In recent years, several factors have led to increasing focus on the meaning of appropriateness of care and clinical performance in the intensive care unit (ICU). The emergence of new and expensive treatment modalities, a deeper reflection on what constitutes a desirable outcome, increasing financial pressure from cost containment efforts, and new attitudes regarding end-of-life decisions are reshaping the delivery of intensive care worldwide. This quest for a measure of ICU performance has led to the development of severity adjustment systems that will allow standardised comparisons of outcome and resource use across ICUs. These systems, for many years used only in the research setting, have evolved to become sophisticated, computer-based decision-support tools, in some instances commercially developed, and capable of predicting a diverse set of outcomes. Their application has broadened to include ICU performance assessment, individual patient decision-making, and pre- and post-hoc risk stratification in randomised trials. In this paper, we review the popular scoring systems currently in use; design issues in the development and evaluation of new scoring systems; current applications of scoring systems; and future directions.