Identification Bias

Because autism was first taken to be a medical condition—a type of mental illness—its ongoing recognition has always suffered from a very bad case of identification bias.

Around forty to fifty years ago, the only cases of autism being diagnosed were those for which the individual was clearly detached from usual behaviors and circumstances, and was clearly in need of support and services. This practice made sense, because at that time autism was assumed to be an acute medical condition and therefore only individuals meeting the criteria of an acute medical condition could be identified as having autism. No one else was ever identified as having autism, and of course autism was thought to be extremely rare.

However, a problem soon emerged: the medical community began recognizing many individuals who possessed most of the same features as those being diagnosed with autism, but who obviously were not suffering from an acute medical condition. This should have alerted the medical community that its initial assumption about autism being a medical condition was now seriously in doubt; but identification bias had already set in, had already become too firmly entrenched, and no serious questioning of that assumption ever took place. When one reads, for instance, Lorna Wing’s work—the work that began the process of bringing Asperger syndrome into the diagnostic fold—one is struck by how the new identifying criteria, although distinguished from the criteria used on the more classically diagnosed cases, was still making the assumption of a major underlying cognitive impairment, was still assuming the presence of a serious medical condition. And so, as terms such as high-functioning autism, Asperger syndrome, and pervasive developmental disorder began to assume more widespread usage, the specter of medical illness never loosened its grip on autism at all—it merely began enveloping a much larger population within its darkening cloud.

That trend continues unabated to the present day. Using an expanding range of diagnostic tools and applying them at earlier and earlier ages, the medical community now identifies nearly one percent of the toddler population as possessing some form of autism, and because of identification bias, that entire one percent is assumed to be suffering from a dire medical condition. With no consideration for whether the expected outcome might be positive or negative, and with no recollection of the statistics from its own diagnostic past, the medical community assumes every child identified as being autistic will require a regimen of early intervention, will need an onslaught of powerful drugs, intensive therapies, and assorted prescriptive treatments. It never occurs to the medical community that its own history of autism identification has now progressed all the way from rare to questionable to absurd. It never dawns on the medical community that nearly the entire one percent now being identified as mentally ill, was only one half century ago being identified with not one sickness at all.

It is not too late to turn back the clock—even some fifty years. Autism has never been a medical condition. That stranglehold of an assumption is only the residue from a lazy history of bias.