First, a couple of bloggers seem to have figured out independently that contrary to some assertions, California Department of Developmental Services (CDDS) data do not support the idea that new cases of autism are decreasing in response to the removal of thimerosal from most childhood vaccines. Interverbal has done a very thorough analysis, including some spiffy graphs, showing that while the total number of cases managed by CDDS is growing at a reduced pace, among children aged 3-5 (the supposed beneficiaries of reduced thimerosal) autism cases managed by CDDS is continuing to grow. Joseph at Autism-Natural Variation also has a crack at the California numbers, and shows why the “thimerosal theory” for explaining increased diagnosis of autism is much less likely to be correct than the theory that increased diagnosis is due to broadening criteria.
Nice job fellas! But while you may have done it better, don’t forget who did it first. I just want to be loved. Is that so wrong?
Second, proponents of the thimerosal theory can now point to a peer-reviewed study in the Journal of American Physicians and Surgeons (JAPS) that says that autism rates are declining as a result of removal of thimerosal from vaccines. But what kind of peer review did this study really go through? How good is the evidence? Orac provides the answers. I highly recommend his respectful insolence. I won't rehash his arguments, but rather offer a few supplemental thoughts:
- Citizen Cain can reveal, after an exclusive investigation, that “peer review” in JAPS means approval by a panel consisting of Phyllis Schlafly, Randall Terry, Jack D. Ripper, and a chipmunk. For you kids out there, Jack D. Ripper is a character in the classic movie Dr. Strangelove, who drinks nothing but rainwater and grain alcohol so that his “precious bodily fluids” won’t be contaminated by fluoridated water. Anti-chipmunk prejudice may explain why JAPS has no scientific standing, and why it isn’t listed in Medline.
- This study, by a father-son team, David and Mark Geier evaluates two data sources to determine changes in autism incidence. One is the CDDS data. If only the Geiers read Citizen Cain, or even those Johnny-come-latelies Interverbal or Autism-Natural Variation, they could have spared themselves a lot of embarrassment.
- The second data source is the Vaccine Adverse Event Report System (VAERS). As you might have guessed from its title, VAERS isn’t a system for tracking autism or other neurological conditions; it’s a system for tracking adverse reactions to vaccines. Therefore, not surprisingly, very few reports of autism incidence are made to this reporting system. According to Figure 1 in the Geiers' article, between 1993 and 2005 nationwide reports of autism to VAERS ranged from zero per quarter to about 78, with most quarters receiving fewer than 40 reports. Since about 1 million little Americans are born every quarter, and since an estimated 1 out of 166 people are thought to be on the autistic spectrum, something on the order of 6000 autistic Americans are coming into the world every quarter. Therefore, VAERS is finding about 2/3 of a percent of autistics. Can you, dear reader, see it might not be a good idea to try to track trends in autism incidence through a system that can't account for even one percent of autistic children? Trends in autism reported to VAERS clearly have far more to do with changes in the fraction of autistic children who are reported to VAERS than with actual changes in the number of autistic children.
What does real science on autism incidence look like? Nothing like what the Geiers have excreted. Real science looks like this. A team of researchers evaluated cohorts of four to six year old children in 1987 and 2002, using "diagnostic interviews, psychometric tests, and medical workups." They found that despite reported increases in pervasive developmental orders, the actual prevalence found in these evaluations had not changed, "suggesting a stable incidence."