MCAT might not be a good predictor of success as a doctor. Is there some other objective measure that does a better job?
I don't know about "objective," but if you look at the study that I had posted, you might say that the last two years of med school begin to predict how well one might do as a doctor. This is because hands-on training (and work) is being performed. Think about this as possibly additional skill sets involving other dimensions: social and physical.
southernhybrid mentions above: emotional intelligence and that would be one aspect, perhaps under a social category. There could also be some factors such as resilience after observing traumas, keeping your cool.
And there's also a kind of procedural dexterity: I remember in basic training, we were taught to shout some verbatim definition for an M16, but then later, shooting an M16 in target practice and scoring well is different from the capacity to shout a definition, and later on shooting an M16 on the battlefield is yet again different from target practice.