Within the autism community, there's a heated debate on whether or not the prevalence of autism has remained stable since Kanner first wrote about the condition in the early 1940s. A few prevalence studies published between 1966 and 1985 estimated the prevalence at about 1 in 2,500. Many years later, other studies have estimated a prevalence rate of autism at 1 in 100 or higher. The latest CDC numbers are 1 in 59. Some persons believe that the main reason for the increase is greater identification of individuals and change in diagnostic criteria. Others believe there is some sort of environmental toxin or toxins, driving a dramatic increase in numbers in a short time frame. There is no consensus on whether or not there are far less autistics born in 1970 and earlier than those born in 1990 and later.
Autism is usually considered a disorder of childhood. However,
autistics are not Peter Pans who remain children forever. Those of us
with autism grow up into adults. Assuming there's been a genuine increase in numbers, the huge increase started with people born in 1990 and later. Many of these individuals are now adults. Children age out of services provided by the IDEA when they turn 21. Some state regional centers provide group home funding once the more severe individuals reach adulthood. There have been published reports of autistics having severe health issues as opposed to the general population as well as much higher rates of suicide. How can we identify these adults, find out what their issues are and help them?
When Hillary Clinton ran for president in 2016, she proposed doing an adult prevalence study on autism.
IACC public member John Elder Robison has written a take on this topic.
In short, JER has addressed the problems of health issues and suicide, etc. in autistic adults. He proposes to spend millions of dollars in taxpayer money doing an adult prevalence study to identify this population and find out what their issues are.
He proposes to combine methodologies of two prominent studies to find the prevalence of autism in adults. First are the reports delivered every two years by the autism and developmental disabilities network monitoring which assesses prevalence in autistic children every two years who are 8 years old at the time the survey is carried out. The second is a survey carried out by T. Brugha, et. al. that assessed the prevalence of autistic adults (16 years and older) in the UK. They found a nearly 1% prevalence, which was about the same as had been reported in autistic children at the time. You can read the survey in its entirety here.
How valid were these two surveys in assessing the prevalence of autism in children and adults respectively? Will Robison's recommendations help us find a valid adult prevalence? Do his talking points have any credibility? I will attempt to answer these questions in this blog post.
On careful scrutiny of Robison's talking points, it seems he's made many factually inaccurate statements. He appears to have not even read the primary sources of the articles he cites.
Brugha et al.  found very high rates of autism among older adults previously diagnosed with learning disabilities, and they found alarge number of autistic people living in group or institu-tional settings. Neither of thosefindings are surprising.
Actually Brugha et. al found 19 cases out of 7,461 persons surveyed in the first phase of his study in a country of 30 million population. Based on a statistical analysis, which at best was educated guesswork on his part, he surmised that there were an additional 55 cases on top of the 19 that were assessed using the ADOS to diagnosis their autism that were in the survey. Most of these had intact intelligence though there were some who responded to the survey by proxy due to being too disabled. These persons did not live in institutional settings but in regular households. I realize though JER's standard for "very high" might be different than mine.
According to Brugha et al. , with our current bestestimates of prevalence, roughly one in 50 men, or onein 75 people of all ages are on the autism spectrum. Thisis consistent with contemporaneous prevalence estimatesfor children, such as the 1 in 56 reported by the CDC’sADDM (Autism and Developmental Disabilities Monitor-ing Network)
Actually Brugha found not one in 75, but 98 out of 1000 or slightly less than 1 in 100. The CDC's 2018 report found 26.6 in 1000 boys or one in 38, rounded to the nearest non-decimal number. /The CDC's number in 2018 was actually 1 in 59 not 1 in 56. These might be considered minor factual errors, but I feel as an autism public policy advisor at taxpayer expense, Robison should get these figures right.
He uses the study by Lisa Croen to state that there were only less than one tenth of the adults there should be in that survey. I don't believe a sample of the general population and a selected sample of people in the Kaiser insurance program is an apples to apples comparison, as there would be selection bias as to who would have this insurance or wouldn't. Adults outside the Kaiser system might have MediCal and seek health at other providers or not have insurance of any kind. Another possibility is that there is not a 1% prevalence or higher in this age group as has been found in children.
With some autisticadults described as tech geniuses while others require24/7 support for basic living, parents rightly wonder whatthe future holds for their autistic children. We owe it tothem tofind some answer
How can an adult prevalence study predict what the future holds for autistic children? How will spending taxpayer money in this manner do this? You don't need a multimillion dollar taxpayer financed study to know that the prognosis for most autistic children is very poor.
In conversation with autistic people, I have often beentold that adults deliberately keep their autism diagnosisout of the medical record, to avoid“having it usedagainst them.”Perceived threats include higher lifeinsurance premiums, higher health insurance premiums,and possible denial of professional licenses or even jobs.Those are all valid concerns and suggest there may be agood number of autistic adults who are aware of theircondition but choose tofly under the radar.
John may not be aware of this, but under the health insurance privacy protection act, it's a federal offense for a hospital or doctor to share medical records without someone's consent. Very few autistic people have children so life insurance would not be a concern. Because of lobbying from autism speaks, insurance now pays for many autism treatments, but it did not in the past, so it's unlikely an autism diagnosis would result in higher insurance premiums for a pre-existing condition. This is even less of a concern since the affordable health care act was passed. Very few autistic people will become eligible for any type of professional licensing. A prospective employer is unlikely to ask someone for their medical records and again under HIPPA it's illegal for a hospital or insurance company to produce these without the person's consent.
A few years ago, researchers believed most autistic peoplewere intellectually disabled. That belief was founded on alimited understanding—they only recognized a small part ofthe autism spectrum as we know it today. Now that our sam-ple is larger, we understand that the average IQ of the autisticpopulation is much closer to that of the general population
Actually only 1% of the population has an intellectual disability, whereas probably at least one in three autistic persons has one. It is probably closer to 40%. This includes findings from the ADDM surveys that Robison takes such high stock in.
In addition to the 1% prevalence of adults in the Brugha study just being a guess, this survey has other problems. It turns out the AQ-20, an abridged version of Simon Baron Cohen's AQ test, was a very poor predictor of who is and who isn't autistic. The blogger autism jabberwocky wrote about this so it would appear the Brugha survey, admitted by the author himself, has no credibility whatsoever.
The authors divided their prevalence data into three age groups, stating there was no difference in prevalence between the groups. But with a sample of only 19 people it does not really have the statistical power to say this.
It appears John Robison just read portions of the abstract of the study without reading Brugha's work in its entirety.
John Also wants to use the ADDM reports as a model for his plan to assess adult prevalence. These reports, published every two years cost the American taxpayers 5 million dollars a year. The reports only cover a cross section of the entire USA. They use states such as New Jersey each year, which as a more generous service delivery, providing an incentive for the parents of children to get an autism dx. This state only has a fraction of the population of the entire USA and thus skews the average. Just a small percentage of the united states instead of the entire country is used to assess these prevalence numbers. Also, a significant percentage, maybe as much as 20% in some years of the autism cases are not actually diagnosed. Rather the epidemiologists who are not trained clinicians and have never even seen the actual 8 year old children rely on health and school records to come to the determination that these children are autistic and provide them as part of the prevalence numbers.
To me it is worrisome that JER can get articles such as these that pass peer review in journals of questionable quality. It troubles me that he can carry out his activities at taxpayer expense.
Another point to ponder is whether or not JER's motivation in making this proposal is to actually produce data that will be of use to service providers and others to help autistic adults. Hillary Clinton's proposal of an adult prevalence study came after a conference call with ASAN. The neurodiversity movement has insisted that autism is a natural variation and 100% genetic. This is despite evidence of studies of fraternal twins showing higher rates of concordance than non twin siblings as well as environmental insults such as thalidomide showing to cause autism. These argue against an exclusively 100% genetic component. If a neurodiversity sponsored study found an adult prevalence rate identical to those of children born in 2000 or later, this would certainly bolster their hypothesis. John has an intimate friendship with Alex Plank and other ND extremists. If you go on his twitter and look at his tweets and replies, you can see some of the most radical ND proponents who block those who disagree with him are having dialogues with him. Could this be at least a factor in his desire for an adult prevalence study?
My own take on an adult prevalence study is that it's like looking for a needle in a haystack. Adults don't present to special ed schools, clinics etc. So it's impossible to do the aggressive case finding that the CDC does to find people over the age of 21 who are on the spectrum. I believe the question on whether or not there are just as many autistics born in 1970 or earlier as opposed to 1990 or later will never be resolved. I don't believe a valid prevalence study is possible to do on this group. There is no point in spending American tax dollars on this endeavor that the British did on the Brugha surveys and study.
Are JER's proposals to spend millions more on taxpayer dollars, based on a model of two different troubling epidemiologic studies worth it? Is this a person who should be making policy recommendations for autism at taxpayer expense? Autism's Gadfly does not think so.