Outcomes and Endpoints

A quality outcome or endpoint will be clinically meaningful, straightforward to collect, and will be sensitive to address research questions of interest. It is also crucial that the measure be feasible to be reliably ascertained in all study participants within the study time frame. Other important considerations in the choice of an outcome are bias (such as self-reporting bias), loss to follow-up, and missing data. The primary endpoint should follow directly from the study's specific aim and primary research question. It dictates the data collection and statistical analysis plan, including calculations for sample size or power. Secondary endpoints can also be chosen to answer secondary research questions or generate hypotheses for future studies. In clinical trials and other prospective studies, types of outcomes include clinical, surrogate, patient-reported, and composite outcomes. In retrospective observational studies, there are often constraints in the choice of endpoint based on measures in existing electronic health record (EHR) data or cohort studies.