Gastrointestinal bleeding accounts for more than 300,000 hospital admissions per year in the United States. Despite the introduction of an array of novel diagnostic and therapeutic approaches, the mortality rate associated with gastrointestinal bleeding has remained stable at 5% to 12%, primarily because of the rising percentage of elderly patients with gastrointestinal bleeding and the high mortality of bleeding that occurs during the course of a hospitalization for an unrelated medical problem. With defined high-risk subgroups of patients and competing diagnostic and therapeutic techniques, the study of gastrointestinal bleeding lends itself almost ideally to the application of outcomes and effectiveness research.