Reduction of Central Line Associated Bloodstream Infections [CLABSI] in a Neonatal Intensive Care Unit (NICU): A 13 Year Experience

Background: Because bloodstream infections are a major cause of mortality in neonates, prevention of CLABSI has been a focus in our neonatal intensive care unit [NICU] with the use of NNIS/NHSN (National Nosocomial Infections Surveillance/National Healthcare Safety Network) to measure improvements since the early 1990's. We present a summary of 13 years of continuous surveillance and process improvement in our NICU since 1995.

Objective: Describe the epidemiology of CLABSIs in NICU and interventions leading to their reduction.

Methods: We used NNIS/NHSN protocols to collect CLABSIs and central line days by birth weight from our NICU (Level II/III). We analyzed available annual data from 1995 to 2007 (due to definition change effective 1/2008, 2008 and 2009 data were excluded). Interventions: In 1994, a multidisciplinary team was formed to address central line (CL) infections utilizing the FOCUS-PDCA process improvement method. Initial interventions (in 1995) included the use of maximum sterile barriers for insertion of CL, a closed system for arterial blood sampling and flushes, preparation of all intravenous (IV) medications under laminar flow, IV tubing configuration change with 24 hour medication attached and extensive hand hygiene campaign. Peripherally inserted central lines were introduced in late 1997. Due to a gradual rise in CLABSI rates, a focus on line maintenance ensued, alcohol hand sanitizer was installed (1999) and initiatives were taken to decrease the dwell time of umbilical catheters. Interventions from 2002-2007 focused on line maintenance, aseptic technique for tubing changes, maintenance of a closed system, hand hygiene and proper technique for obtaining blood cultures. We calculated ‘Standardized Infection Ratio' [SIR], a single metric that takes into account different birth weights for each calendar year. SIR is a summary measure used to compare the CLABSI event among one or more groups of patients to that of a standard population. The standard population reference is computed by appropriately summing up measurement of two national surveillance systems (NNIS 1995-2004 and NHSN 2006-2007). SIR is the ratio of ‘observed' CLABSI / ‘statistically expected' CLABSI derived from the standard population. We also looked at the trends of SIR by time series analysis by baseline [P-1: 1/1995-12/1999], period 2 [P-2: 1/2000-12/2007].

Results: Overall, there was a significant decline in SIR over the study period (6% reduction in SIR each year, p<0.01); P1: 5.8% reduction /year, p=0.53; P2: 4.3% reduction/year, p=0.01) (see figure-1). Initially, the infection rate was 60% higher than that of the standard population (SIR 1.6, 95% confidence interval [CI]:1.03-2.41), whereas it was 60% lower at the end of the study period (SIR 0.43, 95%CI: 0.27-0.69).

Conclusions: Key interventions and utilizing NNIS/NHSN resulted in significant reduction of CLABSI in our NICU.

 

---