I suppose that some didn't underreport the death toll. To address your concern, I'm fairly confident that with proper application of statistics that a rough figure of deaths can be attained for different areas within a margin of error and compared with official numbers. I couldn't do it, but I'm sure others could estimate such a death toll from excess mortality figures and the like. Then again, you get into the issue of semantics and whether COVID was a secondary or primary cause of death and so on.