Pity the investigators at the CDC studying whether thimerosal, the mercury-containing preservative pilloried by the antivaccination movement as the cause of autism and everything that is evil in medicine. Three months ago, they published a high profile article in the New England Journal of Medicine entitled Early Thimerosal Exposure and Neuropsychological Outcomes at 7 to 10 Years, which, as had so many large studies before it, failed to find any correlation between thimerosal-containing vaccines (TCVs) and neuropsychological problems in children. True, it didn’t specifically look at autism (that study is scheduled to be published in 2008), but it was quite a comprehensive study that failed to find any correlation between TCVs and adverse outcomes, with the possible exception of a slightly increased incidence of facial tics. All in all, the results showed some measures of neuropsychological development slightly harmed and others slightly improved, an outcome entirely consistent with random noise.
As I recounted earlier, Sallie Bernard of SAFEMINDS, a key player in the mercury militia, was most displeased with this study. As a sop to autism activists who think that vaccines cause autism, she had been included on a panel of external consultants that provided input regarding the design the study. Clearly, her inclusion was an attempt to coopt her, so that she couldn’t com!lain about the results, having contributed to the study design. Such naÃ¯vetÃ©. Naturally Bernard decided to bolt when it became clear that the study was not producing the results for which she had hoped. The mercury mavens at A-CHAMP immediately released a press release criticizing it. Now she’s managed to get a letter published in the NEJM criticizing the study:
Thompson et al. (Sept. 27 issue) report the results of a study investigating the neuropsychological outcomes of early exposure to thimerosal. As a dissenting member of the panel of external consultants for this study, I object to the authors’ conclusion that there is no causal association between thimerosal and children’s brain function. The sample comprised children who were least likely to exhibit neuropsychological impairments. Specifically, children with congenital problems, those from multiple births, those of low birth weight, and those not living with their biological mother were excluded. The sample was skewed toward higher socioeconomic status and maternal education — factors that are associated with lower rates of neurobehavioral problems and higher intervention rates and that were not measured. The sampling frame included only children enrolled from birth in the health maintenance organization (HMO) and still enrolled after 7 to 10 years, excluding children in higher-mobility families, who tend to have lower academic and behavioral function. Children with neurobehavioral problems may have been less likely to remain with the HMO. Only 30% of families selected for recruitment participated, a low rate for scientific research. Among the families selected for recruitment, 26% refused to participate. Another 28% “could not be located,” which included families that did not respond to multiple recruitment attempts (internal documentation from the study contractor, Abt Associates) — another form of refusal.
I dealt with these complaints before. They’re virtually identical to the much longer complaint that was the A-CHAMP press release. Heck, the authors preempted pretty much every one of these complaints in the Discussion section of their paper! Nonetheless, they calmly respond to Ms. Bernard’s blather:
Bernard raises several points that we wish to address. First, children with low birth weight or serious congenital conditions or conditions developing in infancy that are known to be associated with an increased probability of neurodevelopmental problems were excluded from the study. It would have been difficult to distinguish the possible added influence of thimerosal exposure on neuropsychological deficits among such children. To do so, a larger study with a different design would be required.
Second, our sample was probably skewed toward higher socioeconomic status because participating families were members of HMOs in which coverage was provided by employers. In the study population, thimerosal exposure was associated with both maternal education and maternal IQ. We therefore controlled for socioeconomic factors, maternal education, and maternal IQ in the statistical analyses.
Third, our study was less likely to include highly mobile families because the participants had to have been enrolled in the same HMO during the first year of life and during the time of testing 7 to 10 years later. These criteria ensured that we had all immunization records during the first year of life, as well as access to the participants’ medical records during the time of testing. This enhanced the internal validity of our study but makes the results less generalizable to highly mobile families.
Finally, the 30% participation rate may have resulted in some unmeasured biases. Participation in the study required a substantial time commitment from mothers and their children. Although the 30% participation rate was relatively low, it was higher than we estimated when we planned the study. More discussion regarding participation and other issues can be found in the study technical reports, available on the Web site of the Centers for Disease Control and Prevention.
Bernard’s objection to the study being skewed to a higher socioeconomic status is particularly laughable, given that children of a higher socioeconomic status are more likely to receive the screening that leads to early diagnosis of neurodevelopmental disorders than poor children. Moreover, the reason for having excluded low birthweight children should be obvious and that Bernard even brought it up highlights her ignorance: Such children are more likely to have neurodevelopmental problems completely independent from any external cause, such as the substance under study (thimerosal). Including preemies and lower birth weight children would only contribute to the background noise and make finding true associations more, not less, difficult, contrary to what Bernard apparently thinks. Indeed, excluding such children was almost certainly done to make the study more likely to find an association between vaccines and neurodevelopmental disorders. As the authors respond, to look at that question would have required a larger study with a different design.
Prometheus pointed to Ms. Bernard’s complaints about the design of the NEJM study as an example of extreme arrogance. I agree 100% with him as he asks:
So, who’s the more arrogant; the person with a deep background in biology who states that the current data doesn’t support a connection between autism and mercury or the person with no background in science (apart from Google-based “research”) who insists that there is?
How about somebody with no background in science who presumes to lecture experienced medical researchers about the “limitations” and “flaws” of their study
I’m not infrequently accused of “arrogance” when I deconstruct nonscientific and dubious health care claims promulgated by advocates who really, really believe they work. I can’t help but laugh whenever I see such accusations, because, if there’s one thing that my education has taught me, it’s my limitations. For the most part, I know what I can do, and I know what I can’t do. I know what I understand well, and I know what I don’t understand that well (or don’t understand at all). The topics discussed on this blog are almost always the former, and when a moment of hubris leads me to foolishly venture into the latter, I usually point out that I don’t know that much about the topic at hand. I don’t claim to be perfect, nor do I claim never to exhibit arrogance, and in three years of regular blogging I’m sure I’ve spouted off ignorantly about more than a few things. However, looking back, I think my ratio of ignorant spew to informed discussion is actually surprisingly low, which is even more amazing to me given how much verbiage I’ve produced. I occasionally browse my archives, and I’m surprised at how rarely I come across something that I wrote that now profoundly embarrasses me. (And, no, I’m not going to give you an example here.)
What Prometheus reminds us (and what Mark Hoofnagle reminded us of a while back) is that one defining characteristic of a crank is a consistent and marked disconnect between how the crank perceives his knowledge of a topic compared to their true level of knowledge in which he consistently vastly overestimate his competence, coupled with an inability to recognize competence or incompetence. Moreover, such people are often tolerated or encouraged by our American sense of egalitarianism and anti-elitism. Ãber-crank Robert F. Kennedy, Jr. gave voice to this very sort of thinking when he dismissed complaints about mothers in the mercury militia:
But my experience with these women is inconsistent with those patronizing assessments. Over the past two years I’ve met or communicated with several hundred of these women. Instead of a desperate mob of irrational hysterics, I’ve found the anti-Thimerosal activists for the most part to be calm, grounded and extraordinarily patient. As a group, they are highly educated. Many of them are doctors, nurses, schoolteachers, pharmacists, psychologists, Ph.D.s and other professionals. Many of them approached the link skeptically and only through dispassionate and diligent investigation became convinced that Thimerosal-laced vaccines destroyed their children’s brains. As a group they have sat through hundreds of meetings and scientific conferences, and studied research papers and medical tests. They have networked with each other at meetings and on the Web. Along the way they have stoically endured the abuse routinely heaped upon them by the vaccine industry and public health authorities and casual dismissal by reporters and editors too lazy to do their jobs.
The problem, of course, is that it’s quite possible to be calm and rational–and completely wrong. Being “highly educated” is no guarantee that a person understands a specialized area of science outside her area of expertise or that she can self-educate about that science. Indeed, self-education is very difficult, particularly when those helping you to learn have an agenda that goes counter to current scientific understanding. Most don’t realize that you have to learn to crawl before you can learn to walk and that you have to learn to walk before you can run. There’s a reason it takes so many years of dedicated to master such complex disciplines to the point where it is possible to do credible research or even to understand the research that is done, and even that’s not always a guarantee, as the examples of Boyd Haley and Mark and David Geier demonstrate. Indeed, I would never have the hubris to believe that I could teach myself enough particle physics, for example, to argue particle physics with real particle physicists. Why on earth do so many people think that they can teach themselves enough medicine to argue credibly with real medical scientists? Unfortunately, nearly all cranks, be they antivaccinationists, “intelligent design” creationists, or other, are more like Sallie Bernard or Jenny McCarthy, who thinks that having attended the University of Google is enough to make her credible when she discusses autism science.