The mechanism is that people with the shingles vaccine are less likely to visit the hospital (because they don't get shingles). Because they have fewer hospital visits they are less likely to receive an incidental diagnosis of dementia from a hospital.
There's countless treatments and countless diseases. It is very much worth combing data to find treatments with potential off-label uses, but with that many combinations of treatments and diseases, much more care needs to go into eliminating not just non-causative correlations, but straight-up random correlations that have a very small probability of happening on their own, but are likely to happen in a large enough group of comparisons.