Over the last 20 years, many studies explored how the way information is presented modifies choices. This sort of effect, referred to as "framing effects", typically consists of the inversion of choices when presenting structurally identical decision problems in different ways. It is a common assumption that physicians are unaffected (or less affected) by the surface description of a decision problem, because they are formally trained in medical decision making. However, several studies showed that framing effects occur even in the medical field. The complexity and variability of these effects are remarkable, making it necessary to distinguish among different framing effects, depending on whether the effect is obtained by modifying adjectives (attribute framing), goals of a behavior (goal framing), or the probability of an outcome (risky choice framing). A further reason for the high variability of the framing effects seems to be the domain of the decision problem, with different effects occurring in prevention decisions, disease-detection decisions, and treatment decisions. The present work reviews the studies on framing effects, in order to summarize them and clarify their possible role in medical decision making.