On the importance of assumptions for bulk flow in hemodynamic models of the carotid bifurcation