It's by far not the worst case number. To get the worst case number, we need to know these variables: % of your population that eventually gets infected * % of

It's by far not the worst case number. To get the worst case number, we need to know these variables: % of your population that eventually gets infected * % of all cases requiring hospitalization * % who get no medical attention * Death Rate for people who get no medical attention OR ~0.5 - 1%, whichever is higher. We can start putting error bars around these figures: Best estimates from China/SK & Diamond Princess all indicate that ~20 - 30% of cases at most are asymptomatic and if ~20% of all symptomatic cases require hospitalization, that means the lower limit is ~15% of all cases require hospitalization. If you go with the "herd immunity" approach of 70% of people eventually getting it, then ~10% of all people require hospitalization which is 10 - 1000x overwhelming of the hospital system, no matter how much you "flatten the curve" so 90+% of people will just flat out receive no medical intervention. The only other key number left in the equation is what % of hospitalized cases, if given no attention, will eventually die and the data from Italy seems pretty grim. Italy in a few days, will already have more deaths than ICU beds which means the untreated death rate might be as high as 50% vs 5% for treated. Multiply all of these numbers together and it's not hard to see how 4.3% might end up being a conservative estimate of the death rate. It also becomes obvious that the ONLY variable that matters in this equation is the % of people eventually infected. If you can keep it below 1% forever, you're in a reality far different from if it goes above 30%.