In a randomized controlled trial (RCT), the researchers design their study to measure a particular outcome, which is the overall result or main point of the study.
The primary outcome is used as the basis for determining the sample size of the study. When a study is “powered” for a particular outcome, this means that the study was designed with a certain sample size to measure statistical differences for that primary outcome. For this reason, outcomes besides the primary outcome in an RCT should be interpreted carefully.
Secondary outcomes are usually related to the primary outcome and provide additional information.
For example, an RCT might be looking at outcomes from day 3 vs day 5 embryo transfers. The primary outcome might be the live birth rate, with secondary outcomes being clinical pregnancy rates, miscarriage rates, or multiple pregnancy rate.
According to Pocock et al. (2016), if the primary outcome doesn’t show statistical significance, the secondary outcomes should only be used in providing potential directions for further research. However, there can be exceptions to this, particularly when there is a very small p value, there’s a big difference in the outcome and it’s consistent with other studies.