I don’t always blog about stories or studies that interest me right away. Part of the reason is something I’ve learned over the last eight years of blogging, namely that, while it’s great to be the firstest with the mostest, I’d rather be the blogger with the mostest than the firstest. I’ve learned this from occasionally painful experience, although I’d be lying if I didn’t admit that in part this is a rationalization for the fact that I have a demanding day job that keeps me from jumping all over stories and studies of interest in the way that some bloggers can. There’s also the simple fact that blogging is a global phenomenon, and even if blogging were my full time activity there would always be bloggers in time zones many hours ahead of me who can pounce all over science and skepticism stories that I want to pontificate about. (I’m talking about mainly you, U.K. bloggers. Damn you for being awake while I’m snoozing the night away.)
It’s also sometimes good to let a story percolate a day or two, anyway. It can let me put it into context and, just as importantly for purposes of entertainment value, survey the reaction of the quacks, cranks, and pseudoscientists when appropriate, not to mention give me time to look up the actual study. So it was that I didn’t leap all over the stories that began peppering the media yesterday about how the prevalence of autism is now estimated to be 1 in 50 among U.S. children between the ages of 6 and 17, as reported by the AP and elsewhere. Not surprisingly, the antivaccine cranks at the antivaccine propaganda blog Age of Autism are all over it, with Anne Dachel laying down her usual burning stupid, while others demand that Thomas Insel be fired and Teresa Conrick engages in her usual scientific incompetence born of the arrogance of ignorance and tries to link various observations in science by trying to link the recent report that 1 in 3 adults will die with dementia with the report that 1 in 50 children have autism or autism spectrum disorder. (Yes, vaccine are to blame for both, in adults the influenza vaccine and in children childhood vaccines.)
To understand the importance of the issue of the prevalence of autism and autism spectrum disorders (ASDs) to the antivaccine movement (as opposed to the rational, science- and reality-based community), you have to understand the central dogma of the antivaccine movement. That dogma is that the reason for the massive increase in autism prevalence over the last two or three decades must be something in the environment. Of course, antivaccine cranks being antivaccine cranks, to them there is only one thing that could be causing it, and that’s vaccines. The reason is that the beginning of the increase in autism prevalence just happened to correlate with an expansion of the vaccine schedule. In a massive case of confusing correlation with causation, antivaccine activists, against all existing reliable scientific evidence from well-designed epidemiological studies, insisted (and continue to insist against that evidence) that it must be vaccines that are causing or contributing to what they like to refer to as the “autism epidemic” or, when they’re in a cruder mood, the “autism tsunami.”
The retort to such an obvious logical and scientific fallacy is to point out other things whose rise corresponds to the increase in autism diagnoses. One example I like to use is Internet use. It exploded beginning in the early 1990s and continues to rise today. Then there’s the example of the humble CD. Introduced in 1985 in the U.S., its use skyrocketed for 20 years, although admittedly CD sales are plummeting now as CDs are being supplanted by downloaded MP3 files, as CDs supplanted LPs; so maybe that’s not the best example anymore. However, perhaps the best correlation I’ve found thus far is between organic food sales and autism. Obviously, organic food must cause autism!
In any case, the heart of the antivaccine religion is the dogma that autism prevalence is rising and that the rise is caused by vaccines. Never mind that there are many other factors that cast doubt on the idea that the true prevalence of autism is actually rising, including diagnostic substitution, increased awareness, increased screening, and increased services. The example that I like to use to illustrate this point is worth bringing up again. There is a form of breast cancer known as ductal carcinoma in situ (DCIS). Well, actually, whether it’s really cancer or not is debatable, but it is clearly a precursor to cancer, although the percentage of DCIS lesions that progress to cancer isn’t precisely known. Be that as it may, before 1975 DCIS was a very uncommon diagnosis. Now it is very common, its incident has risen by 16-fold. No one believes that the actual incidence of DCIS has risen by that much. In fact, it’s unlikely that it’s actually risen much at all, but we are detecting much more of it because of the advent of mammography screening programs in the late 1970s and early 1980s. Yes, I know I’m mixing incidence and prevalence, but the example still illustrates a general principle that if you look for a disease or condition intensively, you will always find more of it, often a lot more of it. Always. And if the principle works for something that is diagnosed by an objective test, namely a biopsy, how much more so is it for a condition that has no unequivocal biochemical or tissue test to nail down the diagnosis, like autism, particularly for something whose diagnostic criteria changed considerably 20 years ago to widen the diagnostic criteria?
I hope that puts this report into context. Yes, the apparent prevalence of autism has been reported to be 1 in 50, which is in line with a South Korean study that found it to be 1 in 38. At this point, it is useful to bring up another principle. If you screen intensively for a condition, after an initial rise in incidence and prevalence, you will eventually see a leveling off at something near the “true prevalence” of the condition, and this is what could well be happening. After all, contrary to the ridiculous claim of Julian Whitaker, autism prevalence can’t keep increasing forever until it reaches 100%. It could be that the baseline prevalence of autism and ASDs is somewhere around 1 in 50. Only time will tell whether this is true or not, but it seems reasonable based on what we know now.
So what about the report? Basically, the authors mined newly released data from the 2011–2012 National Survey of Children’s Health (NSCH). This is a telephone survey of 95,677 families. One problem with the study, as has been noted in multiple news stories, is that the response rate was low (23%). While the authors describe going to great lengths to determine whether this biased the results through a phenomenon known as nonresponse bias, I must admit that I’m not entirely convinced. It’s quite possible that parents with children with ASD would be more interested in responding to this survey than parents with neurotypical children. On the other hand, this is not a survey about autism; it’s designed to look many health conditions.
If we accept the reliability of the survey instrument and the authors’ methods, the conclusion of this study was that autism prevalence has increased from 1.16% (1 in 86) in the 2007-2008 survey to 2% in the 2011-2012 survey in children aged 6 to 17. The authors noted that this increase in prevalence was observed across all of the age ranges studied. Moreover, they noted that much of the increase in prevalence was driven by diagnoses made post-2008 of milder cases of ASD, indicating a trend towards less severe presentations since 2008. Also noted, probably not surprisingly, is that the greatest increases occurred in boys, for which prevalence nearly doubled (2007 prevalence: 1.8%; 2011-2012 prevalence: 3.23%) while in girls the increase was less dramatic (2007 prevalence: 0.49%; 2011-2012 prevalence: 0.70%). All of this, the further increase in prevalence coupled with the shift towards less severe presentations, is reminiscent of the story of DCIS and suggests to me, more than anything else, a probable screening effect responsible for the increase rather than a true increased prevalence. Or, as described in the AP story:
“I don’t see any evidence that there’s a true increase in the prevalence of autism,” said Roy Richard Grinker, a professor of anthropology at George Washington University, Washington, D.C.
Grinker said he’s been anticipating a higher count in the United States since he published a 2011 study that found an autism rate of 1 in 38 in South Korean children. “I don’t look at that and say ‘that’s so much higher than the U.S.’ I look at that as ‘the U.S. will catch up.'”
The new study found the biggest jump among older children with milder symptoms, suggesting that their autism wasn’t caught until later in childhood. By definition, symptoms of autism must be present by age 3, affect a child’s communication and social skills, and lead to restricted or repetitive behaviors such as rocking or hand-flapping.
Michael Rosanoff, associate director of research for the advocacy group Autism Speaks, said he thinks the new numbers reflect improved awareness of the condition over the past decade, leading to more diagnoses. Because these children weren’t counted in earlier studies looking at school district support, it also suggests that many children who need help with their symptoms aren’t getting it, Rosanoff said.
It is the children with milder symptoms who are most likely to be affected by a change in the definition of autism that will take place this spring. Rosanoff said this study adds urgency to the need to protect those children.
While the AP quotes actual experts, one can’t help but note the marked contrast with the antivaccine crank blog Age of Autism, which cites not experts but one of the “Thinking Moms” Alison MacNeil, who is the coauthor of a an upcoming book The Thinking Moms’ Revolution: Autism Beyond The Spectrum. MacNeil, not surprisingly is bristling with the arrogance of ignorance. Responding to the CDC’s assessment that the reason for the increase is better diagnosis and screening, MacNeil is having none of it:
I don’t think the public is buying that anymore. They might have a while back, especially when the criteria shifted in the Diagnostic and Statistical Manual, but I think now, what we’re seeing everywhere we go, we’re seeing autistic children.”
“Why do we have so many children descending into autism? What is going on? One in every 50 children is losing the ability to speak, becoming incredibly sick, detaching. There’s something behind this.
Gee, I wonder what MacNeil is referring to when she says there’s “something” behind this. Could it be…
Then there’s the AoA crew demanding Thomas Insel’s head because he doesn’t accept their belief that vaccines cause autism. True, it’s in antivaccine code. To understand the code, you have to realize that whenever an antivaccinationist says “take the autism epidemic seriously and do something” they mean to start believing that vaccines and dubious environmental causes must be the cause of the “autism epidemic.” Anything short of that, particularly accepting the science showing that there is no detectable epidemiological link between vaccines or mercury in vaccines and an increased risk of autism, is heresy.
Meanwhile Teresa Conrick does what she does best and demonstrates her ability to make irrelevant connections between conditions that do not share pathophysiology, obtained mainly by cherry picking studies. In this case, she tries to liken Alzheimer’s disease to autism:
Coincidentally, about 30-40 percent (one out of three) adults receives an influenza vaccine. Is that a coincidence or is it a clue? Since both Autism and Alzheimers are each frequently quoted as being ” A MYSTERY,” is there a pattern to their well kept secrets? Alarmingly, there is much evidence that mercury, and Thimerosal, the kind of mercury in most flu shots, can cause immune and autoimmune issues . Is there a connection to immune issues and Alzheimers? We know Autism has numerous connections to the immune system, with many children and young adults also receiving an autoimmune diagnosis.
Because autism is just like Alzheimer’s and, of course, the evil vaccines cause them both! That reminds me. Perhaps I should take on a previous post of hers to which she links as “evidence” fot the connection. It’s a perfect example of how not to make inferences based on the scientific literature.
In the meantime, I can’t resist concluding with a little tweak to an old “friend” of mine. Two weeks ago, I gave a talk at the National Capital Area Skeptics (NCAS). The talk went well (at least as far as I can tell), and a good time was had by all, including, of course, myself. Some of you were even there, and I thank you for coming! Even though the talk wasn’t about vaccines, that didn’t stop our old buddy Jake Crosby from doing what Jake does and showing up at my talk, no doubt to try to goad me into saying or doing something embarrassing or, failing that, to get himself ejected and then paint himself as some sort of free speech martyr exposing the evil pharma conspiracy. I spotted him right away, lurking in the back of the room. At the end of the talk, not surprisingly he had to ask a question, and his question was apparently based on a post of his from a couple of months ago that I completely missed in which he tried to paint me as a liar because several years ago I said I’d reconsider my rejection of the hypothesis that the mercury containing thimerosal preservative in vaccines causes autism if autism prevalence started plummeting. That was back in the early years after the removal of thimerosal from childhood vaccines, and I proposed this:
I propose as quite a reasonable measure that, if autism rates fall by 50% or more in 2010 or even 2015, I will happily admit that I was incorrect in my assessment and rejoice that such a blow has been struck against this condition. If rates fall by less than 50% but still inarguably statistically significant, I will concede that this would be pretty good epidemiological evidence that there might be a connection, although in that case the connecton would clearly not be nearly as strong as the link claimed by some activists, like J.B. Handley, founder of Generation Rescue, whose website states quite bluntly that “childhood neurological disorders such as autism, Asperger’s, ADHD/ADD, speech delay, sensory integration disorder, and many other developmental delays are all misdiagnoses for mercury poisoning
In retrospect, I think I was way too generous, but there it is. Now, Jake seems to think that I am a liar because he found numbers showing a 40% decline in ASDs among African-American children in Alabama that is almost certainly due to problems in case ascertainment rather than a true decrease. Based on that, he thought I should retract what I said nearly eight years ago. I, of course, do not. Indeed, I repeatedly asked Jake what’s happening to autism prevalence everywhere else, which produced the amusing spectacle of him trying (and failing) to handwave and Gish gallup. I also note that a month before my talk, Jake’s fellow AoA blogger Anne Dachel lamenting that Autism is overwhelming Alabama.
Then this study shows up, suggesting that autism/ASD prevalence is still going up 11 or 12 years after thimerosal was removed from childhood vaccines and may be as high as 1 in 50. I am amused. Too bad this study wasn’t reported two weeks ago. It would have been so much fun to rub Jake’s face in it publicly. I’ll just have to settle for doing so now.
I know it might have been excessively snarky given the circumstances, but later on while trying to question me more after my talk, Jake asked me if I knew who he was. My response was along the lines of, “Of course, I know who you are. That’s how I know you don’t know what you’re talking about.” His only response to that was to tell me I was lying, after which I was done with him. The same can be said of all the antivaccinationists trying to use the increasing prevalence of autism as evidence that it must be those horrible vaccines that have caused it. The problem is, their message is so dangerous that I can’t just disengage, as I did that day, by saying, “We’re done,” and walking away.