Hot break is mostly protein. But the density of wort is a measure of all dissolved material -- of course mostly sugar, but also small contributions from soluble proteins, and even hop acids and oils, and the mineral content of the water. I presume that the removal from solution of a portion of the proteins would be measurable, at least with a hydrometer. I don't know how or if a refractometer calibrated with a sucrose solution would read this. (Maybe I should know, because I too use a digital Plato refractometer.) On the homebrew scale, I have always assumed there is a real effect which accounts to some degree for my observed difference in increase in density as actually measured and as predicted by the evaporation rate. But I also assume that this is all within the margin of my measurement accuracy and resolution (if my eyeball reading of my dipstick is off by a few fluid ounces, and my refractometer has an accuracy of +/- 0.2 °P, any real effect will probably be obscured by error.) But perhaps on a larger scale, you are able to more reliably observe such small variations? This has always been in the back of my mind, and I am curious to know if there is any good scientific study of this. It might even in some way provide an index index of effective boiling, insofar as the precipitation of undesirable proteins is a primary goal of the boil.