Evaluating Possible Confounding by Prescriber in Comparative Effectiveness Research