The interrater reliability was: 3CAT weighted kappa (wk) 0.68 (95%CI: 0.61‐0.75); WHOCAT wk 0.65 (0.58‐0.73); QUANT intra‐cluster correlation coefficient 0.76 (0.71‐ 0.81). Interrater reliability ranged from 0.72 for pneumonia to 0.52 for CDI. We concluded that feasibility, validity and reproducibility of the three MR measures was acceptable for use in HAI surveillance. Part II Surveillance is known to increase awareness, which in itself can lead to lower HAI rates. Nevertheless, intervention programmes are usually needed to optimize practices to prevent HAI. The concept of ‘patient safety’ and the notion that it should be embedded throughout the entire healthcare system was developed in the US and expanded in the late nineties. In 2004 the Institute for Healthcare Improvement (IHI) in the US initiated the 100,000 Lives Campaign to improve patient safety and outcomes. One of its six recommended interventions was the ‘central line bundle’ to prevent CRBSI/central line‐ associated BSI (CLABSI), which has been implemented in many hospitals and national surveillance programmes since. Included in these bundles is hand hygiene during the CVC insertion, as hand hygiene is a corner stone of infection prevention in general. Hand hygiene needs to be performed at various indications and, when attending to hospitalised patients, repeatedly so. Additionally, work pressure and variable personal perceptions of cleanliness challenge hand hygiene compliance. Within the PROHIBIT (Prevention of Hospital Infections by Intervention and Training) study we evaluated whether an extensive CRBSI prevention bundle (CVCi), a WHO‐based HH intervention (HHi), or both in combination (COMBi) would be effective in CRBSI prevention (chapter 7). ICUs from 14 hospitals in 11 European countries participated in this stepped‐wedge cluster randomised study, lasting from 2011 to June 2013. After a 6‐ month baseline, three hospitals were randomised to one of three interventions every quarter. The primary outcome was the prospective CRBSI incidence density. Secondary outcomes were a CVC insertion score (because there were 20 items, compliance was not evaluated as a bundle but as the proportion of items adhered to) and HH compliance. The CRBSI incidence density decreased from 2.4/1000 CVC‐days at baseline to 0.9/1000 (p < 0.0001). When adjusted for patient and CVC characteristics all three interventions significantly reduced CRBSI incidence density. When additionally adjusted for the baseline decreasing trend, the HHi and COMBi arms were still effective. CVC insertion scores and HH compliance increased significantly in each of the three interventions arms. 268 Summary
RkJQdWJsaXNoZXIy MTk4NDMw