I agree that the number of deaths is sufficiently small that you cannot really determine a trend. As we stand a single death occurring an hour after midnight that might have occurred an hour before is enough to substantially alter the analysis.

And that is a good thing, because it emphasises how deaths are.

Hospitalizations are, unfortunately, a different story. While the figures remain low compared to earlier dates, there is a clear upward trend that looks as if it is accelerating. It needs close monitoring.