Who is actually doing this work? I have had an idea for a kind of 'public meta-analysis' site, where we could give simple headline figures for important questions like this based on consensus estimates. A bit like the Cochrane or Mental Elf, but for everything and producing a dataset of actual numbers to be used in arguments like this. It's so frustrating in comment threads like this that nobody actually has reasonable numbers for the fundamental assumptions of the arguments on either side.