An interest in using nosocomial infection rates as an outcome measure to reflect quality of care in hospitals prompted us to consider factors in addition to quality that influence these rates. Approximately one third of nosocomial infections are potentially preventable, and changes in this “preventable” stratum of infections should reflect variations in quality. However, it will be necessary to identify those potentially preventable infections by calculating rates which are adjusted for intrinsic patient risk. Five other factors necessary for nosocomial infection rates to be a valid and reliable indicator of quality include identification of critical indicators (e.g., types of infection) and sampling schemes that most accurately reflect variations in quality; adoption of standardized, objective definitions of site-specific nosocomial infections; adoption of universal denominators across institutions; development of a monitoring system to assess compliance with surveillance and reporting procedures; and the adoption of more standardized training for infection control practitioners.