Alternatively, we could consider allowing triaged questions to get to close queue.
Thing is, the primary reason why inefficient triage is worrying is in the way how this impacts other reviews - specifically, close review.
It is important to understand that while the question is stuck in triage, system blocks it from getting to other review queues. As a result, we get the situation when system correctly and quickly detects most troublesome questions and then, essentially shields them from getting to close queue - for a full day average, go figure.
This is really really weird. The way how things work now, system essentially incentivises askers of poor questions to pay money for fraudulent release of asking limits. Seeing multiple poor questions hanging open and seemingly unattended for days can only convince them that their ban is mere misfortune and that it makes sense to try their luck again.
(Of course this is not how things were intended to work when triage development started. Triaged questions were supposed to be hidden from everyone except for reviewers so nothing like above would be possible. However this turned out infeasible and as a result we have system working opposite to its original intent.)
Given above, one could ask wouldn't it be better to just get rid of triage completely. I considered this and in fact, that would be my next suggestion in case if proposed changes won't help to improve it - because the way how it is now, it really seems to do more harm than good.
Okay, so what good does it do now and why would it be worth giving triage yet another chance to improve instead of just dropping it. My observations on this queue suggest that triage provides a great value of teaching users having under 3K rep. These users can't see close votes on questions - that is, to many of them the whole process of closing may look like obscure and coming out of thin air.
But to triage reviewers, system becomes really transparent. They can learn what questions aren't worth answering and what they should do instead. They learn how to use close flags and how these flags are handled by a system. And those of them who will get to 3K rep will know how to properly use their close votes from day one.
You know, I tried to figure how else people can learn this stuff and nothing I could imagine didn't even come close. The way how system picks inappropriate question for triage and all the mechanics of how this queue works makes it just impossible to beat.
Above, I wrote that the way how things work now, we better drop triage completely. I stand by that because I believe that harm of shielding inappropriate questions from close review for that long outweighs its teaching value. But I think it would be really a pity if we do this without even attempting to improve.
Note: above was also submitted as feedback in recent Review queue workflows - Final release at MSE.