Defensive rating actually is quite significant. I don't see a situation where I'd ever consider either of your choices, though, while trying to build a winning team. It's more like:
efg% 58, D 50
efg% 50, D 90
That's a hard(er) decision. There's almost no situation where I take a tier 4 or higher usage guy with efg% less than 50% no matter what their DR is, and there's very few situations where I would take a guy with much less than 50 D. Progs are different because you rarely can get multiple seasons of every player to line up, so my standards drop a bit, but yeah, that's a rule I tend to stick with. If they are less than 50% efg%, they better be giving me a metric **** ton of something else that I need (like rebounds, assists, etc) and be lower usage. 92-93 Dennis Rodman is an excellent example.
So, back to your hypothetical example, I'm going to assume prog because I would never build a non-prog team where I'm considering either of those players. So, in a prog world, if those are honestly my best two options, I'm probably not building a winning squad that season and will look at someone else that has a better future while not caring what their stats are like this year - the worse, the better. If you're dead set between those two and are in a league where that is your typical player, then I would pick the guy who gave me better boards and/or assists (depending on team makeup/goal) and/or had the lower tov%. If everything else is equal... it doesn't matter because depending on either is a losing effort... but if usage is lower, I'll take the higher D guy. If the usage is higher, I'll take the higher efg% guy.