Patients frequently express strong preferences for medical tests or treatments of their own choosing, even when physicians believe that those interventions are not beneficial. Physicians grant such requests for various reasons. One compelling reason is to avoid confrontation: patient-physician relationships flourish in an atmosphere of trust and goodwill, and physicians rightly worry that disagreement will threaten those relationships. Moreover, explaining why an intervention is not beneficial takes time. For patients with the common cold, granting requests for antibiotics is far less time-consuming than discussing viral microbiology and harms of antibiotic overuse. Although patients' preferences are key factors in clinical decision making, a patient's preference for a diagnostic or therapeutic intervention is not decisive unless a modicum of potential benefit, viewed from a conventional medical perspective, is present.1 When diagnostic or therapeutic choices are consistent with such a modicum of benefit, patients' preferences should drive decisions. In contrast, physicians should not provide interventions that do not meet this criterion.