Although various systems have been developed to identify patients at increased risk of peri- and postoperative mortality and morbidity, little effort has been made in developing tools to reduce this risk. In this issue of Critical Care, Pearse et al. publish two reports related to predicting and improving outcome in high-risk surgical patients. Rather than conducting large, multicentre, randomised, controlled trials, the research group at St George's Hospital in London has persistently and systematically tested the concept of goal-directed haemodynamic management in high risk surgery in their single-centre setting. Their results have been impressive, demonstrating that in this setting, various outcome measures can be reduced with goal-directed haemodynamic management. The impressive positive results of the Pearse studies contrast sharply with the negative results of multicentre studies, such as that of Sandham et al. One reason may be that, like several other successful single-centre trials, Pearse et al. used strict treatment protocols rather than guidelines. In addition, single-centre studies utilize their investigators' knowledge of their patients' risk profiles and familiarity with the care processes and infrastructures of their institutions. An understanding of the organisational and case-mix aspects of pre-, peri-and post-operative management is vital for planning multicentre trials of goal-directed management.