|Statistical power||For adequate power, trials with a single endpoint may be too large to be practical.|
Including several endpoints increases event frequency and may increase power.
|Selecting components of the composite||A single endpoint may not fully document treatment effect.|
Composite endpoint captures several clinical aspects.
Choose components so overall effect is meaningful.
|Evaluating treatment effect||Interpretation can be difficult:|
|Analysis considerations||Emphasis usually on time to first event, which is often less serious (e.g., hospitalization) than subsequent events (e.g., death).|
Always show results for each separate component.
Methods exist for repeat events and for competing risks, but complexity is a challenge.
Win ratio method prioritizes events of greater clinical importance.