What is the method for calculating the non-inferiority margin in clinical trials?

From the Guidelines

To calculate a non-inferiority margin, you should first determine the established treatment effect of the active control compared to placebo from historical data, then decide what proportion of this effect must be preserved by the new treatment, typically preserving 50% of the active control's effect, as recommended by regulatory agencies 1. When calculating the non-inferiority margin, several factors should be considered, including:

• The established treatment effect of the active control compared to placebo from historical data
• The proportion of this effect that must be preserved by the new treatment
• The clinically acceptable loss of efficacy
• The confidence interval of the historical treatment effect For example, if a standard treatment reduces mortality by 10% with a 95% confidence interval of 5-15%, and you want to preserve at least 50% of the effect, your non-inferiority margin would be 2.5% (half of the lower bound) 1. The choice of margin requires a realistic balancing of scientific goals with an achievable sample size, as the trial size is inversely proportional to the square of the margin delta 1. Regulatory agencies often have specific guidelines for different therapeutic areas, so consulting these is advisable, and the margin should be determined before the study begins and clearly justified in the protocol based on both statistical and clinical considerations 1. Some methods for selecting a noninferiority margin include:
• The conventional method, which selects a margin based on a minimal clinically relevant difference
• The effect retention method, which selects a margin based on the proportion of the effect of the active control that is preserved by the new treatment
• The 95%-95% method, which selects a margin based on the 95% lower confidence limit of the effect of the active control 1.

From the Research

Methods for Calculating Non-Inferiority Margin

• The fixed-margin approach is a commonly used method, which consists of two steps: calculating the lower limit of the 1 - α * two-sided confidence interval (CI) of the active-control effect versus placebo, and obtaining the non-inferiority margin as a fraction of the lower confidence limit of the control effect to preserve partial control effect 2.
• An alternative method is to use the point estimate, instead of the lower confidence limit, of the active-control effect 2, 3.
• The International Conference on Harmonization (ICH) recommends that the non-inferiority margin should be chosen in such a way that if the non-inferiority of the test therapy to the active control agent is claimed, the test therapy is not only non-inferior to the active control agent, but also superior to the placebo, and variability should be taken into account 4.

Factors Affecting Non-Inferiority Margin

• Medical specialty is significantly associated with non-inferiority margin size, with larger margins used in certain specialties such as thrombosis, infectious diseases, or oncology 5.
• The size of the non-inferiority margin can also depend on the type of outcome being measured, with larger margins used for outcomes involving mortality 5.

Examples of Non-Inferiority Margin Calculation

• A systematic review of non-inferiority trials found that the overall median non-inferiority margin was an absolute risk difference of 9% (IQR 4.2%-10%), and the median relative non-inferiority margin was 1.5 (IQR 1.3-1.7) 5.
• An example of a non-inferiority trial in oncology used a proposed method to determine an alternative non-inferiority margin, which maintained the target Type 1 error rate 2.