Despite years of research, thousands of researcher hours, and many promising discoveries, one thing that has not been found is a surefire way to have a child diagnosed with autism.  What have been identified, however, are a series of things that make it more likely a child will get autism. 

Many of these risk factors are genetic, though in no case has a single gene been found such that everyone who possesses it becomes autistic.  Some risk factors involve an infection of the mother while pregnant, or exposure to some chemicals before birth.  In rare instances, viruses caught long into childhood have resulted in people developing autistic behaviors.  Thus far, however, there is no gene, chemical, other agent , or combination of these have been shown to result in a diagnosis 100% of the time. 

Likewise, though there have been many physiological abnormalities found in autism; none are unique to the point where every person with autism has been found to have the same difference.  Many are not extremely common in autism (say 20%), but are still much, much more common in autism than in other populations. 

This speaks towards the complexity, and individual nature of autism and leads to one of the most important concept in biomedical treatments; the notion that what is most important is to learn how to treat your child depends on the specifics of that child.  Your mission autism sample size is one.  Trying to tailor medical treatments to a child without evaluation, based solely on a diagnosis of autism, is as ridiculous as basing a series of traditional therapies without first understanding the strengths and weaknesses of the particular child. 

What works for one child, may not work for another.  This means your child may not respond well to a particular supplement, dietary change, or other treatment that a friend's child responded wonderfully to.  Of course, the same can easily be said of traditional therapies like, ABA, or speech therapy.   The only way you can figure out what works for your sample of one is to perform evaluation, and ultimately in some cases, try something out and see what happens. 

This has important ramifications for several of the canards frequently assigned to the realm of biomedical treatments.