We never "use" an initial macro state, it would be way too difficult in any interpretation. The interpretation is around what we imagine
the initial state is. None of the interpretations involve usage, everything we actually use is the same in every interpretation and that's why they all get the same answers.
Decoherence can be used to show that a closed system evolving unitarily can project onto a density matrix of a subsystem that evolves, over time, to be diagonal. So we get that subsystem density matrices can be diagonal, so projections of pure states can be mixed states. But we never know the quantum state of the measuring device, so we simply don't know if it even has one-- this is purely a choice of our imagination to make. Decoherence allows us avoid contradiction when we imagine that macro systems have a quantum state, but it is not evidence that they do. And above all, it begs the key issue in the measurement problem-- how does a diagonal density matrix for the substate turn, unitarily or otherwise, into a definite measured outcome?
We can never find any initial state for the whole, if a measurement is involved. There is never any measurement that has a well defined initial state for the environment, that's why we need an interpretation of the environmental interaction. Copenhagen just says that part of the interaction in the measurement creates a "collapse" which need not involve unitary evolution of the entire system. It is close to the ensemble interpretation, in that neither asserts there is a unitarily evolving quantum state for the whole, but the ensemble interpretation does not take the "collapse" literally because the whole mathematical structure applies only to the ensemble, whereas Copenhagen suggests that something inherently non-deterministic is occuring.
I'd say there are two very different "measurement problems" that tend to get confused. One is, how does a unitarily evolving quantum state of the whole project into a diagonal density matrix for a subspace, and the other is, how does a diagonal density matrix turn into a definite outcome. The various interpretations hinge on the answer to the latter question, and I don't see any progress on that issue at all-- I see it as entirely a subjective choice for the philosopher/physicist. Decoherence has made interesting progress on the former question, but in my view that was always the easy question.