I also suspect Vitamin A is similarly harmful in isolated form compared to the much more complex natural sources of it but don't really have any evidence on that.

I would bet against the specific prediction without rejecting the principle. That is, I expect when maximising a given function along the lines of "useful when you don't have enough and minimally harmful when you have more than enough" there is a specific isolated form that is better for supplementation than complex natural sources. Where my in principle agreement comes in is that I consider it highly unlikely that the actual isolated form used in an arbitrary multivitamin supplement to be the most desirable one.

The most obvious reason I would expect to find that at least one of the forms of Vitamin A to be better supplemented in isolation than just adding more of a combination is that some of the forms are largely inactive until they are converted by the suitable enzymes. That gives you a potential rate-limiting buffer in the case where you are already eating too much of the stuff and adding more would otherwise be bad.