Goes number 1; Taylor number 42
Gallagher CM, Goodman MS, Hepatitis B vaccination of male neonates and autism diagnosis, NHIS 1997-2002. J Toxicol Environ Health A. 2010;73:(24):1665-77.
http://www.ncbi.nlm.nih.gov/pubmed/21058170
The central claim of the paper is that "Boys vaccinated as neonates had threefold greater odds for autism diagnosis compared to boys never vaccinated or vaccinated after the first month of life."
The autism group had 33 kids total. Of these, 9 of 31 (29%) were given the HepB vaccine. Compare this to 1,258 of 7,455 (17%) of the non-autism group who were given the HepB.
9 out of 31.
Are the red flags up yet? They should be.
Take for example kids aged 17 in the 1997 survey. When were they born? That’s right, 1980.
When was the Hepatitis B vaccine introduced? 1991. According to Mr. Kirby himself, the HepB vaccine didn’t get fully implemented until about 1996.
A lot of the kids were born before the “epidemic” of autism. No one disputes that the number of people identified with autism has gone up significantly in the last 30 years.
So, pretty much anything that changed in that time would “correlate” with autism.
Prometheus, a pseudonymous blogger who is a professor of biological sciences, noted:
I just did a couple of quick statistics on the data in this poster.
The first thing that pops up is that the prevalence of autism in this group is 1 in 238 (in the groups with HBV data, the prevalence was 1 in 241). Seems a bit low.
Secondly, there is no statistically significant difference between the two groups (HBV/no HBV) in their autism prevalence (p = 0.07).
The poster gives no details about how the data was “controlled for confounders”, so it is unclear how they might have “massaged” the data into significance. No matter – with no statistically significant difference between the groups, the odds ratio is meaningless.
the authors found after their regression that the odds ratio for autism for non-Hispanic whites was 0.385.
This is polar opposite to the administrative data (which usually shows very high “odds” for whites) and against what actual prevalence studies show (no variation by race/ethnicity).
Something is very odd with that study.
In January 2011, the paper again came up for discussion in the comments at Science-Based Medicine. One of the regular commenters wrote:
The key, of course, is that the premise of the article rests on the danger of thimerosal, so there are immediate problems with the mechanism by which the authors suppose the hep B vaccine is linked to autism — and predictably, in the discussion section, the authors cite the ludicrous work of the Geiers as supporting evidence for their findings. The data are drawn from the NHIH — so it’s interview data with no independent confirmation of autism diagnoses, and there appears to be a lot of relevant information missing. For example, although they found about 193 cases of autism in the sample they used, only ~45 of those actually had usable vaccination records.
So far, the study looks more than weak.
I looked for papers published after 2010 that might replicate the results; there are none. However, there is a 2011 Polish paper that failed to find an association between thimerosal exposure from vaccines and autism.
Does this paper "demonstrate that vaccines can cause autism"? No. Not even the Hep B vaccine.
Comments