By Jake Crosby
He may not receive the same level of attention as Paul Offit, but what Columbia sociologist professor Peter Bearman lacks in publicity, he more than makes up for in autism research money. Millions of your tax dollars are being given to him through the NIH under the guidance of the Interagency Autism Coordinating Committee, whose chair, Dr. Thomas Insel, is the brother of a vaccine developer, Dr. Richard Insel. (When Vaccine Development is a Family Business. The Conflicted Role On Vaccines and Autism)
Is Bearman’s research worth the money being invested in it? This question was addressed on Age of Autism once before by Katie Wright, whose answer was pretty conclusive – “nonsense.” Kevin Leitch of Leftbrain/Rightbrain - perhaps the only place I’ve seen Peter Bearman receive a large amount of attention outside his own academic circles - leapt into the fray addressing none of Wright’s points while making a stink about how she simply won’t accept Bearman’s “science.” Little did I know at the time that this very debate would jump off the Internet and onto the Brandeis campus.
On March 24th, I attended a lecture at Brandeis University’s Heller School for Social Policy and Management held by the Lurie Institute for Disability Policy. The title of the event was “Understanding the Increased Prevalence of Autism.” The honorary speaker was none other than NIH-funded Columbia sociologist, Professor Peter Bearman.
Having read Age of Autism’s post about Bearman, Peter Bearman, Autism and the House of Nonsense, I knew what to expect, and none of what Bearman said really shocked me or was different from what I’d heard from him in the past. He went through his research - the increased risk of autism between same-sex twins and a corresponding decrease of autism between opposite-sex twins, his claims that autism is due at least in part to diagnostic substitution of the “mental retardation label,” the “social clustering,” the increased parental age, and the close spacing of births.
With the close spacing of births, Bearman did acknowledge that it was perhaps related to folate-depletion, suggesting a nutritional basis for autism spectrum disorders that might perhaps play a role. While nutrition may well be involved in the pathology of autism, historical epidemiology suggests these findings are hardly significant. Closely spaced births was a much bigger phenomenon 100 years ago when the average family size was considerably larger while autism prevalence was not even high enough to be detectable – if there was any autism back then at all – much less at epidemic levels like it is today.
After hearing Bearman talk for a good 45 minutes, the audience was allowed to ask questions. I brought up his paper in the International Journal of Epidemiology and its claims that diagnostic substitution could account for a 25% increase in the rate of autism. I told him it was based on a flawed extrapolation of a sub-group of about 9% of cases who switched from a diagnosis of mental retardation to a co-morbid or sole diagnosis of autism. I also mentioned that the conclusion was based on generalizing this group to the autism caseload as a whole.
He started to answer, then paused and said out loud, “How should I answer this?” He then gave a long-winded response about how difficult it was to quantify the increase and that he felt his methods were right, but never explained why. In his response, he made two amazing admissions:
1. The increased prevalence of autism is real. It was good to know that even Bearman admitted that the epidemic denial of anthropologist Roy Grinker – a frequent commenter on Bearman’s work – was bunk.
2. Four out of five of the reviewers of his paper had the problem I had about its incorrect extrapolation. “And now,” he said, referring to me, “five out of six,” in a half-joking manner.
Bearman still concluded, “But I think it’s a good paper.”
That is just not an acceptable answer. He might as well have admitted that the NIH was giving him millions for his opinions, because that in essence is what he just said. NIH funding is supposed to produce scientific papers, not editorials. I could just as easily get someone’s opinion by reading the comments on a HuffPo article.
Yet he still had the nerve to call studies of environmental factors “junk” in his response to another audience member’s question, after admitting that four out of five peer-reviewers have essentially said just that about his own work. Then he said that another paper of his that is up for peer-review at different journal will “probably get rejected,” which was followed by snickers from the audience – to which I thought “it’s funny to them that his research is garbage?”
Bearman also delved into the increased risk of parental age, saying this caused “de-novo mutations” in genes that was driving the increase in autism and making it more heritable. He showed us a graph of autism concordance between same-sex twins going up and between opposite-sex twins going down. Historically, twin studies have been used to claim autism is a genetic condition while spuriously omitting the fact that concordance between twin births is instead due to related environmental exposures and susceptibility. Bearman’s analysis was no exception here. He never mentioned that the older a parent is, the more exposure they’ve had to a particular substance, and the more exposure they’ve had to such a substance, the more likely it is to affect their unborn children.
With that in mind, I asked Professor Bearman if perhaps the trend in twin concordance by gender was due to the increase of something in the environment that puts a certain group of people at risk, noting that autism effects more boys than it does girls. He did not address my point about susceptibility factors – simply saying that the 4:1 ratio of boys to girls affected with autism has remained stable and offered no evidence for why “de novo” genetic mutations were more likely a culprit than environmental insults beyond the correlation to older parents.
What struck me as most incredible was his response to a student who brought up Wakefield’s case series and The Lancet’s retraction of the paper. The student, as repeated ad nauseum in the popular press, incorrectly called the paper a study linking vaccines to autism. I called out correcting her, stating that it never did.
Bearman responded that because autism exists in clusters as opposed to the “global usage of vaccines” that he does not think the vaccine-autism connection warrants further study. Of course, vaccination rates fluctuate just as autism rates do. It is absolutely bizarre that he would make such a statement, especially when his own research made almost no effort to see whether vaccination rates and autism rates were related.
Bearman also stated that vaccination exemptions were highest in places with high autism prevalence as a way of doubting the connection. Well, of course they would be; vaccination exemptions are a response to the soaring autism diagnoses. This should, if anything, intensify the urgency for vaccine-autism research.
“But what if that’s caused by fear of vaccines-induced autism?” – I asked.
He never answered my question, but did acknowledge, “I know you disagree with me.”
But Bearman’s disagreement also contradicts what he told a different audience in a video featured in Katie Wright’s article. To that audience, he said autism rates and vaccination exemptions are completely unrelated, and that autism organizations are scaring people away from vaccines – as if autism clusters form around autism organizations and not the other way around. I guess even Bearman figured that such an assertion was absurd and changed his story. Yet, the story he replaced it with was no less absurd. Evidently, he is also disagreeing with himself in the Columbia video lecture.
I wanted to discuss the spatial clustering – his own brainchild of autism study design. I noticed he said that he found significant clusters between households that share kindergarten and daycare centers, parks and malls – all places where, “people talk.” He supposed that the shared public places resulted in parents talking to one another about autism, making their children more likely to get diagnosed. Of course, places that people tend to go to more often are probably more likely to correlate to where people live – which further raises the plausibility of environmental exposure in relation to locale. Coal burning plants come to mind.
Bearman also did not mention some of the other things parents might talk about, like getting their children vaccinated, like eating foods that may be high in certain substances of interest (such as the levels of mercury in fish), or dental care (such as dental amalgam) – all of these are things that, I would assume, people talk about that could influence exposure to the cause or causes of autism.
Bearman further asserted that pediatricians had a negative affect on autism diagnoses because they are always trying to delay them! That’s quite strange – my pediatrician was the first to spot that I was delayed in my language development. Yet Bearman made his assertion without anything to back it up – no data, not even anecdotes.
He further claimed that pediatricians were not associated with autism clusters and that dentists were more strongly associated with autism clusters, since people don’t talk about getting autism diagnoses “in dental chairs.” What he left out was that in some dental clinics, amalgam fillings are often given which emit mercury! He also seemed to leave out fact that dentist’s offices - like all doctor’s offices – do have places where people talk. They’re called “waiting rooms.”
In a further attempt to try to back up his claim that spatial clustering directly relates to autism diagnoses, he noted that the clustering was more likely to happen among higher-functioning children than lower-functioning children. But that is probably for the same reason that there is a negative connection between proximity to an autism household and mental retardation, which can only be generalized to the small fraction of people who, depending on their circumstances, could qualify for an autism label, a mental retardation label, or both. Similarly, the small fraction of people who have a borderline autism diagnosis is likely to be higher functioning. But issues with generalizations aside, the associations of functioning levels and mental retardation diagnoses to autism clusters are still very weak.
A Bearman paper claimed close proximity to a person with autism was associated with a 10% decreased risk of a mental retardation diagnosis – just barely statistically significant (meaning, likely attributable to more than just chance). The upper limit of that finding’s confidence interval was 1.00 – a 0% risk of acquiring a mental retardation diagnosis and living near a person with autism. Meanwhile, the positive association with developing autism and living near someone else with autism was much stronger and well within the range of statistical significance.
The high functioning autism claim was very strange; rather than simply divide up the groups into equal proportions – high-functioning, low-functioning and moderate – Bearman created a top 20%, bottom 20%, while putting the bulk of people with autism in the middle. The differences in both the high-functioning and low-functioning groups were not statistically significant when compared to the rest. And the 20% high-functioning group was barely statistically significant compared to the spacing risk in the 20%-lowfunctioning group, which was not statistically significant. Had these groups simply been equally divided into thirds as they should have been – rather than 20%, 60%, 20% - there would have been no statistically significant differences among the functional levels of autism.
And yet, it was on the basis of both of these supposed findings that Bearman dismissed environmental factors, despite having previously shown the audience a real-life example of an autism cluster in a neighborhood next to a construction site, which he acknowledged was an environmental omissions source for “metals.” The relationship between environmental exposures and autism clusters would also link to the correlations found with older parents who are more likely to have greater exposure to environmental insults, but not the “de novo” mutations Bearman asserted that older parents were causing – mutations that have yet to even be discovered.
Bearman sure did not follow Occam’s Razor – the principle that the more simple explanation is the correct one - in his analyses. The idea that “de novo” mutations behind the increase do not parallel at all with “social clustering,” unless he is making the less direct claim that parents are influencing each other to have children later. Environmental exposures, however, would link both of these directly.
Before I was able to ask Peter Bearman about any of these additional concerns, the question and answer session was ended - cut short by 10 minutes, even though there was no shortage of questions from the audience. Evidently, there was a little too much scientific skepticism about Bearman’s work to allot the full Q/A time the audience was promised.
One of the most amazing remarks Peter Bearman made during his lecture was his prediction that autism rates will go down and that it will be result of economic difficulties that will in turn reduce autism diagnoses. Yet, Bearman’s own research - the one graph he left out of his exhaustive presentation - showed an increasing rate of children getting diagnosed at earlier ages. If he knows something about the future of autism rates that the rest of the audience does not, this would surely seem to contradict his prediction of an autism decrease caused by economic crises, if a decrease were to happen. Certainly, the reduction of thimerosal from vaccines as a cause would not sit well with the man funding him - Tom Insel, or his brother Richard who co-developed a thimerosal-preserved vaccine.
Nonetheless, Bearman shares that special passion with the IOM and its members like NVAC Vaccine Safety Working Group co-chair Marie “"We are not ever going to come down that [autism] is a true side effect [of vaccines]" McCormick of explaining the unexplainable. Perhaps that’s why the CDC took notice and is now giving him additional funds to study assisted birth technology and its possible relationship to autism. If history is any indicator, it’s not science, but Peter Bearman’s opinions that will bring him more of the taxpayers’ millions.
Jake Crosby has Asperger Syndrome and is a contributing editor to Age of Autism. He is a 2011 graduate of Brandeis University with a BA in both History and Health: Science, Society and Policy. In August, he will attend The George Washington University School of Public Health and Health Services where he will study for an MPH in epidemiology.