Actually, if you think detection can't come close to answering, you're probably mostly buying the antivaxxer accounts--and I'm being quite charitable there. Admittedly, a good part of it also comes from changing the diagnostic criteria, which is a Problem because as my dev psych professor brutally put it, it's the end of the goddamn bell curve--literally so, as a certain amount of autistic behavior is within the range of normal, which means that if you want to increase the number of people with autism the easiest way is to literally lower the threshold.
This folds into the fact that, frankly, the US school system has a perverse incentive to have as many students as possible diagnosed with a disability--they get money for it--and it's one of the easiest ones to game this way. (Yes, this qualifies as practicing medicine without a license and is harmful but since when did that ever stop the public schools?)
The least nasty player in all of this is that we've lost the perverse incentive to avoid diagnosis, as the stigma's decreased and, well, forced sterlizations & euthanasia is no longer anywhere near a problem as it was around 90 years ago. (Some of these court cases are still being settled though.)
Now, on the rest of it? If the clinic fails to check the health before giving the vaccine, GTFO. Vaccines should never be given to somebody not healthy. This is why I was delayed on one of mine when I had to get mine redone--I walked in and was bounced to the ER for a serious infection.
I don't know what you mean by the '17 vaccines,' but then I live in an area where the list is...about five, all of which are cheap and out of patent. I don't actually trust Gardasil, but that's because the way they did the challenge in testing it is absurd. (This is basically the gold standard, and is deliberately infecting somebody who has gotten the vaccine with what the vaccine is supposed to be against. Very few diseases are so bad as to warrant using a faux antigen--arguably, none are, as if they're dangerous enough to make it ethically doubtful to deliberately infect somebody who volunteered and knowingly consents to doing this test, it ought to be also unethical to sell it as a vaccine because it'll still get tested this way, just on people who did not knowingly volunteer to do so...and you can cover for it a lot more easily.)
That said, we actually do have some pretty strong proof that autism is probably a defect in brain development--last I checked we'd pushed back the earliest age at which we can diagnose it to before you should be getting any vaccines, namely six months. At least some cases can be traced to exposure to diseases at the prenatal stage, as the whole idea that the womb is a sterile, clean environment is hilariously wrong--let's take rubella, for example.
People exposed to it in the womb can end up with autism.
Ever asked what the R in MMR is for? I did, or rather I read the info sheet the clinic had to give me before accepting my consent to get it (again). It's rubella. Oh, and it turns out the vaccine is not for-life, regardless of what was previously thought; herd immunity just covered for a lot of people's immune systems 'forgetting,' sort of the equivalent of nobody noticing if antique computer viruses are no longer protected against by antivirus programs because there's little chance of somebody encountering an infected system anymore...