enjoying a sight. Rather than signaling a parent that the baby wants help in doing something, he or she may take the adult's hand and put it on the object that needs opening or moving, as if the hand itself were a tool. Interestingly, these are behaviors that can be seen only when the child is with someone else--- they would not be apparent if we just watched the child playing alone.
Under the right circumstances, those behaviors would be pretty obvious, it seems, and they would allow for quite early diagnosis. So, what's the problem? What does early diagnosis have to do with an increasing rate of diagnosis of autism?
Steven Sheinkopf, writing in the "Brown University Child and Adolescent Behavior Letter" ("Early screening and diagnosis of autism: Advances and challenges", October 2009), has discussed some relevant research into problems of early diagnosis. Sheinkopf points out that autism can be diagnosed at better than a chance level in children as young as 18 to 25 months. However, these diagnoses are subject to the occurrence of a large number of false positives-- cases in which a child is diagnosed as autistic, but later on shows non-autistic development. According to Sheinkopf, 20-30% of children diagnosed as autistic before age 3 years will not receive the same diagnosis when re-evaluated between ages 3 and 4.
Why so many false positives? The fact is that there are many overlaps between behaviors of typically-developing toddlers and those of autistic toddlers and preschoolers. Typically-developing children leave these behaviors behind as they continue to develop, but autistic children do not. For example, typically-developing toddlers may echo others' words or may refer to themselves as "you" (as in "Give you dat cookie!"). They may also show the repetitive behaviors like flipping light switches over and over, and the intense interests in specific objects or actions so characteristic of older autistic children. Typically-developing children may also show slow expressive language development, but catch up rapidly at some later point. Of course, there are also many typically-developing toddlers who do NOT show these autistic-like behaviors, but there are many who do and thus can receive a false positive diagnosis of autism, which will be corrected if they are re-evaluated in another year or two.
Thus, our concern about providing early intervention creates a situation where some children who will later appear to have typical development are initially diagnosed as autistic, adding to the numbers of such diagnoses. For this reason and the other reasons discussed above, we need careful research before we assume that there is an epidemic. And, if it turns out that there is no epidemic, it makes little sense to look for something that has changed and caused an increase in autism. Of course, none of this makes it any less important to understand the causes and effective treatments of autism.
