People get sick, go to the hospital and die and are never confirmed as having Covid-19 because there aren't enough diagnostic tests. So it becomes a guessing game for the doctor who fills out the death certificate.
Though I can not speak for every hospital, I would feel confident in saying that this is not true and the number of people dying
in hospitals where they weren’t able to be tested would be infinitesimally small. Though there may not be as much testing available generally, those shortages will typically be diverted to other lower risk situations. The amount of resources consumed by a COVID+ patient or one under investigation is massive. As an example, one of the hospitals I work at where I do inpatient (not Utah), we had an outbreak of COVID-19 on our unit and it’s been a disaster. At some point I’ll post more about that, but not now. When testing first came out and was scarce, we had a couple people develop symptoms but without any known exposure and we tested everyone on the unit. Later, we did have a staff member who had a known exposure and retested and got positives. As a result of this, we’ve had to completely shut down the unit to any new admissions or discharges. We’ve had to isolate the positives, and retest the negatives. When one of the negatives then tested positive, we started this process all over again. So if someone is in the hospital, it’s imperative to know if they’re there with COVID, because if they’re not, they need to be off a COVID unit, and you can start to free up things like negative pressure rooms, and place them somewhere where staff can redirect that PPE to COVID+ patients.
I’d suspect the only possible exception to this is if local public health authorities in an epicenter diverted all testing away from places where there’s a high pre-test probability and moved those into areas with much lower pre-test probabilities as a containment strategy, but that’d be a very interesting, and controversial, strategy.
If people are dying at home, then that may be a different story.