Except 98% of cases are not yet resolved in the US (recovered/discharged/dead).
Out of 362 closed cases in the US - 237 died (65%). NOT saying this is the mortality rate. Not at all. Only an example to how ineffective calculating a mortality rate can be right now. Is using closed cases best? Unlikely, too little volume of data and mild cases have not been reported at all. Is using total open cases best? Unlikely, it takes weeks to resolve. There is no best answer here.
On one hand - there are only 372 closed cases out of 18,393 total. How many more will die before having a resolution? This thing can take something like 2 weeks to fully resolve. 59% of open cases have come only within the last 5 days.
On the other hand- how many infected were never tested and not included in the denominator at all? Likely a great number of people.
Last "what if" - how many died, but did not have an autopsy to reveal it was COVID. I'd say it is likely not many, jmo. But how many very poor or rural or stubborn individuals passed away at home without ever knowing the cause? It will happen as well.
At this point it just seems a little disingenuous for the twitter account to act like there is a solved mortality rate for the U.S., when most cases won't be resolved for weeks. Once you have a large number of cases that have ran their course, then maybe you can at least begin to extrapolate some range for the mortality rate.
NOT that a 1% mortality rate is anything to scoff at. Significantly worse than the common flu, by an order of magnitude of 1.