I’m asking why a random sampling of hospitalizations for non Covid reasons would be significantly different than the general population with respect to likelihood of Covid infection.
For example, consider a hospital with 100 patients of which 90 are negative and 10 are positive. Of the 10 positive, 5 were admitted for covid and 5 for some other reason. Thus, 10% of all patients are positive, but the "TV" statement still holds.
That being said, the number of people who have had covid appears to be much larger than people seem to think. https://covidestim.org/us/MA estimates that 73% +- 20% of the MA population was infected at some point before Jan 8 2022. Like everything else, take all estimates with a grain of salt, but I have no reason to doubt those numbers.