In a randomized trial of a new antidiabetic drug, some patients allocated to the drug stop taking it because of side effects, and some placebo patients start metformin. To preserve the validity of the randomization when estimating the drug's real-world effectiveness, the analysis should be:
- A Per-protocol analysis
- B As-treated analysis
- C Intention-to-treat analysis ✓
- D Subgroup analysis by adherence
Correct answer: C. Intention-to-treat analysis
Explanation
Intention-to-treat analysis keeps every randomized participant in the arm to which they were assigned regardless of adherence, preserving the prognostic balance created by randomization and reflecting pragmatic effectiveness. Per-protocol and as-treated analyses analyze only compliant participants, which reintroduces selection bias because non-adherence is usually related to tolerability and severity. Subgroup analysis by adherence does not restore randomization and invites confounding by indication.
Reference: Gordis Epidemiology, 6th ed.
High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP
Written and medically reviewed by the StethoPrep medical team.