Mr. Fordtran has presented an elegant and thorough argument that the incidence of colitis due to Clostridium difficile may be increased because of our patterns of medication use. The Pharmacy and Therapeutics Committee at Baylor University Medical Center is charged with monitoring drug use for safety, efficacy, and appropriateness. Proton pump inhibitors (PPIs) have been widely used in a variety of circumstances and to the casual prescriber are free of significant side effects, drug-drug interactions, allergies, and other untoward consequences. But their power to affect cellular biology is unquestioned, and thus it is reasonable to expect some risk as well as benefit, as is the case with all medications. As emphasized in Dr. Fordtran's review, PPIs may be implicated in increasing the incidence of C. difficile colitis, though the evidence for this assertion remains scant. At Baylor University Medical Center, PPIs have recently been deleted from intensive care unit order sets, where they were considered an option for routine stress ulcer prophylaxis. Sucralfate and H2 blockers remain alternatives, though sucralfate can be difficult to administer to a critically ill or unconscious patient and may interfere with the adsorption of concomitantly administered drugs. The Pharmacy and Therapeutics Committee struggles with the various ways that some “prewritten” orders are used. Prescriptions may enter the system as personal order sets that are preloaded into Baylor's clinical computer system (BCON), as well as through care paths, protocols, or preprinted order sheets. These orders may have been nationally approved by various professional organizations, extensively considered by local quality improvement committees, or simply brought in as a personal favorite of a single physician. While prewritten orders can improve physician efficiency, it is especially important that medication orders encourage “best” prescribing practices, based on current evidence, agreed-upon guidelines, and local policies. It remains an arduous but important task to review and approve these orders and to keep them updated and properly distributed. As we move toward electronic prescribing and computer-assisted physician order entry, we will have further opportunities to guide and direct medication use. With prescribing electronically facilitated, it will become even more important that this guidance is based on the best current evidence. Thoughtful analysis, as shown in Dr. Fordtran's review, is not only enlightening and welcome but also critical to ensure that technology is used to improve patient care and not to inadvertently automate suspect decision-making.