"Identifiying the lost generation of adults with autism" (Baron-Cohen & Lai) is outrageously misleading.

This post relates to this paper here: https://www.ncbi.nlm.nih.gov/pubmed/26544750

This review by Dr Meng-Chuan Lai and Dr Simon Baron-Cohen (SBC) should be retracted for being outrageously misleading.  It starts with an assertion of a highly dubious massively important controversial “fact” with no citation whatsoever in support.

The opening sentence (of both abstract and text) asserts to us that as of 2014-15, 1% of 10-year-olds are autistic, and 1% of 30-year-olds are autistic, 1% of 50-year-olds are autistic, and so on, for “all ages”.  And the title of the paper reinforces that untruth with the notion that there is therefore a “lost generation” of older autistics who have mostly remained unrecognised as autistic and thus “lost”.  The un-explicated implication being that there has not been any increase of autism but instead only an increase of awareness or diagnosis.  And this paper sets out from its beginning by grossly misleading the reader into assuming that that is known to be a fact, indeed established to such an extent it doesn’t even need any citation in support. 

And that title and first sentence are not incidental to the paper but are its central premises. 

I sent emails to the authors, querying the justification for that first sentence.  I got no reply from SBC. 

Dr Lai sent a series of replies.  The full correspondence is too lengthy to include here in full, but here are some main excerpts.

Dear Robin,

Thank you for your message. For clarification, when we write “Autism spectrum conditions (panel 1) comprise a set of neurodevelopmental syndromes with a population prevalence of 1% across all ages.”  we are NOT referring the term “ages” to different chronological years (i.e., 1970s, 80s, 90s, etc.) but are referring it to the ages of the individuals; that is, we are referring to epidemiological evidences that recent cross-sectional studies in children, in teenagers and in adults all tend to show a prevalence around 1% (depending on studies but can range from around ~0.5% to ~2%). We, therefore, have no intention to argue in the paper whether the prevalence of autism is constant or not across different chronological years.

Hope this clarifies the question/mis-understanding.

Regards,

Meng-Chuan

Dear Meng-Chuan,

 Thanks for your reply.  Unfortunately it appears to me to be completely bonkers.  Your phrase “all ages” includes obviously those who are 40 or 60 years old.  And even older.  As of 2015, a person who is age 40 was necessarily born in 1975 or earlier, and a person of age 60 was necessarily born in 1955 or earlier.  That is, such ages of persons are impossible to disentangle from the huge change of apparent incidence (per birth-year cohorts) which is the basis of the notion of the autism increase.  I fail to see any way you can separate ages (of “all ages”) from specific years of birth. 

 In response to my request for any evidence to support that claim of your first sentence, you didn't send any, but instead included a sentence contradicting it.  And that notion of “1% across all ages” is fundamental to the point of your entire article, namely the notion of a “lost generation” of older autistics. 

 (If I may do a bit of your homework for you here, perhaps the cited Brugha et al. could be supposed to constitute supporting evidence on the matter.  But Brugha et al. is fundamentally unsound.  They went to great lengths for establishing reliability of the assessments.  But no establishing of the (infinitely more important) validity of the assessments, that is establishing that the measures at age x were equivalent to the measures at age y.   And that is because it is logically impossible to do that validity verification (except maybe over many years).  It's a bit like if I put a steel ruler in a hot fire and declared that I can see that its own scale shows it to be still exactly 12 inches long, so “therefore” the heating hasn't changed its length.  For this reason Brugha's (and any similar studies) amounts to no evidence whatsoever for countering the evidence of Hertz-Picciotto and others.) 

[....] (I am aware of many recent younger autistics being under-recognised [....].) 

 But your present article is simply unacceptable (seriously misleading) and has to be retracted. 

 Sincerely, Robin P Clarke

Dear Robin,

Thank you for the opportunity for further clarification. If I understand your points correctly, the main critique is based on your perception: "that notion of "1% across all ages" is fundamental to the point of your entire article, namely the notion of a "lost generation" of older autistics".

I am sorry if this is how you perceive our arguments. However, this is not what we have in mind when we wrote it. You are right that the paper mainly discusses under-recognized autism in adults - indeed the sole focus of the review, and from a clinical perspective. We are asked by the journal to provide general epidemiological background (based on published studies), and this is where the ~1% number comes from - and we provided the number in adults, referencing the Brugha paper which is one of the few published study focusing on adults.  We DO NOT mean to argue "there is a constant 1% prevalence over the years so there is a lost generation".  We simply refer to the current epidemiological findings that adult prevalence at the time being is also around 1%, and many of them have not been diagnosed in the past.  I agree it is difficult to get the ground truth (due to issues with validity of diagnosis, changes of diagnostic concepts [that we have discussed in the review], and not knowing whether there is true incidence change); the prevalence number cited here is simply giving a general picture to the readers, from available published epidemiological data. 

It is possible that there is increased incidence over the years, but this does not contradict the possibility of people being under-recognized as a child. We focus on the latter and have no intention to discuss (either support or refute) the former.

I hope this helps clarify where this review comes from.

Regards,

Meng-Chuan

Still not having heard anything from SBC himself, on 9^th January I posted a letter to him by Signed-For-Delivery (KP597377771GB at royalmail.com) as follows.

Dear Dr Baron-Cohen,

I am writing here in respect of your article "Identifying the lost generation of adults with autism spectrum" in Lancet Psychiatry 2015.

I am concerned that I have not yet heard from you about this article.  I hope you are not unwell.

I will send herewith a copy of my correspondence with your co-author Meng-Chuan Lai.  His replies get even more absurd and indefensible than the original article (indeed raising doubt about his sanity).  This matter of the huge tragedy of the manyfold increase of autism is very serious.  Publishing of sloppy, wantonly misleading writing about it is completely unacceptable and unethical.  It is all the more unacceptable that you yourself put your name to such a grossly misleading and unworthy document, given that you are widely trusted as “the leading expert” on autism.  And further that this is the very opening of a “review” being published in the Lancet no less, thus likely to be “authoritatively” parrotted by future generations of  students as supposed established knowledge. 

I included two of your email addresses in my correspondence with Meng-Chuan (sb205@cam and editorial@molecular) but have not heard any comment from yourself on the matter.  I do not know whether you are aware or not of these issues (though as co-author you should have been aware of the opening sentences of your own article anyway). 

Accordingly I am now sending this letter by signed delivery requesting that you clarify your position on this very important matter.  Do you agree with myself and others that the article is unacceptably misleading and must be retracted?  Or what?  

I look forward to hearing from you as soon as practical.  You have one of my email addresses at the top here, and another is rpclarke@autism. 

Sincerely, Robin P Clarke

I have still not received any reply to this letter, by email, post or any other means.  Why not?

Ann Dachel posted an “open letter” to Dr Baron-Cohen on the AgeofAutism website (Dachel, 2009).  Her letter ended as follows. 

“I have only two questions for you:
 1.  [....]
 2.  [....] I'd like you to show us the 30, 50, and 70 year old adults who display the same symptoms of classic autism that we see in children, the non-verbal adults in diapers, banging holes in walls and spinning in circles.  [...]”

The reply from SBC went on quite a bit, but did not give the slightest answer to those questions.  Again, why not? 

And this hugely misleading from SBC comes in a context of his twice blocking my own evidence about the increase.  When I sent to his Molecular Autism journal my epidemics paper, he replied with an assertion that “The advice we have received is that the methodology would not get through critical peer review from our journal”.  And yet this is a pseudo-reason, because it is the editors themselves (i.e. himself) who choose the peer reviewers and thereafter decide whether or not the peer reviewers have shown the paper to be inadequate or not. 

And a second occurrence of blocking by SBC occurred when I sent to his journal my paper about road traffic pollution as a supposed cause of autism.  He refused to publish it unless I cut out all mention of my alternative explanation (in terms of reduced ventilation of the mercury vapour emitting from non-gamma-2 dental amalgams).