Abundant evidence suggests a relationship between antibiotic resistance and use, including animal models, consistent associations between resistance and antibiotic use in hospitals, concomitant variation in resistance as antibiotic use varies, and a dose-response relationship for many pathogen/antibiotic combinations. Much of the evidence has come from studies performed in single hospitals. Most multicentre studies on resistance have not included data on antibiotic usage. Despite this substantial body of evidence, some studies have failed to demonstrate an association between antibiotic resistance and use, suggesting other contributing factors such as cross-transmission, inter-hospital transfer of resistance, a community contribution to resistance, or a complex relationship between resistance and the use of a variety of antibiotics. A multicentre study, project ICARE (Intensive Care Antimicrobial Resistance Epidemiology), implemented in 1994 by Centers for Disease Control and Prevention and Rollins School of Public Health, Emory University, has found dramatic differences in the patterns of antibiotic usage and resistance in US hospitals. The findings suggest that antibiotic usage is the major risk factor in development of antibiotic resistance in hospitals but the relationship can be complex with additional factors involved. Understanding the problem of antibiotic resistance in a hospital cannot be achieved without knowledge of the hospital's pattern of antibiotic use.