The hysteresis loop area gives the actual power loss (neglecting eddy currents in the core), but it's generally infeasible to manually calculate that for any arbitrary situations. You'd basically need a perfect nonlinear transformer model... Rather the manufacturer usually gives loss as a function of Bp-p, which is much easier to deal with, and I assume is very closely related with hysteresis losses. I'm sure there's some heavy math behind it, but probably only the core material engineers are qualified to figure it out.