That would apply if the numbers stayed steady through 1979 and beyond. That chart shows the aggravated autism-equivalent of of a cancer cluster. No one can write that off as a normal distribution.i have a background in quality engineering, every thing has a range or distribution, autism is no different.
I have worked in quality control, too.
If 1:1000 products failed quality checks (not counting ASD1), a company might live with that, but 1:200-250 would be alarming (and demand a resolution).
Preventing severe co-morbidities won't turn the person into an NT, just a healthier autistic.