You're still breeding resistance, just slightly slower.
In the best case, the full treatment kills all of them, or enough that your immune system can eliminate them completely. Resistance doesn't develop because it doesn't get a chance. They're all dead.
Worst case, you kill enough to weed out those with no or weak resistance, and then spread the infection to others where it can continue to evolve.
So it comes down to (a) how much are you concentrating/selecting for these resistances, and (b) how likely are you to pass it on such that it propagates.
I think it's reasonable to assume that a lower bodily population makes it less likely to pass on, which would be one good reason. Another might be that, by heavily reducing the population, you may reduce the genetic diversity enough that they're susceptible to other factors (ie: when not facing antibiotics, they get starved out by the non-resistant strains because the resisting mutations come with costs attached). By leaving a larger population alive, there's more scope for hybridisation leading to overall stronger species.
I must admit I have little formal knowledge in this area, so it's possible these effects don't exist, or have only limited impact. If I had time it would be interesting to look into it further, it seems like this topic is too obvious not to have been considered before.