918 Effect of Targeted Surveillance on Catheter-Related Bloodstream Infection Incidence Rates in Quebec Intensive Care Units

Sunday, March 21, 2010
Grand Hall (Hyatt Regency Atlanta)
Patricia S. Fontela, MD, MSc, PhD(c) , McGill University, Montreal, QC, Canada
Robert W. Platt, PhD , McGill University, Montreal, QC, Canada
David Buckeridge, MD, PhD , McGill University, Montreal, QC, Canada
Madhukar Pai, MD, PhD , McGill University, Montreal, QC, Canada
Caroline Quach, MD, MSc , Montreal Children's Hospital, Montreal, QC, Canada
Background: Catheter-related bloodstream infections (CR-BSI) are a major problem in intensive care units (ICUs). In 2003, the Institut National de Santé Publique du Québec launched the Surveillance Provinciale des Infections Nosocomiales (SPIN) program -- a continuous surveillance program that aims to monitor ICU CR-BSI incidence and mortality rates. It is hypothesized that programs such as SPIN may help to reduce CR-BSI incidence rates. However, methodological limitations of previous studies have precluded accurate estimation of the effect of such surveillance programs.
Objective: To determine the effect of a targeted surveillance program (SPIN) on CR-BSI incidence rates in Quebec ICUs.
Methods: Before-after study including 56 ICUs that continuously participated in the SPIN program for at least 2 consecutive years between 2003 and 2009. The intervention was defined as the first dissemination of the SPIN results to the participating ICUs. Primary outcome was the change in the pooled CR-BSI incidence rate that happened between the periods before (year 1) and after (year 2) the intervention. We calculated the standardized incidence ratio (SIR) and its 95% confidence interval (95%CI) using the CR-BSI incidence rate from year 1 as the reference rate. Data analysis was repeated after stratifying data by academic profile (teaching vs. non-teaching ICUs), ICU type (adult, pediatric and neonatal) and volunteer status (ICUs that joined SPIN before participation became mandatory in 2007 vs. ICUs that only joined SPIN after it became mandatory). CR-BSI incidence rates were calculated according to the National Healthcare Safety Network (NHSN) specifications.
Results: Between 2003 and 2009, 46 adult, 4 pediatric and 6 neonatal ICUs participated in the SPIN program for at least 2 consecutive years. For all ICUs, we obtained a SIR of 0.94 (95%CI 0.82-1.07). Among non-volunteer ICUs, the SIR was 0.68 (95%CI 0.48-0.93) compared to the SIR of 1.03 (95%CI 0.89-1.19) obtained for the volunteer subgroup. Non-teaching ICUs yielded a SIR of 0.82 (95%CI 0.55-1.17), while the SIR for teaching units was 0.98 (95%CI 0.84-1.13). Finally, after stratifying data according to ICU type, the SIRs for adult, pediatric and neonatal ICUs were 0.90 (95%CI 0.77-1.05), 1.06 (95%CI 0.58-1.78) and 1.07 (95%CI 0.77-1.45), respectively.
Conclusions: Preliminary results show that targeted surveillance is associated with a decrease in ICU CR-BSI incidence rates observed among units that joined the SPIN program after it became mandatory. The latter is likely a surrogate for not having a local surveillance program in place prior to the implementation of mandatory participation.