to 
Since July, my team at @cmudelphi has been tracking the dropping case fatality rate, deciphering what share of the drop might owe to actual drops due to treatment (vs testing, younger infectees, etc). Drops of 55-60% appear plausible! Learn more here:https://delphi.cmu.edu/blog/2021/01/28/unpacking-the-drop-in-covid-19-case-fatality-rates/ …
-
In the HFR definition, shouldn’t we be looking at excessive deaths? And factor # of deaths caused by people who’s treatments or surgeries were delayed. Wouldn’t that give a truer representation? (Not relying on the # of reported cases; which have their own uncertainty.)
Maybe it’s a really stupid point. It was an issue in the U.K. gov reporting. It turned out the # of deaths were significant higher than what was being reported. Independent assessors were able to figure out a more accurate representation by looking at excess deaths.