From the Guidelines
The non-inferiority margin should be calculated by preserving at least 50% of the established treatment's effect, considering both statistical and clinical factors, as recommended by regulatory agencies 1. To determine the non-inferiority margin, start by establishing the minimum clinically important difference (MCID) based on clinical expertise and patient perspectives. Then, review historical data on the standard treatment's effect compared to placebo to establish the treatment effect that must be preserved.
- The margin should be smaller than the MCID and consider both statistical and clinical factors, such as the potential risks of the new treatment and regulatory requirements 1.
- A common approach is to preserve at least 50-80% of the established treatment's effect, as seen in the CONVINCE trial where the noninferiority margin for the composite of stroke, MI, or CV death was set at a hazard ratio of 1.16 1.
- For example, if a standard antibiotic has shown a 10% advantage over placebo for infection clearance, and you want to preserve 70% of this effect, your non-inferiority margin might be 3% (allowing the new treatment to be no more than 3% worse) 1.
- The choice of margin requires a realistic balancing of scientific goals with an achievable sample size, as a smaller margin may require a larger sample size to demonstrate non-inferiority 1.
- Regulatory agencies, such as the U.S. Food and Drug Administration, accept the "putative placebo" approach to establish a noninferiority margin and provide guidance for its determination 1.
- The margin should be established before the study begins and clearly justified in the study protocol to ensure scientific validity and regulatory acceptance 1.
- The sample size of a noninferiority trial is typically determined to ensure a high level of power, with the alpha error rate defined as the probability of concluding noninferiority under the null hypothesis 1.
From the Research
Non-Inferiority Margin Calculation
The non-inferiority margin is a crucial parameter in non-inferiority trials, as it determines whether a new treatment is not significantly worse than an active comparator 2. The calculation of this margin is based on clinical and statistical reasoning, and regulators recommend that it be defined based on historical evidence of the active comparator 2.
Factors Affecting Non-Inferiority Margin Calculation
Several factors can affect the calculation of the non-inferiority margin, including:
- The choice of margin is crucial, as a large margin may lead to the acceptance of a truly inferior treatment 3
- The magnitude of a clinically relevant difference between groups should be carefully considered, taking into account the benefits, harms, costs, and potential side effects of both study interventions 3
- Medical specialty is significantly associated with non-inferiority margin size, with larger margins used in certain specialties 4
Methods of Defining Non-Inferiority Margin
There are several methods for defining the non-inferiority margin, including:
- Historical evidence of the active comparator 2
- Fixed-margin method 5
- Point-estimate method 2
- Synthesis method 2
- The method of defining the margin is often not reported, and the reporting of this critical information is essential for the judgment of non-inferiority trials 5, 6
Reporting of Non-Inferiority Margin
The reporting of the non-inferiority margin is often poor, with low rates of reporting at registration and inconsistent reporting between registration and publication 6. Mandatory reporting of the design and the non-inferiority margin at registration could enhance transparency and favor more reliable results 6.
Non-Inferiority Margin Sizes
The size of the non-inferiority margin can vary, with a median absolute risk difference of 9% and a median relative non-inferiority margin of 1.5 reported in one study 4. The choice of margin size is critical, as it can affect the interpretation of results and the acceptance of a new treatment as non-inferior 3.