bpy.ops.object.duplicate() preserves the relative relationships between duplicated objects— Properties originally set to any of the original objects in parenting, modifiers, constraints, and drivers get set to their corresponding duplicated objects.
What would be the most efficient way to replicate this behaviour with low-level calls? Traversing the objects' hierarchies to find and replace all constraint, modifier, and driver fields that refer to the original objects would surely nullify much of the performance gain from using low-level calls in the first place.
