Showing posts with label statistics. Show all posts
Showing posts with label statistics. Show all posts

Tuesday 28 March 2023

Are any autism statistics credible? Most are not.

The best place for many statistics

In 2021 the UK government carried out a census and in that census there were some voluntary questions about sexual orientation. A surprisingly large 92.5% of respondents, aged 16 or over, answered the question on sexual orientation and just 3.2% identified as gay, lesbian, bisexual, or another sexual orientation (LGBT+). 

Even in the gay capital of the UK, a city called Brighton, only 14% identified as L or G. 

2021 census: What do we know about the LGBT+ population?

Go back 40 years and many LGBT people would have undoubtedly lied about their orientation, or just refused to answer the question. In most Western countries this is no longer the case – the pendulum has swung very much in favour of all minorities.

When my elder son was applying for jobs, the first part was always online and one of the first questions he faced is what is your sexual orientation – that is meant to be a sign of progress, apparently.  Having started work he discovered that the real key factor of the selection was, not surprisingly, whether you will fit in and what you like to drink down at the pub. I guess they want meat-eating colleagues who like a drink, or two.  Not a place for teetotal vegetarians.

If you watch CNN, the BBC or read news from a liberal source you might think that 20% were LGBT.  I would have guessed 10%.  In research targeting young people you also get vastly inflated figures, because their views are shaped by social media which rarely reflects reality.


Back to Autism

In the current diagnostic framework autism is split into level 1 (least affected), level 2 (moderate) and level 3 (most affected).

In some countries only a small number of very specific people can make a diagnosis, whereas in others a much broader group can hand out a diagnosis. In some countries you can effectively buy the diagnosis you are seeking.

In some countries only kids with an autism diagnosis can get free early intervention. Some doctors are diagnosing autism in a toddler with an intellectual/development delay because they know he is likely to benefit, even though he does not technically qualify.

In Australia having a level 1 diagnosis does not automatically entitle you to any free services and it was recently reported that doctors are changing the diagnosis to increase the child’s entitlements. 

Children are being diagnosed with autism more severely to secure NDIS funding

When I visited our local special school many years ago to meet their piano teacher, I learned that the school had a large Roma population and so some autism parents did not want to send their children there.  I asked why there were so many Roma kids. I was told they do not have autism, they are classified as having MR/ID, which then entitles the parents to financial support.

Many readers of this blog told me that if they improve their child’s cognition they will lose supports, so they keep on using an old IQ test result.  


So do I believe this recently published chart from the US?


It tells us that in 4 just years the number of kids with autism and MR/ID has doubled to  more than 1%.

It is not credible.

Here is the source:




Here is a chart showing autism prevalence in specific states in the US in 2020.

California leads the pack.


“Only” 22% of kids with ASD in California also had intellectual disability (MR/ID) compared to 48% in Arkansas.

But, that is because kids are twice as likely to get an autism diagnosis if they live in California.  This is because they are diagnosing many minor cases that would not get diagnosed in Arkansas.






Current incidence of autism

If 4.5% of 8 year old Californians have autism, that likely equates to something like 7% of boys and 2% of girls.

We know that about 20% of school children have some special educational need.  We know that about 4% of Americans will develop bipolar disorder, about 0.5% will develop schizophrenia and approximately 9% of adults in the U.S. have some type of personality disorder.

Let’s assume 10% of Americans are LGBT, well maybe better say Californians, to keep the peace.

We know that many LGBT people experience bullying and exclusion which will give some people symptoms that do indeed overlap with those of mild autism and hence they may choose to identify as autistic, get a diagnosis, join the autism club and make TikTok videos.


Future incidence of “autism”

I would predict the published incidence of “autism” in 8 year old Californian boys will reach 20% in the next decade.

Am I worried? Not really. It’s nonsense.

Any increase in genuine, severe, non or limited-verbal autism, with IQ<70 is a problem.  It could and should be addressed, but it will not be.

By the way, this kind of autism does not need the new name “Profound Autism”, that has been proposed by the Lancet Commission.  It already has several names, including Classic Autism and Autistic Disorder.  It just needs medical treatment! Go from level 3 to level 2, maybe passing some Australians going the other direction, seeking more money from NDIS.


Statistics in Autism Clinical Trials

All clinical trials involve statisticians, lots of data and hence lots of charts.

Unfortunately, almost all autism clinical trials are flawed from the outset.  There is no singular autism, but rather hundreds of biological variations that produce symptoms that appear to overlap with this fuzzy autism spectrum.  

Lump in all these different types of autism and of very different levels of severity and give all the kids the same therapy.

When a subgroup does respond, ignore it because it is too small; the overall clinical response does not satisfy the goal/endpoint of the trial.  The trial is branded a failure.

In the bumetanide trial as an example, what percentage of parents actually gave the pill every morning to the kids for the duration of the trial?  If the trial was during the school term, there will have been problems with needing to pee on the way to school and for the first 2 hours at school. After complaints from school and accidents in the car, what percentage really gave the pill every day?

As some readers of this blog have proposed, “I’ll just give it at the weekend”.

Combine the behavior of the parents, the school bus driver and the class teacher with the trial using a dose 50% too low, is it a surprise the statistics show that the phase 3 trial failed? Not to me and many others.

Statistics on Treatable ID and single gene autisms

We are told that syndromes leading to treatable types of intellectual disability (MR/ID) are so rare that it is not cost effective to screen children for them.

We are also told that there are numerous single gene autisms, but that they are ultra rare.

Since almost no children with autism are routinely screened using genetic testing, there is no way to know just how rare treatable ID or single gene autism actually are. 

The current statistics on the incidence rates are nonsense.


Garbage in, garbage out.

Since it was first coined in the world of computer science in the 1950s, the phrase “garbage in, garbage out” has been a popular metaphor for flawed, or nonsense data input that produces flawed or nonsense output, aka “garbage.”

Mark Twain popularised the phrase "lies, damned lies, and statistics" to describe the persuasive power of statistics to bolster weak arguments.

Or, as I would say, if you need a statistician to prove your point, you probably don’t have one.

Thursday 30 August 2018

Éric Fombonne on Sloppy Autism Statistics

The reason I have a short list of talented researchers as my Dean’s List for this blog, is because of the generally low standard of much you can read about autism, even sometimes from governmental bodies like the CDC (Centers for Disease Control and Prevention) in the in the US and NICE (National Institute for Health and Care Excellence) in the UK.

I suppose these days it’s just called Fake News

My contribution is to highlight the researchers I think are worth paying attention to.
When it comes to the prevalence of autism, Éric Fombonne is a researcher who has more than his share of common sense.  Fombonne is a French psychiatrist and epidemiologist who also worked in the UK and Canada before moving to the US.
He recently gave the interview below, which highlights glaring errors/weaknesses in reports which are picked up by the mass media and put forward as facts. 

“… the CDC does not attempt to assess everybody in a population. Instead, it looks in the medical and special education records of each child in a certain region to determine whether a child meets criteria for its surveillance definition of autism. Children with no relevant notations of social problems in their records are assumed not to have autism.”
“In a study designed to validate these, researchers found that between 20 and 40 percent of children who met the CDC definition of autism did not actually have autism”
“Another issue is that of every 100 children the CDC researchers determine to have autism, only 80 have a reference to autism specifically in their medical or school records. So, one in five 8-year-old children the CDC decides has autism had never been picked up as autistic by any professional. At age 8, how is this likely?”
“I led a team that verified autism diagnoses prior to inclusion in a neuroimaging study. Trained researchers performed independent state-of-the-art assessments of over 200 children with an existing autism diagnosis. At least 30 percent of these children turned out to not have autism. It was mind-boggling.”
“In my clinics 25 years ago, I remember explaining to parents who had no clue about autism why their child qualified for the diagnosis. Things have now reversed. Nowadays, some parents and professionals push for that diagnosis and resist a ‘not autism’ conclusion because it may come with less support.”

I did a while back do my own review of autism prevalence, based on two studies carried out by Éric Fombonne.
My conclusion then was that prevalence is about 1% and that you can split it roughly into thirds.
·      One third has severe autism, which we can also call SDA (Strictly Defined Autism), or DSM3 autism.  DSM 3 autism is what medical professionals called autism 30 years ago, when it was rare. At that time it appears that many people with autism were only given a diagnosis of mental retardation (MR), which recently became intellectual disability (ID). The number of people diagnosed with MR/ID has fallen substantially as the number of people diagnosed with autism has increased, a clear case of substitution. 
·      One third have true Asperger’s, by which I mean above average IQ, no speech delay in childhood, but with substantial issues getting on with typical people.

·      The remaining third are people without MR/ID or a high IQ, may or may not have had a speech delay, but face significant challenges interacting with the wider world.  Surprisingly many people who today hold an Asperger’s diagnosis have below average IQ.

While someone with Down Syndrome (DS) will fit the diagnostic criteria of MR/ID and quite possibly autism, I think giving multiple diagnoses is unhelpful.
It is clear from the statistics that doctors stopped diagnosing people with MR/ID and instead switched many of these people to an autism diagnosis. 
Autism really should be seen as just a catch-all observational diagnosis, pending an accurate biological diagnosis, like Rett syndrome, Mitochondrial Disease etc.

The increase in prevalence of SDA
Due to the broadening of the definition of autism from DSM3 to DSM5 and, as Fombonne highlights, the very poor application of those ever moving diagnostic guidelines, we are left unable to draw any reliable conclusions about any increase in prevalence of SDA.
I certainly cannot prove it, but I believe the prevalence of SDA has increased substantially and most likely will continue to rise as society moves further from where it came from. This takes us back to the idea of the holobiont/hologenome.

The hologenome concept of evolution, considers a human as a community, or a holobiont - the host plus all of its symbiotic microbes. The collective genomes of the holobiont form a hologenome.

This becomes very relevant in human disease because in modern life humans have become separated from part of their evolutional holobiont (symbiotic microbes).  As a result all kinds on immune disease have become more prevalent.

To reduce the incidence of future SDA, steps could be taken now, just like public health tries to reduce the incidence of future heart disease.
You cannot stop random genetic mutations that lead to features diagnosed as autism, but you can influence many of the environmental factors and very likely you could take protective measures to increase the robustness of the mother and baby’s defensive mechanisms.

Friday 30 May 2014

Lies, Damned Lies and Autism Statistics

When used correctly, statistics are extremely useful to diagnose and solve all kinds of problems.  In the wrong hands, meaningless or mixed-up data can be portrayed as impressive statistics, on which other people faithfully rely, to form their opinions.

Clearly there is nothing new in this; the phrase “lies, damned lies and statistics” was popularized by Mark Twain (1835-1910).  The advent of cheap computers and desktop publishing has made it possible for just about anyone to crunch some numbers and make impressive looking tables and graphs.   Along comes the internet and all of a sudden, somebody’s idea can go viral and be quoted back to them later, as a “fact”.

This blog is based mainly on the blue skies research, being carried out at leading universities around the world.  I say “leading universities” because there are now so many universities/institutes, that not all academics are equal.  In the case of autism, it seems that the more someone publishes, the less likely it is to be worth reading.

So we have to filter out the bad science and just focus on the good.

Every now and again, I get redirected to the world of other people searching for autism treatments and then the floodgates open and I am awash in a sea of wonder cures, false promises and lots of statistics; sadly these are almost entirely the type of bogus statistics that Mark Twain referred to.

I do my best not to look at these statistics and endless articles in newspapers, based on them. But some basic statistics do matter, if you want to figure out your type of autism.

Autism affects males more than females

This seems to be agreed everywhere.  It seems that about 80% of cases are male and 20% female.  To me, it seems highly relevant that the 20% of females are skewed towards the more severe end of the spectrum.

In Retts syndrome, almost all cases are female; but this is because the male fetuses do not survive until birth.  So it seems again true that the female brain is better protected than the male brain; you would hope that this would have caught someone’s interest as an avenue of thought that might lead to a therapy, sadly not.

Is the Incidence in Adults lower than in Children?

Since autism is a lifelong disorder, why is it that people always focus on young children?  Where are all the adults?  The sad answer to this is that the adults tend to get forgotten.  As time passes, they are without their vocal parent advocates.

 Conclusions: Conducting epidemiologic research on ASD in adults is feasible. The prevalence of ASD in this population is similar to that found in children. The lack of an association with age is consistent with there having been no increase in prevalence and with its causes being temporally constant. Adults with ASD living in the community are socially disadvantaged and tend to be unrecognized

The above paper, from 2011, is based on research among adults in the community.  The researchers found the same incidence of autism, about 1%, in all age groups; it was just that in the older age groups the affected people had never been diagnosed.  They were either tagged as intellectually impaired at one end of the spectrum or “odd loners” at the other end.

If the incidence in adults is the same as in children, then all talk of an autism epidemic would be nonsense.  The US CDC autism incidence statistics would also be nonsense, rather than just a misrepresentation of data.

What percentage of people with autism are non-verbal?

This is a question that worries parents of very young children with autism.  But before looking for statistics, beware what people mean by non-verbal.  You might think this means kids who never talk, never say “Mum/Mom”, or “Dad”; but in fact plenty of parents refer to their child, who is not using 5 word sentences, as being “non-verbal”; they really mean “not as verbal as I’d like”.

Here is some data quoted by the Simons Foundation

“Roughly 25 percent of people with autism speak few or no words. A generation ago, that figure was closer to 50 percent. Most researchers agree that the decline is due to the recognition of more people with milder forms of autism, as well as to the advent of early intervention programs, that have helped more children develop language than in the past.”

What percentage of people with autism have mental retardation / intellectual disability?

This question is raised often by people, who feel the autism bandwagon has been hijacked by some of the very articulate people with Asperger’s.

The WHO (World Health Organization) does state that it thinks that 50% of people with autism have MR.  But of course what counts as autism to the WHO will not match what counts as autism according to the US DSM.  The WHO mean a more severe autism.

From the academic world we have this:-

“Autism used to be considered a rare disorder with a population prevalence of about 0.04%, of whom 70–80% had a significant learning disability. More recently, the extended spectrum of autistic disorder gives a population prevalence of at least 0.6%, of whom 70–90% are of normal learning ability. So far, the evidence is that this shift can be explained by changing concepts and diagnostic boundaries as well as by the wider recognition of autistic-spectrum disorders rather than by any real substantial increase (Fombonne, 2003).”

Back to the Research

Now we come back to the focus of this blog, which is to find safe and effective drug treatments that improve autism.

While most of the blue skies research in this blog comes from top universities like Harvard, MIT and Stanford, when it comes to applied research we do drop down a notch of two.

At some point we enter the zone of pseudo-science and then complete quackery.
Where to draw these lines is subjective and it may be that an over-zealous practitioner may be mixing elements of good science, with elements of quackery.

Here again, we are confronted with statistics, this time from the impressively sounding Autism Research Institute (formerly linked to the DAN! doctor people, who then became MAPS doctors).

At first you might be thinking, WOW!  Just look at all those interventions to try.  Indeed you could spend many years, and much money, working through all these drugs, supplements and diets.

Mixed in the list are some really good ideas (like melatonin), but there are also some things that have been shown in serious clinical trials to be totally ineffective, like secretin and fish oil.

Have all the participants heard of something called the placebo effect?
All diets, it seems, regardless of which one, work wonders for 45% to 71% of people. 

As expected, chelation comes out well, but they miss the real point.  Chelating agents are all anti-oxidants; so if you suffer oxidative stress, chelating agents will all help you, but it has nothing to do with “detoxifying” you from heavy metals.

If you would like more from the same source:-

But beware; this is not an expert scientific review of the literature, done by an Ivy League university (any more than my blog is).  It is interesting and it does have some good points, but it is not quite the authoritative resource it makes itself out to be.   The author is at Arizona State University and has a Ph.D. in Materials Engineering, which does not mean he is not right, but it is an odd qualification to be co-leader of the Scientific Advisory Committee of the "Autism Research Institute".

I expect for some of the practitioners using these methods, there are financial incentives, like renting you a home hyperbaric chamber, and others are just well intentioned people trying to help desperate parents; but these are Mark Twain statistics.

I am with Dr Chez, if you have a non-verbal child, steroids may make a lot of sense to try.  The data from ARI would not support this, but of the 204 cases they report on, what where their symptoms? What kind of autism was it? Regressive/Early onset.  What was the objective? Treat SIB, stereotypy, hyperactivity?  Nobody knows.  This is Mickey Mouse medicine and part of the reason mainstream medicine continues to assume autism is untreatable.

My Conclusion

I will continue to focus on the fundamental research from very clever scientists, who you can identify from their educational background and where they now work.  The highly prolific publishers, sometime expert witnesses and people trying to sell you something, I will try to ignore.  

Whatever type of autism you are interested in, if you follow the underlying science, you have a great chance of making progress.