yea problem with that data is the post hoc procter hoc fallacy which can be a big concern on indirect data like that. For example once could explain the data by saying more unvaccinated people are careless about the virus since they dont think its deadly (thus why they arent vaccinated) and thus are a much higher incident simply because they social distance less and arent as mindful.
Thats why newer direct evidence tests, like measuring viral load directly, is far more telling.