Keywords: Composite endpoints, win ratio, competing risk
Composite endpoints are created to facilitate interpretation of clinical trial results by reducing the number of comparisons (Type 1 Error), increase power and having a smaller sample size but could also obscure differences in different directions. One key goal of creating composite endpoint is to reduce the repeated measures from each subject to a single outcome and this simplistic approach leads to the use of univariate tests. Many of these procedures assume data are missing completely at random (MCAR) and are inappropriate in longitudinal trials.
In most cardiovascular trials, composite endpoints do report patient’s first event, this has some inherent limitations. We shall discuss the concept of ‘win ratio’ and competing risk to and demonstrate their use with a clinical data. In this presentation, we will discuss when to choose a composite endpoint and their strength and weakness. Most patients reported outcomes like Health Related Quality of Life include multiple domains or subscales. Example of these includes Physical Component Scale (PCS) and the Mental Component Scale (MCS) of the SF-36. Use of different weights that place different emphasis on each of the domains or subscale will be discussed.
Examples of the use of composite endpoints will be provided for selected therapeutic areas like cardiovascular, neurosciences, respiratory and immunology.