The reason it moved up is pretty simple. The calculation was: deaths to date / confirmed cases to date. As the pandemic spreads, more people will be infected and still sick, and as long as that is still on the exponential portion of the sigmoid curve, the portion of those who are sick/dead/recovered will remain roughly constant. This hides the optimistic assumption that everyone sick will recover. That was 3.5% or so before, now it is higher. And it changes place to place.
Prior to last week, we were only testing people who had definite coronavirus symptoms, AND were in a high risk group. So the death rate in that testing set would logically be HIGHER than it would be for a larger population which includes all people.
In other words, we were only testing people with the highest likelihood of death. AND one of our initial vectors found its way into a nursing home with his preponderance of high risk patients. AND our mortality rate was the same as the worldwide average.
So, as we now start testing people who are symptomatic that have a higher likelihood of recovery, then obviously the mortality rate should be expected to drop.
So all our rates are bogus.
No, they are accurate reflections of where we are at a particular moment in time. They only become bogus if somebody tries to assume that they will accurately predict the final outcomes, which, at this point, could still be higher or lower than where we are today.
Speaking of where we are today (ignore the 22nd, since the days’ data is not entirely in yet.)
Anyway, the new cases for both the US and the world are pretty similar. Trend line is frightningly up.
However, mortality is another story. Worldwide mortality is rising, ours is falling.
I find this interesting. YMMV.