I had a patron come in asking about material (not just books, but articles too) about 'Optimal Sleep Time'. You've probably heard of this. It's been the focus of a few studies where the result is that high schooler's do better in school when the school starts later, and kindergartners do better earlier, when the school starts early.
Unfortunately, a search (not just through several databases, but google and google scholar) of "Optimal Sleep Time" only gave me a couple results, much to my surprise.
After finding the few scant articles that specifically mentioned the words "Optimal Sleep Time" for my patron, I started to brainstorm other ways to search for the information she was looking for.
The one that got the most results? From google and our databases? "High school sleep", and a few variations of that. So, after probably fifteen minutes of searching for what we'd consider the 'proper name', a few keywords pulled up more than anything.
With her second topic, I just basically ignored the 'proper name' and did a keyword search (early intervention at risk kids), and a few variations of that as well. (Of course, that didn't have a proper name so much as the Optimal Sleep Time did). I found her enough articles to make her very happy.
So, using the very specific isn't always a good idea, even though, logically, if you're looking for that one thing, it should be. Go a bit broader. Do a keyword search. That's also how I find a lot of the obscure books that patrons ask me about. If they don't remember much (like just a plot), a keyword search can usually get me a lot further than most anything else.