hemodynamic monitoring, administration of antibiotics, dialysis or, if none of these, other uses]; and, if present, infection criteria, infection date and microbiological results. CVC surveillance was continued for 28 days or until a CR‐BSI occurred, the CVC was removed, the patient was transferred to another hospital, died or when active (life‐ supporting) treatment was withheld. When ICU patients with a CVC in situ were discharged to another ward in the same hospital surveillance was continued. Mortality was an optional variable, recorded by some hospitals. Hospitals participated voluntarily and received feedback reports with case‐mix‐ adjusted infection rates. These reports were usually discussed by the infection control committee with ICU and other staff as appropriate. Quality assurance included annual workshops for participants and visits from an external validation team. Data were recorded locally but were checked for completeness and consistency on receipt by RIVM. Infection rates were expressed as the number of infections per 1000 CVC days. Time at risk (in days) was defined as removal date minus insertion date. Poisson and Cox regression in SAS 9.2 (Cary, USA) were used to calculate relative risks. All variables were fixed except for whether the patient with CVC was in the ICU on each specific day which was determined daily and was therefore time dependent. Length of ICU stay prior to CVC insertion was analysed as a categorical factor. Risk factors were tested for proportional hazards and plausible interactions. Risk factors with P ≤ 0.20 in the univariate regression analyses were initially included in the multiple regression models. The model was reduced with manual backward elimination. Confounding was investigated in this model. Statistical significance was defined at P ≤ 0.05. RESULTS Data from the first two years of surveillance were of poor quality and were therefore excluded. Nine hospitals (9% of all Dutch hospitals), including one university medical centre, collected data for one to four years between 2002 and 2009. Most hospitals collected data from adult ICU wards. In total, data were collected on 2565 patients, reflecting 3750 CVCs and 29,003 CVC‐days; 9.5% of these patients were not in an ICU, and 56.4% of the catheters were inserted in men. The median age of patients was 67 years. The median CVC duration was six days (interquartile range 4‐10). Overall, 1.6% of CVCs [95% confidence interval (CI) 1.2–2.0] resulted in CR‐BSI, representing 2.0/1000 CVC‐days (95% CI 1.6–2.6), and 2.3% of patients (95% CI 1.8–3.0) developed CR‐BSI. The average incidence rate varied between hospitals from 0.9 to 3.0/1000 CVC‐days. Of the 59 infections that occurred, 44% were based on both (semi) quantitative culture of the 3 47 Catheter application, vein and pre-insertion ICU stay affect CRBSI risk
RkJQdWJsaXNoZXIy MTk4NDMw