Tupelo3, I agree with every word you write bar one, the word "too" as in:
"Too many people act on correlated studies ..."

The word "too" implies a comparison with some available alternative, as in the sentence: "Too many people drowned when the ship sank because the lifeboats were locked." The problem for PwP is that there are no lifeboats that will take us to safety, though there are plenty of lifejackets that stop us from drowning for a long time.

I am reminded of a story, it may be apocryphal, about a mountaineer climbing in the Himalayas. When night arrived he found himself above 8000 metres. He reckoned that at this altitude and with such low temperatures he would probably die if he stayed here. He noticed that an icy slope led down to his camp. Unfortunately, this slope was too steep to climb down. He could slide down it, but he might hit rocks before he got to the bottom, which would surely kill him. What would you do if you were in his predicament?

The point I'm making is that in some circumstances making a decision on the basis of partial information is rational.

This is not to say that every factor correlated to Parkinson's should be explored willy nilly. At the least, there should be a high expectation that the therapy is safe and does no harm. As far as I know, this is the case with blueberries.

The challenge comes in trying to combine weak statistical data with weak causal data in an as effective a way as possible.

John