Newcastle. However, this statement ignores that the P value was of 0.4413, which statistically denies the existence of any trend. So, we are left with a (small) hope and with a (relatively big) frustration. The hope is demonstration that some quality parameters related to management of patients with decompensated cirrhosis can be improved by providing a care bundle, although the relevance of improving some indirect indicators remains conjectural. The frustration comes from seeing that despite the effort done, implementing a dedicated policy reinforcing early delivery of evidence-based care did not improve hard end-points. The lesson for the future is that probably we ought to select better the centres where we need to intervene, concentrating the effort where performance indicators, such as in-hospital mortality, are unusually high.