I am in the beginning stages of relativity and space-time, but for distances between galaxies we determine this and cross check this with a few different measurement, i.e. Type 1a SN, Cepheids, Tully-Fisher, etc. so "space-time" isn't really needed. As for the question to "when" they are I don't think is quantifiable because when dealing in SR & GR there is no absolute time, it is relative to the observers refrence frame. It is true that when we observe say Andromeda, it's photons are from 2.5 MLY ago and since then it's location has indeed changed, but this does not mean we cannot know "where" it is in the present time. What we can do is we can infer it's "new" location first by determing its velocity via doppler shifts in its spectrum, and then combining that with time to solve for the distance. There are a few other aspects to it such as the radial & tangential components of the galaxy in question. Some one correct me if I am wrong. I hope I haven't sent the OP in the wrong direction.
Joe