Understanding and interpreting composite outcomes in clinical research

Abstract Composite outcomes combine multiple endpoints into a single measure. They can capture overall treatment effects and increase statistical power, but have important interpretive challenges. Components of the outcome may vary in clinical importance, frequency, and treatment effect. This Progress Note describes three types of composite outcomes used in clinical research, discusses challenges in interpretation, including when outcomes are dominated by less important components and when treatment effects diverge across components. Practical guidelines for designing, analyzing, and interpreting studies that use composite endpoints are presented. Understanding these principles will help clinicians critically evaluate studies with composite outcomes.