Shouldn't mortality rate be determined by total dead out of total recovered? Doing it the other way, total dead out of total infected, assumes everyone infected will recover.
Bottom line is it's hard to get good estimates and we should let epidemiologists do their job.