Development of a statewide collaborative to decrease NICU central line-associated bloodstream infections

J Perinatol. 2009 Sep;29(9):591-9. doi: 10.1038/jp.2009.18. Epub 2009 Mar 5.

Abstract

Objective: To characterize hospital-acquired bloodstream infection rates among New York State's 19 regional referral NICUs (at regional perinatal centers; RPCs) and develop strategies to promote best practices to reduce central line-associated bloodstream infections (CLABSIs).

Study design: During 2006 and 2007, RPC NICUs reported bloodstream infections, patient-days and central line-days to the Department of Health, and shared their results. Aiming to improve, participants created a central line-care bundle based on visiting a potentially best performing NICU and reviewing the literature.

Result: All 19 RPCs participated in this quality initiative, contributing 218,096 patient-days and 56,911 central line-days of observation. Individual RPC nosocomial sepsis infection (NI) rates ranged from 1.0 to 5.8 NIs per 1000 patient-days (2006), and CLABSI rates ranged from 2.6 to 15.1 CLABSIs per 1000 central line-days (2007). A six-fold rate variation among RPC NICUs was observed. Participants unanimously approved a level-1 evidence-based central line-care bundle.

Conclusion: Individual RPC rates and consequent morbidity and resource use attributable to these infections were substantial and varied greatly. No center was without infections. It is hoped that the cooperation and accountability exhibited by the RPCs will result in a major network for characterizing performance and improving outcomes.

Publication types

  • Multicenter Study

MeSH terms

  • Benchmarking
  • Catheter-Related Infections / prevention & control*
  • Catheterization, Central Venous / standards
  • Cross Infection / prevention & control*
  • Hand Disinfection / standards
  • Humans
  • Infant, Extremely Low Birth Weight
  • Infant, Newborn
  • Intensive Care Units, Neonatal
  • New York
  • Quality Indicators, Health Care*
  • Sepsis / prevention & control*