I am hipwr223's (uh, didn't that used to be 308? hehe) brother in law, and I have a young son with autism. We got him into an ABA program about 2 months after his second birthday and he is now seven.
We've tried it all - speech therapy, physical therapy, diets, etc. (never swam with any frigging dolphins, though). We never saw any difference with any of it. ABA is effective, but the greatest factor for outcomes is the severity of the disorder in a particular child. Some will do well, others not as well, but all will benefit from an ABA program.
We've run home programs and started two schools, one in south florida and one in new jersey (with PCDI). A school is definately the way to go - if you have one, it will attract better teaching talent.
A word of caution: YOU CANNOT DEPEND UPON THE PUBLIC SCHOOL SYSTEM TO MAKE THE RIGHT CHOICES FOR YOUR CHILD. They will make the right choices for the public school system - by matching available resources and ability to the number of students. The vast majority of these programs are just well intentioned babysitting, run by special ed teachers who may know less about autism and it's treatment than you. A liberal arts degree is no qualifier for treating a neurological disorder.
A second word of caution. If you choose not to immunize your child, have you asked yourself the following question: What is more likely to happen - that my child will develop autism or similar disorders if immunized, or will suffer serious disease because they were not immunized? As a parent, you have a choice, but hopefully you can get the info to make an informed one. (look up Bernard Rimland on the net if you want more info).
I used to be a product manager at the pharmaceutical company that makes the MMR II vaccine (Merck) and have had the chance to discuss the vaccine issue with the researchers. When you look at the epidemeology of autism and it's occurance within the studied populations for the MMR II vaccine, there is no positive correlation between the incidence of autism and the vaccines. This does not mean that there is no link, just that a retrospective analysis of the data doesn't show an obvious one.
Personally, I believe there is a link because the incidence of autism in the population has increased. We need a prospective epidemeology study to see if it's really there or if the data is fooling us. We also need a large vaccine study with the incidence of developmental delay as an end point. I don't know of any companies that are currently doing this.
As for seratonin, Rimland's web site has some material on it. Also, there's a group that formed a few years ago that funds basic research for autism and has some projects working on neurotransmitters. A search of some popular med literature databases will show some studies where SSRIs (drugs like prozac, which allow the seratonin in the brain to last longer) were given to children with autism, but results have been inconclusive (and the study populations are very, very small).
Well, that's all the treatment I can talk for one night. My fingers are tired. See ya.
The Remnant