As many of the comments here have indicated, the problem with the numbers for these various categories (positives, recoveries, deaths, etc.) is a lack of consistency and enforcement in how they are gathered. If the source numbers are not reliable, the various rates extrapolated from them have to be considered unreliable. For example, at a given hospital, the methodology used to determine positives, or even deaths, can vary from day to day, or within a single day based on who is doing the counting in one shift versus another shift.
The sad fact is, we really have no handle on any of these numbers. The damage we've done to our economy, and all the associated increase in mental and physical health problems that have resulted from sheltering in place, were very likely unnecessary. Instead of imploding our economy, had we taken steps to protect those in the higher risk categories (yeah, I'm talking at you, murderer Andrew Cuomo!) and enforced some of the common sense social guidelines we're only now enforcing (in spite of a lack of "science" to prove their effectiveness) we might very well have already achieved herd immunity, or at least been well on our way toward achieving it.