Abstract

The use of composite outcomes is becoming widespread in clinical trials. By combining individual outcome measures into a composite, researchers claim a composite can increase statistical precision and trial efficiency, expediting the trial by reducing sample size and cost, and consequently enabling researchers to answer questions that could not otherwise be answered. Another rationale given for using a composite is that it provides a measure of the net effect of the intervention that is more patient-relevant than any single outcome measure. Critics, on the other hand, argue that the use of composites threatens the scientific objectivity of the trial by introducing new risks, and that in practice components are inconsistently defined, unreasonably combined, and inadequately reported. Philosophical scrutiny can help shed light in these disagreements and disentangle ethical and epistemological issues of composites in clinical trials. I examine common use of composites in cardiovascular trials, and show how composite results can be problematic, even misleading, if the proper range of scientific decisions and their consequences are ignored when judging the composite. Philosophies of science that are principled on normative versions of the argument from inductive risk can help explicate issues with composites. I conclude by suggesting ways of remedying composite issues in clinical trials.