(WHO) recommendations (HHi) targeting the entire ICU‐unit [12]; and 3) both interventions combined (COMBi) (more information in Supp. Methods). CRBSI was the primary outcome, process indicators, i.e. a CVC insertion score and hand hygiene compliance, were secondary outcomes. Interventions targeting quality of care were implemented at ICU‐ward level (cluster level) whereas ICU patients were monitored for CRBSI and healthcare professionals for compliance. Central venous catheters and CRBSI Surveillance was performed prospectively by the local PROHIBIT study nurse. Tunneled or peripherally‐inserted central catheters, as well as CVCs already present at hospital admission or in situ for only one day were excluded. CVCs were followed up from insertion to removal. Dwell‐time was censored in the case of CRBSI, beyond 48 hours after ICU discharge, and after fatal outcome. CRBSI was measured as an incidence density (per 1000 CVC‐days). Bloodstream infection (BSI) was defined according to the US Centers for Disease Control and Prevention (CDC) criteria for laboratory‐confirmed BSI. CVC relation was defined using the ECDC HAI‐Net criteria for microbiologically‐confirmed CVC‐related BSI (CRI3) [13], with the added criterion of the timely resolution of clinical symptoms after CVC removal or after starting antimicrobial therapy (Supp. Methods and results). Process indicator surveillance CVC insertion compliance was measured by direct observation using a list based on the CRBSI prevention protocol developed at the University of Geneva Hospitals (Supp. Fig. 1) and expressed as the proportion of fulfilled protocol items divided by the total number of protocol items (CVC insertion score). Hand hygiene compliance was measured by direct observation according to the WHO “My five moments for hand hygiene” concept [14]. Observations of CVC insertion were randomized for date (weekdays) and place of insertion (ICU, operating theatre [if ICU patients regularly received the CVC there]) and were performed during daytime. Hand hygiene observations were randomized for date (weekdays), time slots (08‐12:00, 12:00‐16:00 and 16:00‐20:00), and ICU beds. OSIs and study nurses were advised not to disclose any outcome or process indicator results during baseline. During the intervention, quarterly feedback reports on CRBSI and process indicator(s), according to the allocated intervention, were made available to OSIs by the project coordination team.
RkJQdWJsaXNoZXIy MTk4NDMw