Follow-up studies
Given the tremendous potential public-health significance of a link between the MMR vaccine, a number of scientists immediately followed Dr. Wakefield’s suggestion to make a “further investigation” of the purported link 
between the MMR vaccine and autism. Let me list and describe a few of the more prominent papers:
- In 1999, just a year after the initial Wakefield paper, The Lancet published an article from Dr. Brent Taylor and 6 co-authors which examined a much larger group of patients: 498 individuals diagnosed with autism drawn from a special-needs register in the UK. Although they found a steady increase in the incidence of autism in the overall population, they found no sudden increase in the rate of autism after the introduction of the MMR vaccine in 1988. They also reported that there was “no difference in age at diagnosis between the cases vaccinated before or after 18 months of age and those never vaccinated” and that there was “no temporal association between onset of autism within 1 or 2 years after vaccination with MMR.”
- In 2001, three years after the Wakefield study, an article was published in the New England Journal of Medicine on a huge group of 537,303 children from Denmark. After adjusting for factors such as gender and birth weight, the relative risk of a diagnosis of autism for vaccinated children was 8% lower than for unvaccinated children. Similarly, the adjusted relative risk for a diagnosis of autism-spectrum disorders for vaccinated children was 17% lower than for unvaccinated children. Additionally, there was no significant clustering of autism diagnoses in any particular interval after vaccination.
- In 2008, another small study which attempted to replicate Dr. Wakefield’s original study was published in PLoS ONE. In it, 25 children with autism and gastrointestinal disorders were compared to a group of children with gastrointestinal disorders alone. The researchers found that “[gastrointestinal] symptom and autism onset were unrelated to MMR timing [i.e. when the MMR vaccine was administered]” stating that “This study provides strong evidence against association of autism with … MMR exposure.”
- In 2014, an analysis of all the published scientific evidence collected so far on the association of vaccines with autism was published by researchers at the University of Sydney, analyzing data from 10 studies which included over 1 million individuals. They concluded that “There was no relationship between vaccination and autism”, “There was no relationship between vaccination and ASD (autism spectrum disorder), and “there was no relationship between [autism/ASD] and MMR.”
These are only a few of the more prominent studies on this subject. A more extensive list can be found here or here. These studies found no evidence that there is a link between the MMR vaccine and autism and provide positive evidence that withholding the MMR vaccine from children has no effect on autism rates.
Criticism of the original study
So what accounts for the contradictory results of Dr. Wakefield’s original study and the follow-up studies I’ve discussed? A number of scientists have asked that question. Two scientific objections seem the most relevant to me. First, the original study included an incredibly small number of children– only 12. No matter how compelling the results, medical and scientific studies need to be repeated on large groups to ensure that observations are not due to coincidence. The discrepancies between Dr. Wakefield’s 1998 study and later studies could be due in part to the far larger numbers considered in the later studies, which were simply less subject to statistical error. Second, the first study relied heavily on self-reporting from parents regarding the timing of the onset of behavioral regression. But this is an unreliable guide. For example, if you are a parent, can you remember at what age each of your children caught their first cold? Had their first fever? First threw up? I’d imagine that my own guess might be accurate only to within a few months; it certainly would not be accurate to within a few days or weeks. Consequently, it’s questionable whether the timing between the MMR vaccination and the onset of behavioral symptoms could be ascertained accurately merely by asking the parents. At best, we’d have only their best guess.
Another possibility is, of course, that the follow-up studies were wrong. The primary difficulty with this assertion is that the methodology employed in other studies was far more rigorous than Dr. Wakefield’s. First, they included far more children, so if we doubt the statistical significance of their results, it seems we would have to doubt Dr. Wakefield’s results even more. Second, they relied on far more objective measures of reporting: clinical diagnoses of autism by professionals, vaccination dates, government records, etc. While these methods are also subject to error, they seem less susceptible to error than simple recollection, often many months or years after the fact. A more common explanation is that the follow-up studies were all influenced, either consciously or unconsciously, by the pharmaceutical industry. Given that pharmaceuticals are a highly lucrative business, scientists have a financial incentive to produce results that exonerate large pharmaceutical companies, the same companies which award many of the research grants that fund these same medical researchers. As I said at the outset, I’m not denying that this is a possibility. Bias is real and scientists are no less susceptible to temptation than any of us. But I also stated that I would not assume bias at the outset, prior to considering the evidence. Now that we’ve discussed the evidence, is there a possibility of a comprehensive, systematic bias?
Neil Shenvi - Apologetics 