Primary Goal of Non-Inferiority Trials
The primary goal of a non-inferiority trial is to demonstrate that a new experimental treatment is not worse than an established active control by more than a prespecified small amount (the non-inferiority margin), typically when the new treatment offers advantages in safety, cost, convenience, or tolerability that may justify accepting a small potential loss in efficacy. 1, 2
Core Objective
- Non-inferiority trials aim to show that the experimental treatment is not unacceptably worse than the active control, rather than proving superiority 1
- This is achieved by demonstrating that the lower bound of the confidence interval for the treatment effect lies above the non-inferiority margin 1, 2
- The design is used when placebo-controlled trials would be unethical because an effective standard treatment already exists 3, 4
Clinical Context and Rationale
Non-inferiority trials are undertaken when the experimental treatment is expected to provide non-efficacy benefits that outweigh a potential small loss in efficacy, including: 1, 5
- Reduced toxicity or better safety profile 1
- Lower cost 1
- Easier administration or simpler regimen 1, 4
- Shorter treatment duration 4
A critical caveat: Non-inferiority designs should not be used when non-efficacy benefits are absent, as loss in efficacy cannot be justified without compensatory advantages 5
The Non-Inferiority Margin
- The non-inferiority margin (delta) represents the maximum acceptable difference in treatment effect that would still allow the new treatment to be considered clinically acceptable 1, 2
- This margin must be prespecified before viewing results to avoid inflating the alpha error rate 1
- The margin should account for both statistical considerations (based on historical evidence of the active control's effect) and clinical considerations (the fraction of the standard treatment's effect that must be preserved) 4, 6
Statistical Framework
The hypothesis testing in non-inferiority trials is reversed from traditional superiority trials: 2
- Null hypothesis: The new treatment is inferior by at least the margin
- Alternative hypothesis: The treatments are equivalent or the difference is less than the margin
Common Pitfalls to Avoid
Major deficiencies identified in published non-inferiority trials include: 1
- Claiming non-inferiority when statistical criteria were not met (occurred in 10% of oncology trials reviewed) 1
- Failing to prespecify the non-inferiority margin (occurred in 7% of trials) 1
- Switching from superiority to non-inferiority after viewing results (only 6% appropriately planned this in advance) 1
- Using margins potentially larger than the assured effects of active controls 1
- Relying solely on intention-to-treat analysis, which can bias toward finding non-inferiority when high dropout rates occur 1, 2
Both intention-to-treat AND per-protocol analyses should be conducted to ensure robust conclusions, as ITT analysis alone may lead to spurious non-inferiority conclusions 1, 2