Objective: Identification of catheter-related bloodstream infection (CR-BSI) risk factors and determination of whether intervention related to identified risk factors would reduce CR-BSI rates.
Design: Prospective, observational, interventional and interrupted time-series study.
Setting: Pediatric Intensive Care Unit (PICU) in a university hospital.
Methods: During a 7-year period, 609 central venous catheters (CVC) were placed in 389 patients. CR-BSI risk factors were determined by multivariate analysis during two periods (January 2000-November 2002 and January 2003-April 2007). An intervention to reduce identified risk factors was performed after the first period. CR-BSI rates per 1,000 catheters-days were compared during the two periods.
Results: The CR-BSI rate was 11.94 [(95% CI 7.94-15.94)/1,000 catheter-days during the first period]. Weight [OR 0.96 (0.91-0.99)], parenteral nutrition (PN) [OR 3.38 (1.40-8.19)] and indwelling time (IT) [OR 1.08 (1.02-1.14)] were CR-BSI risk factors. Practice changes aimed at reducing PN and IT were introduced. PN decreased from 49.8% [95% CI (49.7-49.9)] to 26.7% [(95% CI 26.6-26.8)] (p < 0.001), and IT dropped from 9.92 (95% CI 9.09-10.75) to 8.13 (95% CI 7.47-8.79) days (p < 0.001). The CR-BSI rate was reduced to 3.05 (95% CI 0.93-5.17)/1,000 catheter-days. During the last period, PN and IT were no longer CR-BSI risk factors. Type of catheterisation (guide wire exchange) [OR 6.66 (1.40-31.7)] was the only CR-BSI risk factor.
Conclusions: PN and IT were independent CR-BSI risk factors during the first period. An intervention focused on PN and IT reduction resulted in a sustained decrease of CR-BSI rates in our PICU.