The claims for or against an autism epidemic simply cannot be proved given the evidence available...The article gives a good overview of the statistics and the difficulties of classifying children who fall somewhere in the autistic spectrum, given the shifting definitions of the disease.
The article also mentions the beginning of Cedillo v. Secretary of Health and Human Services, one of many cases in the US courts claiming that childhood vaccinations lead to autism. I have yet to see any direct support for causal claims in the scientific literature, although I have seen serious direct evidence against it. There are some indirect connections made regarding emerging symptoms and time of vaccinations, but those behavioral symptoms coincide with the time in life when those behaviors develop in children - rendering them visible - and that makes it impossible to draw any conclusions. Recent work has identified some behavioral markers of the disease that might be visible earlier in development, and if these become more reliably established, that could put this debate entirely to rest. I am not politically motivated to see either argument be true and I am open to being wrong, but I have yet to see the connection between autism and vaccinations compellingly made.
Many people want this to be true because they so fervently wish for an explanation, and I can hardly blame them. If the theory about the involvement of vaccinations is true it will be conclusively borne out by evidence over time that establishes high probabilities of connection. Conclusions of this sort cannot be drawn based on anecdotes, on indirect connections however compelling, or even on the evidence of a single good experiment. Knowledge is gained in the sciences from posing hypotheses and testing them and then having that test made again by someone else, and then having another lab ask a slight variation on the question and testing it again, and then altering the group of people slightly and testing it again. One experiment does not establish cause, the evidence accumulates. And autism makes it particularly difficult because the ever shifting nature of the disease gives us a moving target. It's interesting to see this debate move out of the lab and into the court. DSM- IV and ICD-10 (diagnostic manuals) provide their diagnostic criteria, but the symptomatology of autism is heterogeneous (manifests itself differently in different people) and its underlying neurobiological mechanisms (some marker that says you have this disease in the way that the presence of the polio virus in your blood positively establishes your diagnosis of polio) are not understood. Autism is not just a disease, it's a cause celebre - diagnoses are a part of our culture now. I'm not sure how fighting the cause of one explanation for autism in court (it is fairly widely accepted that there is not only 1 cause), is possible given the fact that we still don't know exactly what autism is. It's like having witnesses testify regarding a murder they didn't actually see. I don't know if judges decide or juries rule in this type of court, but will they have uncovered the truth at the conclusion of this case in a way that has eluded leading scientists researching autism for the last decade?
In this context, a brief mention of Uta Frith and Elisabeth Hill's excellent book Autism: Mind and Brain which gives a very good summary of current work of many of the leading scientists building an understanding of autism. I'll discuss some of the more compelling work in a later post, but its chapters are written by the scientists who are conducting this work and it is well written, and careful not to draw any massive conclusions. I'd say the book is for an audience who can read scientific articles of a moderate but not highly technical nature, it's not a layman's volume.