Switching noninferiority and superiority trial (ENG)

When conducting superiority or non-inferiority clinical trials, there may be cases where, depending on the analysis results, you change from a superiority trial to a non-inferiority trial, or from a non-inferiority trial to a superiority trial.

Statistically, since both superiority and non-inferiority trials are closed tests, there is no problem with making such changes. However, when changing from a superiority trial to a non-inferiority trial, issues may arise regarding the justification of the non-inferiority margin. Therefore, it’s advisable to pre-specify the non-inferiority margin during the planning stage of the superiority trial. At this point, the non-inferiority margin should include information demonstrating superiority over placebo.

Conversely, when changing from a non-inferiority trial to a superiority trial, there are few difficulties, but superiority trials emphasize the analysis of the full analysis set derived from intention-to-treat (ITT) analysis. Therefore, switching to a superiority trial is possible when ITT analysis has been given high importance in advance.

Additionally, although not directly related, there is often confusion regarding the setting of type 1 error values. In two-sided tests, the type 1 error is set at 0.05, while in one-sided tests, it’s set at 0.025. To conclude, the sample size calculations end up being the same whether you use a type 1 error of 0.05 for two-sided tests or 0.025 for one-sided tests. This follows the ICH E9 guideline.

Reference : Points to Consider on Switching between Superiority and Non-inferiority

Source: Original Korean article | English translation by Claude 3.5 Sonnet (New)