Some thoughts ..
(a) Vmca normally is a work up for static values and then checked to make sure that the dynamic failure is not more critical
(b) the sea level value is an extrapolation from the altitude work
(c) playing with low speed asymmetric thrust ... ie near stall ... is silly at best for non TP folk .. so, for normally aspirated birds, he who plays with this at altitude is even sillier ..
(d) for most light twins there is no performance left to play with at or about Vmca so the pilot ought not to be anywhere near Vmca in most Types for any normal or routine OEI situation
(e) "Vmca" demonstrations have very little to do with Vmca as such although they provide an exposure to the static Vmca situation
(f) real world Vmca is very much bank angle dependent. Generally it may be assumed that the book figure is based on 5 degree bank (only because the test is not allowed to use more bank ..). If one doesn't use this figure the Vmca goes up .. and, if the bank is allowed to go the wrong way, it goes up rapidly ..
(g) if you must expose the student to the yawing problem .. rather than just discussing it over coffee, then consider arbitrarily introducing a conservative delta by having the instructor restrict rudder travel. A colleague once had to do a lightie endorsement to renew an instrument rating .. the instructor thought it appropriate to have a look at Vmca .. my colleague (who knew lots about Vmca in any case) thought otherwise .. so he restricted rudder travel himself .. confused the instructor greatly ... but got my colleague the tick in the box without exposing himself to fearful risk.
At the end of the day, Vmca is a line in the sand certification boundary condition and has little to do with real world flying .. so, why go there ? .. it's a bit like playing marbles on the middle of a busy road ..