Recent public attention has been directed toward the risks associated with certain drugs that are in wide-spread use. As with most drugs after FDA approval, assessment of the risks and clinical benefits of these products becomes very complicated because the relationship between drug use and patient outcomes is often confounded in complex ways that can not be addressed by conventional statistical techniques. Recently, researchers at Harvard School of Public Health (HSPH) have developed advanced statistical methods (referred to as marginal structural models - MSM) that adjust for such complex confounding practice patterns. In this grant, "Epoetin Therapy and Survival of Hemodialysis Patients," MSM techniques will be applied to explore the causal relationship of epoetin therapy for treatment of anemia on survival for hemodialysis patients. The application of these methods will provide a state-of-the-art estimate of the effect of epoetin on survival.