Question 2 of 9 Advanced Scenario
How would you select a primary endpoint?
Short answer
Choose a primary endpoint that directly answers the trial's main clinical question and can be measured reliably at a prespecified time. Define exactly what constitutes the outcome, its clinical relevance and how intercurrent events and missing observations will be handled before looking at results.
Why
I would start with the condition, target population, intervention and intended benefit. An endpoint needs clinical meaning as well as measurement reliability and feasibility. For a symptom-treatment trial, patient-reported severity at 12 weeks may be more relevant than a convenient blood biomarker, provided the instrument is valid and the timeframe matches the treatment effect.
Specify the variable, aggregation or comparison metric, follow-up time and any adjudication or masking, plus the estimand framework for events such as discontinuation or rescue treatment. Consider plausible effect size and variability for a justified sample size, and define multiplicity handling for co-primary or multiple important outcomes. Missing outcome risk must be addressed prospectively.
A surrogate endpoint may be useful but is not automatically a validated substitute for patient benefit. A post hoc switch to a more favorable endpoint threatens credibility unless appropriately documented and interpreted.
A strong answer covers
- Discuss measurement timing, statistical power and multiplicity