Composite outcomes and making sense of the ‘win ratio’


Take Home Messages
  • Composite outcomes are commonly used in clinical trials to reduce sample size and increase statistical power.
  • The traditional approach to evaluating composite outcomes fails to account for relative clinical priority and recurrent events.
  • Hierarchical composite endpoints have been devised to overcome this, with the ‘win ratio’ being a method of analysis that uses pairwise comparisons.
  • Understanding its advantages and limitations is important for clinicians as the adoption of the win ratio increases.

Global improvements in healthcare provision over time has resulted in event rates in clinical trials, such as mortality and hospitalisations, to decline in frequency. Adequately powering these studies to detect meaningful differences between treatment arms (if a true difference exists) therefore mandates recruiting larger sample sizes at the expense of time, labour, and costs. Using composite as opposed to single endpoints has been popularised to overcome this barrier. The development of hierarchical endpoints and the method of the ‘win ratio’ has increasingly gained favour as a more sensitive way to evaluate composite outcomes. We discuss the nuances of composite endpoint analysis, how the win ratio works, and review examples of recent clinical trials using this methodology.