Analysis Pitfalls & Trial Integrity

This lesson is part of Premium

Intention-to-treat (preserves randomisation, conservative for superiority) vs per-protocol. Regression to the mean. Composite endpoints (can be driven by softest component). Surrogate vs patient-oriented outcomes (POEMs vs DOEs). Subgroup analysis dangers (multiplicity). Loss to follow-up thresholds (>20% threatens validity). Last-observation-carried-forward. Exam tip: ITT analysis is the default for RCTs; dropping non-compliers breaks randomisation.

Premium unlocks all 40 lessons, every calculator and the timed mock.