In a very real-world sense the "optimal dose" is 100% about finding hospitalization/health data. The "optimal dose" in phase II is about the microbiology and isn't very precise in terms of how it translates to real world effects.
But while a decrease in hospitalization is important, I disagree that it's the only thing that we should be optimizing against. I'm not sure what you mean by health data, but safety and immunogenicity (which might be upstream of hospitalization) are incredibly important to consider in the choice of dosage as well, and that's what optimal dose in the context of trials means. This is the context which I meant it in, and I'm just borrowing this terminology from other places, e.g.:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4944327/ https://pubmed.ncbi.nlm.nih.gov/29420118/
Once its approved already at a set dose, the "ideal dose" for future use is probably not the same as the "optimal dose" that maximized probability of approval.
"We propose a Bayesian utility approach (BUART) to randomized Phase II clinical trials which uses a first-order bivariate normal dynamic linear model for efficacy and safety in order to determine the optimal dose and study population in a subsequent Phase III clinical trial. "
> The important thing for the vaccine was too limit death and hospitalization. And phase II didn't tell us clear info on that.
Phase II is primarily to test that the "candidate is safe, sufficiently immunogenic, and maybe protective"[1]. It's not clear to me that testing hospitalization percentage is in scope.
[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4944327/