They are all huge broken OO designs! At least, that is my experience.
The original architect considers the domain - usually poorly since he/she is only human - and says "lets have an OO structure that looks like this". And with the first enhancement request the OO design breaks because the designer thought the world worked in a different way and enshrined that thinking in to the OOD.
OO almost always turns out to be an impediment. I once inherited a project that had been a poster-child for OOD, where the architect had innovated with "design patterns". OMG. Compared with similar projects of similar complexity, this particular project ended up costing about 100% more, or $20M, and was of course massively late.
I like your comment about "thinking about OO problems". Could we say OOD is software engineering for software engineering's sake? Oh, and for the worthies who sell books and CASE tools?
You are not alone!