Saturday, March 20, 2010: 10:30 AM
Centennial I-II (Hyatt Regency Atlanta)
Background: Monitoring adherence to evidence-based central line insertion practices (CLIP) is a fundamental component of multifaceted interventions proven to reduce central line-associated bloodstream infection (CLABSI) rates in intensive care units (ICUs). Monitoring and feedback of CLIP can help identify improvement opportunities resulting in increased adherence. In 2008, the National Healthcare Safety Network (NHSN) launched CLIP adherence monitoring. Adherence was monitored by a third-party during insertion or by the inserter after insertion (self-report). Objective: 1) To examine temporal changes in CLIP adherence during the first year of NHSN reporting; and 2) to test whether temporal changes vary by location type (ICU vs. non-ICU) and monitor type (third-party vs. self-report). Methods: Adherence to the NHSN CLIP bundle was defined by adherence to all of the following practices: hand hygiene, appropriate skin antisepsis, dry antiseptic before skin puncture, and maximum barrier precautions. We assessed adherence over a one-year period for locations reporting at least six months of CLIP data. To describe temporal changes in overall adherence, we used logistic regression to model CLIP adherence by time (in months); we then specified separate models for each location and monitor type. We used a multivariable interaction model to statistically test whether location type or monitor type modified the impact of time on adherence. Results: Between 7/1/08 and 6/30/09, 51,468 central line insertions were reported to NHSN from 433 unique locations. Overall adherence during this time period was 91%. Odds of adherence increased by 8% per month overall (OR=1.08, 95% CI=1.07–1.10). In ICUs (n=45,619 insertions in 226 locations), odds increased by 9% per month (OR=1.09, 95% CI=1.08–1.10). In non-ICU settings (n=5789 insertions in 13 locations) there was no increase in the odds of adherence over time (OR=0.98, 95% CI=0.95–1.01). For both self-assessment (n=28,979 insertions in196 locations) and third-party monitors (n=22,489 insertions in 237 locations), there were significant increases in the odds of adherence over time. In the multivariable interaction model, there was a significant interaction between time and location type (p=0.002) and no interaction between time and monitor type. Conclusions: This preliminary assessment suggests that overall CLIP adherence increased over the course of one year of monitoring and that this improvement was strongly influenced by adherence improvement in ICUs. In non-ICU settings, where CLIP monitoring and feedback may be less familiar, no improvement over time was noted. Significant improvement over time was noted when adherence was assessed by either third-party observers or inserters indicating that despite potential for reporting bias, self-report may still encourage improvement in adherence to CLIP.