The way I understand it is by knocking down the bees population curve at a critical time (before they reach critical mass) you can avoid that inevitable period where there is more mites then bees when they slow down brood rearing. Randy Oliver's work on varroa population dynamics can help you to understand this. There is however no hard and fast peer reviewed studies that I can find about varroa populations and brood breaks. All the studies of this effect are for treatments. He is also coupling this with his OTS queen rearing techniques as I understand it, you are creating a brood break of about 30 days, as they develop a queen. But assuming that there are eggs, and the mites will have somewhere to reproduce it turns out to be a brood break of 10 +/- days (I think, dont quote me on that). Furthermore, he goes on to point out that when you have your queen begin to lay all the varroa will rush into the first 20-30 cells about to get capped. In worker cells 1-2 mites can live comfortably and any more then that they will suffocate and die.
I think this is a good method worth some serious investigation, but would be very dependent on your season, and when they reach critical mass in brood population. Hes in Michigan and im in Ontario so it "should" be similar.
Another interesting point he made was that by going queenless before your flow, the bees do not have to feed brood and you will actually have MORE honey. However, if you have multiple flows then you would have a severely reduced population of bees and that would affect production.