Term
Primary endpoint
The single pre-specified outcome a trial is designed and statistically powered to test.
Before a trial starts, its protocol names one outcome as primary: the measurement that decides whether the trial succeeded. The sample size is calculated so the trial has adequate statistical power for exactly that comparison.
A trial can report favorable secondary outcomes and still have failed, because only the primary comparison was protected from chance and multiplicity by design. This is why evidence grading on this site reads the primary endpoint first, and why a press release that leads with a secondary result deserves suspicion.
On this site: How grades read endpoints, in the methodology ↗