A 42-year-old woman with type 2 diabetes mellitus is enrolled in a randomized controlled trial comparing a new SGLT2 inhibitor versus metformin for glycemic control over 12 weeks. The trial protocol specifies that participants will be stratified by baseline HbA1c (≤8% vs >8%) and randomized 1:1 to each arm using computer-generated random number tables. After randomization, 5 participants in the SGLT2 inhibitor arm develop genital mycotic infections and withdraw. The metformin arm reports no withdrawals. The principal investigator decides to replace these 5 participants with new recruits to maintain equal group sizes. What is the primary methodological flaw introduced by this post-randomization replacement strategy?

Explanation

## Methodological Flaw: Post-Randomization Replacement **Key Point:** Replacing participants who withdraw after randomization violates the fundamental principle of intention-to-treat (ITT) analysis and introduces selection bias into the trial. ### Why ITT Principle Matters The intention-to-treat principle requires that: 1. All randomized participants remain in their assigned group for analysis, regardless of adherence or withdrawal 2. Analysis is based on the group to which participants were randomized, not the treatment they actually received 3. Withdrawals and adverse events are documented and analyzed as part of the assigned arm ### The Problem with Replacement Replacing withdrawn participants with new recruits: - **Breaks randomization integrity:** The new recruits were not part of the original randomization sequence and may differ systematically from those who withdrew - **Introduces selection bias:** The decision to withdraw (due to adverse effects) becomes non-random; replacement participants lack this selection pressure - **Inflates efficacy estimates:** Adverse event-related withdrawals are removed from the treatment arm, artificially improving the safety/efficacy profile - **Violates ITT:** The final analysis no longer reflects the real-world experience of the treatment in the randomized cohort ### Correct Approach | Scenario | Correct Action | |----------|----------------| | Participant withdraws after randomization | Keep in assigned arm for ITT analysis; document reason for withdrawal | | Missing outcome data | Use appropriate imputation methods (LOCF, multiple imputation) if <5% missing | | Adverse event leads to discontinuation | Analyze as part of safety profile; do NOT replace | **Clinical Pearl:** The CONSORT statement requires reporting of all randomized participants in a flow diagram, including reasons for withdrawal. This transparency allows readers to assess bias risk. **High-Yield:** ITT analysis is the gold standard for RCT analysis because it preserves the benefits of randomization and reflects real-world treatment effects, including tolerability issues.