Correlation studies are fine, but I do worry about national GDP as a confounding factor.

Maybe poorer countries have lower life expectancies, (or whatever PYLL measures is higher), and work longer hours.

I would like to see this graph broken down by quantiles of national GDP.