The hospitalization data is so bad at the moment from the combination of Omicron being so widespread and very few states separating the difference between hospitalizations from covid and hospitalizations with covid that it is really a useless stat at this point.
I'll even go a step farther. I bet hospitalization data shows the Omicron decline before regular testing data does since there appears to be less delay in the hospitalization data. In other words it is so widespread that when it starts to decrease the hospitalization data will actually serve as a better surveillance than the testing data.