Its strange to me that they don't report COVID19 hospitalizations.
Deaths are the most reliable statistic, but they're a month late. By the time the death-rate is going up, the virus was already spreading for a month. #Cases are fast moving, but unreliable. If you don't have enough tests, then you literally can't count the number of people effectively. (And even here in my State, our %Positive went to 30%+ sometimes and our casecount went out-of-whack with reality. So I know for a fact that those case-counts can become unreliable during surges)
Hospitalizations are the happy medium between the two. As long as you have additional hospital beds, the hospitalization count is accurate. Hospitalizations aren't quite as delayed as death counts: they're only delayed by maybe 2 weeks or so (dependent on the culture of course).
You need "many eyes" into the problem, because every statistic has its own flaw. You can't just trust one statistic. If %Positive is low (5% or so), then rely on Case#. If %Positive goes above 5%, switch to hospitalizations. This strategy has been "back-tested" using death# (a month later, when the deaths show up, treat the death# as the ground truth).
Yeah, Death# still has some bits of controversy over. But its the closest thing to reality that tracks the pandemic.
