To the Editor.
—The investigators of the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT)1 are to be commended for meticulous research and their commitment to improving the care of seriously ill patients. Unfortunately, SUPPORT was based on at least 3 assumptions, each of which can be questioned.Their first assumption is that patients want information to compute expected values and make a maximally informed choice. In SUPPORT, patients and their families reported satisfaction with their care, even while they noted that physicians had failed to discuss prognosis. The medical literature suggests that patients seek information without necessarily wanting to make a decision; when they do wish to decide about end-of-life issues, they may not be computing expected values.2-4 Information sharing is not an adequate model for understanding patient preferences, and it is a poor strategy for changing outcomes.Another assumption is that