[Randjuke]

I agree with your points, you must have been typing when I was reading. Meta-analyses are notoriously difficult to interpret and only as good as the individual studies that make them up (we are told nothing about the methods of the individual studies). To blame the difference in mortality on the shareholders expecting a profit is quite a leap to a conclusion as no study tested for that as a variable. As you say, the researchers most likely had a desired conclusion in mind before they undertook their analysis.

[B.R. Burton]

I didn't read the post, but I don't think 2 percent is enough to say definitively, either way.