As with the CVC insertion score, increasing hand hygiene compliance was associated with decreasing CRBSI incidence density: the incidence rate ratio (95% CI) per PP increase of hand hygiene compliance over the entire study was 0.99 (0.98–1.00). After adjustment for the proportions of patients with bloodstream infection at the time of insertion and of patients with prolonged ICU stay before insertion, this association did not remain significant, with an incidence rate ratio of 1.00 (0.99–1.01) for the entire population, and for the CVCi‐ and COMBi arms. However, the adjusted association was significant in the HHi arm (Supp. Table 6). DISCUSSION Our results demonstrate that the introduction of a best practice CVC insertion strategy, a WHO‐based HH promotion strategy, and the combination of both, significantly improve process indicators, and reduce CRBSI incidence densities. When taking into account a decreasing trend during baseline both the HH program and the combined HH and CVC insertion strategy were still effective. The low baseline rates in the CVCi arm limited the power to demonstrate the same effect in this arm. This is the first multinational randomized multicenter CRBSI prevention study providing sufficiently powered information on both outcome and process indicators. Many studies have reported successful CVC insertion or hand hygiene improvement initiatives [18‐22]. However, most CRBSI prevention studies reported outcome data only, without mentioning process indicator data, as evidenced in a recent systematic review by Ista et al. [18]. In a randomized Canadian study, compliance with CRBSI‐prevention measures increased from 10% at baseline to 70% during intervention in the intervention ICUs, while compliance increased from 31% to 52% in the control ICUs [20].The study did not report CRBSIs and numbers of CVCs were low. Non‐randomized before‐and‐after studies, partially using retrospective data, have reported CVC “bundle” compliance in the range of 20%‐37% [23], 55.2% [24], 74% [25] and 90%‐100% [21, 26, 27]. An Australian multicenter study reported variation of bundle compliance between hospitals ranging from 0 to 100% (personal communication from McLaws) [21]. Most studies used an “all or nothing” approach where the outcome was met if all items of the bundle were fulfilled. Our CVC checklist consisted of 20 items, which is beyond the usual bundle promotions. Thus, in order to be discriminatory and to capture gradual quality improvement over time, we analyzed CVC insertion as a score. Our baseline hand hygiene compliance of 48% is similar to the 52% identified by the MOSAR study in 13 European ICUs [28], and to the 40‐50% reported by a review summarizing 65 ICU studies [29]. The effects of our intervention are in line with the
RkJQdWJsaXNoZXIy MTk4NDMw