| osocomial Bloodstream Infections (BSIs) create a serious health problem in hospitals all over the world. Nosocomial BSIs are associated with significant morbidity, mortality, and case-fatality rates, as well as increased length of stay (LOS) and healthcare costs. Although a vast number of studies have been conducted around the world to examine the epidemiology of nosocomial BSIs, few have used a U.S. national probability sampling scheme. Hence, the purposes of this study were to generate U.S. national estimates of the incidence and severity of nosocomial BSIs, estimated LOS and total charges, as well as to identify risk factors for nosocomial BSIs among adults hospitalized in the U.S. The conceptual model proposed by Soule, Larson, and Perston (1995) was adapted for this study. This conceptual mode was modified for this study to incorporate specific risk factors for the acquisition of nosocomial BSIs. This study used cross sectional data based on admission and discharge data from a single year (2003). Furthermore, this study used a matching procedure (case-control design) for the cost analyses. The study used an existing dataset called the Nationwide Inpatient Sample (NIS). The NIS for the year of 2003 provides a probability sample of approximately 8 million inpatient stays from 994 hospitals in 37 states. In 2003, the weighted incidence and mortality rates for nosocomial BSIs projected to the U.S. population were 21.6 incident cases and 4.4 deaths per 1,000 admissions. The weighted case-fatality rate of nosocomial BSIs was 20.6%. The weighted average of LOS for nosocomial BSI cases was 16.0 days beyond the first 48 hours of admission and the average total charges of... |
