More “proof” that vaccines cause autism? Seriously. This paper sucks.
Today I would like to walk you through the latest steaming turd of a “study” being bandied about by the anti-vaxxers, which they tout as the holy grail: an allegedly peer-reviewed article showing that vaccines unquestionably cause autism. The paper is written by one John B. Classen, and is such a horrific pile of non-science that it is hard to know where to begin a critique. But try, I shall. In the meantime, one cannot help but wonder if the “peer review” were conducted by the author’s mother or perhaps a 7-year-old child. Regardless, below is a list of reasons why this paper cannot be used for anything other than parakeet cage lining.
1. The introduction omits any information contrary to Classen’s hypothesis, which appears to be that vaccines cause both diabetes and autism.
Classen’s introduction makes the rather bold assertion that vaccines may cause diabetes. This line of thinking is allegedly supported using references from the author of the paper, and ignores several studies showing no increased risk for Type 1 diabetes whatsoever with vaccination. In fact, a ginormous study conducted by Duderstadt et al. among over 2 million individuals, a significantly reduced risk of Type I diabetes was observed with several vaccinations, including hepatitis B, MMR, and yellow fever vaccines. Another study by Blom et al. has also shown that having received an MMR vaccination was associated with a lower risk for Type I diabetes.
In addition to the published literature, Johns Hopkins convened a panel to address the very issue of whether vaccines might cause Type I diabetes. Their conclusion? They do not. As an added bonus, Classen actually participated in this meeting, though the panel chided him for his over-interpretation of ecological data, as well as stating that, for at least one of his presentations “[Classen’s] analytic methods were incorrect and a careful analysis of data from 10 years of follow-up has revealed no significant differences in the incidence of type 1 diabetes mellitus in children who received one vs. four doses of Hib vaccine”. The evidence from Classen’s latest paper demonstrates that he again appears to fall into the same trap of misinterpretation and flawed data analysis. In short, Classen’s introduction is one big epidemiological no-no. He apparently ignores data that do not support his hypotheses, he relies almost exclusively on his own work, and he continues to do the very things to which the Hopkins panel rightfully objected. Red flags, I see you waving.
2. The methods are a disaster.
It appears that Classen attempted to perform a dubious ecological study comparing prevalence rates of Type I and Type II diabetes and autism, by race. In short, he copied data from this JAMA paper for rates of diabetes by race, and then added on another column of data from this paper for autism prevalence among children aged 8 years, by race. From here, things begin to degenerate quickly, as can be handily ascertained because these data are quite easy to check. I will use the correlations between Type 1 diabetes and autism as an example. If you conduct the analysis by age group, examining the association between Type I diabetes and autism rates by race, there are no statistically significant findings—not even close. The magic p-values Classen presents are only obtained when you perform a sleight-of-hand whereby the Type I diabetes data from all age groups are counted together in one lump, and the data for autism at age 8 are counted three separate times for comparison. This might be considered to be intellectually dishonest at worst, and statistically inept at best. But regardless of the faux findings, it wouldn’t matter even if they were legit, which leads us to the third fatal flaw with this paper. Remember, all that these “analyses” fail to show is that two diseases—Type I diabetes and autism—tend to be higher in certain racial/ethnic groups when compared to others.
3. The conclusions are not supported by the data.
This may be the worst criticism that one scientist can fling at another. But in this case, I’m afraid it must be said, particularly as Classen himself admits that his conclusions are based in part on stuff that was not even included in the publication. The final sentence in this paper is as follows:
It is the belief of the author, based in part on the data present in this manuscript that the epidemics of Type 1 diabetes and autoimmune autism are more likely than not to share the same etiological cause.
That cause, according to Classen, is vaccination. Perhaps the most striking aspect of this outrageous conclusion is that Classen did not include vaccines as a variable at all in his deeply flawed analysis. Yet somehow, he “believes” that these data support the idea that because certain diseases are diagnosed more often in one racial group than another, it can only be caused by vaccination. He ignores a plethora of other potential confounding factors, which are actually too numerous to be listed in full, but may include socioeconomic status, access to healthcare, or genetic predispositions.
4. The author has a clear conflict of interest.
Oh yes. The author is the CEO of Classen Immunotherapies, Inc., which, to his credit, he discloses the paper. But we can’t quite get away from the fact that on his company webpage, it states that Classen Immunotherapies, Inc. is
dedicated to understanding the chronic effects of vaccines and developing safer immunization methods for those who choose to be immunized.
Classen also peddles his patented “adverse event discovery and disclosure method” on his website. Thus, we have a gentleman who potentially stands to profit directly from sowing fear of vaccinations. This type of person generally stimulates cries of “Pharma Shill!” from the anti-vaxxers, yet for some reason he has been accepted into the warm, yet stupid, embrace of the anti-vax community. Curious indeed.
Thanks for stopping by, Reader.