1.57)); however, the trend indicated a 24% increase in CRBSI risk when all bundle items were complied with. When evaluating the partial bundles instead, the association of compliance with the insertion bundle was not associated with CRBSI risk (HR 0.84 [0.56‐ 1.25]). Compliance to the maintenance bundle was unexpectedly associated with a higher CRBSI risk (HR 1.31 [1.03‐1.67]), i.e. the CRBSI risk was 31% higher when the maintenance bundle was complied with. To evaluate the effect of ongoing participation, the number of participation years was added to the multivariable model. This resulted in a comparable hazard ratio for the maintenance bundle (HR 1.28 [1.00‐1.62]), and revealed a significant decrease with ongoing participation (HR per extra participation year 0.93 [0.87‐0.99]). This reduction of 7% per year was an average with individual hospitals demonstrating a more erratic pattern over time. Adding calendar year to the model instead of the number of participation years had a similar effect. When evaluating the first four years and the remaining participation years separately, i.e. in two separate multivariable models, the effect of each extra participation year was significant in the first period of four years (HR 0.87 [0.78‐0.97]) but not so thereafter (HR 0.91 [0.73‐1.12]). In the ICU: In the ICU the CRBSI incidence was lower than outside the ICU (Figure 1). Mean overall bundle compliance for CVCs in the ICU (for the entire or part of the catheter duration) was 67.6%, whereas this was 52.9% for CVCs that were in place outside the ICU only. This disparity resulted from a difference in the maintenance bundle compliance (69.7 and 49.2%, respectively), whereas the insertion bundle compliance was comparable (93.2 and 93.7%, respectively). In the ICU, neither compliance with the overall bundle, nor the separate insertion and maintenance bundle (together in a model) were independently associated with the CRBSI risk (HR 1.14 [0.80‐1.64], 1.05 [0.56‐1.95] and 1.13 [0.79‐1.62], respectively). Ongoing participation was associated with a significant reduction of the CRBSI risk (HR 0.90 [0.82‐0.98]). Calendar year instead of the number of participation years demonstrated a similar association (HR 0.89 [0.82‐0.96] and explained the results slightly better. When evaluated separately for the first four and remaining years this risk reduction was significant for the first participation years only (HR 0.78 [0.65‐0.94] and 0.99 [0.76‐1.29]) respectively). Outside the ICU: For CVCs outside the ICU, compliance with the overall bundle seemed associated with an increased CRBSI risk, although the association was not statistically significant (HR 1.36 [0.96‐1.93]). A model with the separate insertion and maintenance 9 225 The effect of a CRBSI prevention bundle in the Netherlands
RkJQdWJsaXNoZXIy MTk4NDMw