Nosocomial infection rates for interhospital comparison: limitations and possible solutions. A Report from the National Nosocomial Infections Surveillance (NNIS) System.

Improvement in quality of patient care has received increasing attention in the last decade, with emphasis in infection control and interhospital comparison of infection rates. One of the main objectives of the National Nosocomial Infections Surveillance (NNIS) system is to provide hospitals with comparative nosocomial infection data that at least partially adjust for patients' intrinsic and extrinsic risks for infection. This article summarizes the methods and results of analyses from the NNIS system and describes their application to future surveillance in US hospitals. We emphasize the importance of nosocomial infection surveillance data that adjust for specific infection risks in order to provide better interhospital comparison of infection rates. Traditional rates that do not provide such adjustment include the crude overall nosocomial infection rate of a hospital or service and site-specific infection rates by service. Because these inadequately adjusted rates are potentially misleading, they should not be used for interhospital comparison. This article describes several new infection rates, including device-associated, device-day infection rates for intensive care units and high-risk nurseries, and an NNIS surgical wound infection risk index. These rates appear to be better for interhospital comparison. NNIS data also suggest the importance of examining interventions (devices and operative procedures) that increase patient risk for infection. Failure to use appropriately adjusted rates and to examine the intervention experience may make interhospital comparisons meaningless or even misleading.