Simulation in anaesthesia training

one possible explanation, sedation usage was not an independent factor in this study. Sedation usage was an observed parameter and neither controlled by the study design nor prone to changes of standardized procedures. However, sedation usage depends on the physician’s intention to leave a patient on the respirator. We agree that a disadvantage of this study is that we could not report differences in the VAP rates. However, there are no general accepted rules for VAP diagnosis. 6 This makes it difficult to compare VAP rates between different institutions. It may make it also impossible to detect small changes in VAP rates. We discussed this topic in the manuscript but want to emphasize again that this study did not include all ICU patients but only with a duration of mechanical ventilation of .24 h. Additionally, we intentionally did not report VAP rates per 1000 ventilator days. This parameter is prone to error when comparing patient populations with very different duration of mechanical ventilation. We also agree with O’Brien and colleagues that high compliance rates need to be achieved. We certainly did not want readers to conclude that one should aim for low compliance rates. As it was pointed out, quality control is a fastidious and time-consuming process. A large multicentre trial applying sepsis bundles could also demonstrate only a limited increase in bundle compliance after one training cycle. Nevertheless, there was a measurable effect on outcome. Last but not least, only a few studies about the experience with the ventilator bundles have been published. For this reason, we found it worthwhile to share our experience with the ICU community.