This post serves as a gathered collection of thoughts surrounding the paper recently published in the British Journal of Psychiatry, Prevalence of Autism-spectrum Conditions: UK School-based Population Study (Baron-Cohen, Scott, Allison, Williams, Bolton, Matthews, Brayne, 2009, hereafter referred to as PAC). Let me begin by stating that this post is not intended to be a commentary on PAC's methods or accuracy—epidemiology is well outside my knowledge domain and I can leave that analysis to others. My intention here is merely to consider the reported findings of the paper and to discuss how they, if confirmed, must call into question much of the conventional wisdom regarding autism, including conventional wisdom accepted by much of the autism research community.
The two main findings reported in PAC are 1) for the population studied (school children aged 5 to 9 living in Cambridgeshire), the rate of diagnosed autism-spectrum conditions is roughly 1.0%; and 2) for the same population, the true prevalence of autism—that is, the percentage of those who meet diagnostic criteria for an autism-spectrum condition, whether or not they have been officially diagnosed—is roughly 1.5%.
For the purpose of this discussion I am going to assume these findings will be confirmed, and furthermore, I am going to assume they will be found to be roughly applicable for many populations outside the study—for instance, the populations of the industrialized nations such as Canada, Great Britain, and the United States. This last assumption is of course a bit uncertain, because there are reasons to believe the Cambridgeshire population might possess unique characteristics affecting its autism rate; but as a rough measure, I think it not unreasonable to assume Cambridgeshire autism rates are similar to autism rates for many other geographic locations.
Finally, I am going to add one more assumption that is not a finding of PAC itself: I am going to assume the true prevalence of autism has remained steady over the last thirty years or so—that is, the true prevalence of autism (as defined in PAC) was 1.5% around 1980, and has remained near that level to the present day. Not everyone would agree with this assumption, of course, but there are many autism researchers who seem to indicate it is both plausible and likely, and those who do not agree with this assumption—that is, those who posit a recent “epidemic” of autism—face the daunting challenge of matching this “epidemic” assumption to everything else that is known about autism. Only time will tell which understanding is most correct, but the intention of this particular discussion is to consider the ramifications of assuming autism's prevalence has remained steady over recent decades.
In summary then, here are the findings and assumptions being addressed in this post, stated in such a way as to indicate they are being taken as accurate and broadly applicable:
The current rate of diagnosed autism (autism-spectrum conditions) for children aged 5 to 9 is 1.0%.
The current true prevalence of autism (diagnosed and undiagnosed) for children aged 5 to 9 is 1.5%.
The true prevalence of autism for children aged 5 to 9 has remained steady (near 1.5%) for at least the past thirty years or so.
These assumptions, taken as a group, lead to a number of corollaries, corollaries which have not been widely discussed but which should be considered, for they are surprising—surprising in that they defy much of the conventional wisdom already established around the subject of autism. These corollaries would include:
The vast majority of autistic adults are undiagnosed.
Autism is (on average) too non-disruptive to be considered a disorder.
Autism is too common to be considered a disorder.
The so-called “advantage” of early diagnosis, treatment, and intervention is highly questionable.
All these corollaries follow directly from the above-listed findings and assumptions. Let me discuss each corollary in turn:
The vast majority of autistic adults are undiagnosed. “Autistic” in this context means someone who could have received (or did receive) a diagnosis of autism as a school-aged child under the conditions and criteria set forth by PAC, notwithstanding how such an individual might then present as an adult. If we consider the circumstances of thirty years ago (and using the findings and assumptions listed above), this means 1.5% of the school-aged population around 1980 qualified as “autistic” under this definition, and since mortality rates appear not to differ significantly between autistic and non-autistic populations, that same 1.5% rate of autism prevalence still roughly holds for the adult population currently approaching the age of forty. By extension, we can expect a similar rate of autism prevalence throughout the entire adult population.
There are a variety of estimates given for the rate of diagnosed autism some twenty to thirty years ago. For convenience, I am going to assume the rate of diagnosed autism in school-aged children around 1980 was roughly 1 in 1000. That appears to be consistent with many of the estimates provided, and besides, the exact number is not nearly as important as the scale of difference from current rates, for any large scale difference is going to demonstrate the point that most autistic adults today are undiagnosed. For instance, if we assume the diagnosed rate of autism was 1 in 1000 in 1980 while the true prevalence of autism was holding steady at 1.5%, this means that for every school-aged autistic child who was diagnosed with autism in 1980, there were approximately 14 school-aged autistic children who were going undiagnosed. This is a tremendous gulf of difference from the present-day ratio arising out of PAC (two diagnosed for every one undiagnosed by the assumptions of this discussion; three diagnosed for every two undiagnosed by the more exact statistics presented in PAC).
It is true that some of these undiagnosed autistic children from 1980 would have received an alternative diagnosis (such as mental retardation or childhood schizophrenia), and it is also true that some of these individuals would have received an autism-spectrum diagnosis later in life. But no statistics regarding diagnostic substitution or adult-age diagnoses can account for the majority of these children who would have qualified as “autistic” back in 1980 (by PAC criteria) but who did not receive an actual diagnosis at that time. There is only one conclusion that can possibly match to the numbers and assumptions: the vast majority of autistic individuals went undiagnosed as children in previous years, and have remained undiagnosed as adults through the present day.
Autism is (on average) too non-disruptive to be considered a disorder. There are two ways of looking at this particular corollary: the first arises from consideration of the previous corollary alongside a contemplation of what must be the status of undiagnosed autistic adults who are living today, and the second arises from an examination of the expectations held for most present-day school-aged children who qualify as autistic under the conditions and criteria set forth in PAC.
If the vast majority of autistic adults living today are indeed undiagnosed, then their circumstances cannot be all that dire. If their circumstances were dire, they would be much easier to find. But the institutions are not overflowing with these individuals, and they do not seem to have congregated in other discernibly poor locales. There is only one possible location where so many undiagnosed autistic adults can remain so perfectly well hidden, both to themselves and to others, and that is within the normal confines and conditions of general human society. That is to say, they are living quite ably and quite abundantly among us. Individual outcomes will of course have varied greatly, from extremely negative to quite good, but such a range of outcomes is no different from what might be anticipated for the non-autistic population. When we examine those school-aged children from around 1980 who could have received a diagnosis of autism (by the criteria of PAC) but who did not actually receive such a diagnosis, and when we consider the circumstances that the majority of those individuals must find themselves in today, we are forced to conclude either that their autistic characteristics did not preclude them from assimilating into general adult society or that any early difficulties these individuals might have faced, without treatment and without recognition, simply eased with time. Either way, to apply the term “disordered” to such a large, well-assimilated group of individuals would necessitate a complete overhaul of modern-day semantics.
And the conclusion turns out to be no different when applied to today's autistic children, such as those who were studied in PAC. Starting with the ones who have been diagnosed with an autism-spectrum condition, it must be noted that the expectations are quite promising for a large percentage of these children. Although terms such as “high-functioning,” “Asperger syndrome,” and “mild autism” are too inconsistently applied to be statistically useful, they do indicate there is an anticipation with many diagnosed autism-spectrum children that they will be able to progress through school, possibly attend college, obtain work, and in general assimilate in some fashion to the typical circumstances of general human society. And if these are the expectations for many of the diagnosed autistic children, it must be even more so for the many autistic children who have remained undiagnosed. Although the PAC authors do not describe in detail the life circumstances of the children they categorize as undiagnosed, if I am reading between the lines correctly, the suggestion seems to be that these are children who for the most part possess diagnosable characteristics of autism but who have not been regarded as being troubled enough to be brought in for an actual diagnosis. If so, then one has to surmise the expectations are again quite good for these children, certainly no different in kind than the expectations held for most non-autistic children, and when we add the size of this group to the number of diagnosed autistic children for whom expectations are also quite positive, the notion that most autistic children are somehow in the grip of a dreadful disorder rings clearly false, for it does not match the anticipated outcomes for the large majority of these children.
None of this is to deny the circumstances of autistic individuals who face greater challenges, for they certainly exist too and are in need of a greater understanding, but the findings of PAC, along with the assumption that true autism prevalence has remained steady over recent years, leads to an inevitable conclusion that for the vast majority of autistic individuals, their autism is not so disruptive as to prevent eventual assimilation into general human society, and is not so disruptive as to warrant use of the term “disordered.”
Autism is too common to be considered a disorder. At 1.5%, autism's prevalence approaches nearly the same scale of prevalence as is seen in left-handedness and homosexuality, and as history has suggested with these and other lifelong conditions, an attempt to medicalize and demonize such a large portion of the human population leads only to unresolvable tension, rampant misunderstandings and needless suffering. Ultimately, such a course cannot be sustained.
Indeed, if a 1.5% rate of autism prevalence holds throughout the entire human population, we are talking about nearly one hundred million people all told, the large majority of whom (by the numbers alone—you cannot find one hundred million people residing in institutions) must be participating relatively productively and relatively well within the circumstances and conditions of general human society. To describe such a large segment of humanity as “disordered” and to classify its unifying condition as somehow “tragic” runs counter to all forms of medical and scientific logic. Humanity has been down this misguided path before; it should know well enough by now not to go down that path again.
The so-called “advantage” of early diagnosis, treatment, and intervention is highly questionable. If the PAC findings are accurate, and if the true prevalence of autism has remained steady over recent decades, then it is only quite recently that the majority of autistic children have begun to be exposed to diagnosis, treatment and intervention; before now, most autistic children went undiagnosed, untreated and unintervened, and as the discussions of the previous corollaries have demonstrated, most of these children must have been able to develop, progress and assimilate into general human society with a relative degree of success. It is against that background that judgments of diagnosis, treatment and intervention should always be made; but in reality, that background is almost always ignored.
When a research team announces that a form of treatment can lead to positive outcomes (usually defined as removal of autistic behaviors) for say 50% of the autistic children being studied, this news is typically greeted with great fanfare, for the unstated assumption behind such joy is that without treatment, without intervention, these children would invariably experience extremely negative outcomes, full of “terrible” autistic behaviors and leading inevitably to eventual institutionalization. But the findings of PAC clearly demonstrate that this unstated assumption is not just a little bit false but indeed is grotesquely false. The normal outcome, the expectation, is that autistic children will develop, progress and assimilate to human society, and this includes (most especially includes) those autistic children who do not experience diagnosis, treatment and intervention. Rather than greeting announced results of “50% positive outcomes” with great fanfare, we should instead be responding with extreme horror: how could a proposed treatment possibly be doing so much harm?
Indeed, when considering the findings of PAC, one would almost have to make the cheeky assertion that the most promising form of autism treatment is to remain undiagnosed. I do not myself want to be quite so cheeky, because I do believe early recognition of autism has the potential to create benefits for autistic individuals. But if the form of those “benefits” continues to be treatments and interventions aimed at combating a pseudo disorder, then I am forced to remind the autism research community that the findings of PAC suggest quite strongly that most autistic individuals were much better served in an era in which they went entirely unrecognized.
These four corollaries—derived only from the findings of PAC, along with the additional assumption that true autism prevalence has remained steady over recent decades—reveal an image of autism that runs diametrically counter to the image offered by autism's conventional wisdom. From the findings of PAC, autism emerges as a condition both common and healthy, and a condition that leads in most instances to a predictable developmental course (albeit a course somewhat different than the norm), a course of growth and eventual assimilation into general human society.
The more commonly held vision—that autism is instead a tragic brain disorder leading, without treatment and without intervention, to frightening consequences—this conventional wisdom has arisen out of autism's volatile diagnostic history, having taken hold at a time when only the most troubling cases were being recognized and autism was considered to be a circumstance exceedingly rare. But as autism's diagnosable characteristics have become more crystallized, and as it has become more apparent that those characteristics can be applied to far more cases—including many cases not nearly so troubling—autism's conventional wisdom has remained nonetheless quite stubborn, has refused against all logic to be dislodged, and thus continues to hold sway within nearly the entire autism research community.
Surely the PAC findings must be the final straw.
There is only one conceivable way to accept PAC's findings while continuing to cling to autism's conventional wisdom, and that is to negate the added assumption, to claim that instead of holding steady over recent decades, autism's true prevalence has been rapidly increasing. And because I suspect autism researchers will remain unwilling to let go of autism's conventional wisdom, that negation will now be put forward with a much greater urgency, and the holy grail hunt for the source of autism's sudden surge will continue to expand both far and wide—from proposed floods of fetal testosterone to endless laundry lists of suspicious environmental toxins. But as researchers are attempting to bridge that enormous gap between purported cause and established effect, allow me to offer instead the more promising prospects of moving decidedly on. Letting go of autism's conventional wisdom requires only very little and entails not a single moment's wait for new discoveries. Letting go of autism's conventional wisdom requires only the confirmation of PAC's reported findings, along with an acceptance that autism has been significantly present within the human population for a very long time.
