I'd reckon it's "doable" but not "easy", putting it in the realm of professional epidemiology. People are also dying less in car crashes, pollution-induced asth

I'd reckon it's "doable" but not "easy", putting it in the realm of professional epidemiology. People are also dying less in car crashes, pollution-induced asthma, etc. There's quite a bit that may need to be accounted for. I suspect analyses will have decent error bars a year or so from now, but it's unlikely we'll get precise realtime estimates from this methodology. Would love to be proved wrong.