Objective: We sought to identify the presence or absence of international variation in central venous catheter-associated bloodstream infection (BSI) rates and to examine associated infection control practices that might underlie the differences.
Design: The Evaluation of Processes and Indicators in Infection Control (EPIC) study was conducted as a prospective surveillance study.
Settings: The study took place in intensive care units (ICUs) from 14 countries, which were from the Asian Pacific (3), Europe (7), Middle East (2), and South America (2), in addition to 41 US hospitals.
Methods: We compared the National Nosocomial Infections Surveillance catheter-associated BSI rate between the non-US and US units. We also compared the following organization factors between the 2 groups: hospital factors (ownership, average daily census of patients); ICU type (medical vs surgical); number of beds; and infection control-related factors (number of staff, number of hours spent on study ICU surveillance, years of experience, number of inservice sessions on line infection, number of blood cultures drawn/1000 patients).
Results: We found no significant difference in catheter-associated BSI rates between non-US and US hospitals (5.02 +/- 0.75 vs 3.82 +/- 0.42/1000 days, respectively; P =.27). Non-US hospitals were more likely to be government-owned (10/14 vs 7/41;P <.001) and to have larger daily patient census (795 +/- 84 vs 276 +/- 47 patients; P <.001). There was no difference in ICU type or number of beds. Infection control committees were present in all US and non-US hospitals. No significant differences were found in the number of staff involved in surveillance in the study ICU, years of experience, hours spent on surveillance, or the provision of inservices on line care. The use of barriers during line insertion also did not differ.
Conclusions: Catheter-associated BSIs in patients in the ICU were not significantly different between non-US and US hospitals. All hospitals had infection control committees, and there were no significant differences in time spent and numbers of persons involved in ICU surveillance activities. These findings suggest that many aspects of the standards of care do not differ between the 2 groups.