Testing is the key to improvement.
When you order a test you have to first ask “Why?” And then, “How will the answer change management?” Testing could potentially give data but with testing there are “Alpha errors” and “Beta errors”. Death is a “hard fact”. There are no “false” positives or negatives. Sure, testing gives us a better idea of “where the virus is now” but the death rate is just a week or so behind that and the data is irrefutable.
The Roche announcement seemed to have the promise of greatly increasing the testing possible.
So i would argue testing doesn’t mean much and does little to help make decisions. But I have spent decades wrestling with the complexities of “testing asymptomatic people”. In reality the “helpful” part of “testing” is going to be at autopsies.