Justification for Determining the Delta Should Be Provided in the Noninferiority Trial
The most appropriate recommendation to improve the proposed study design is that justification for determining the delta (noninferiority margin) of 12% should be provided. This is critical because the noninferiority margin directly impacts the clinical interpretation of results and patient outcomes related to VTE prevention following hip surgery.
Analysis of Study Design Issues
Noninferiority Margin (Delta)
- The prespecified delta of 12% lacks justification in the study design
- Noninferiority margins should be:
- Clinically meaningful and justified based on previous studies
- Small enough to ensure that any difference less than the margin would be clinically acceptable
- Based on the preservation of a substantial fraction of the active control effect 1
Other Design Elements
Alpha Level (Type I Error)
- The study correctly used a 2-sided alpha of 5%
- While some argue that a 1-sided alpha should be used in noninferiority trials, the 2-sided approach is more conservative and provides additional information about potential superiority 2
Analysis Population
Statistical Power
- The power of 90% is appropriate and exceeds the minimum standard of 80%
- Higher power reduces the risk of Type II error (failing to detect a true difference)
- The sample size calculation appears adequate (803 needed, 792 enrolled)
Missing Data Handling
- Last observation carried forward (LOCF) was used, which is acceptable but not optimal
- More sophisticated methods like multiple imputation might be preferable but wouldn't fundamentally change the study validity
Clinical Implications of Noninferiority Margin Selection
The choice of noninferiority margin has direct implications for patient outcomes:
- A margin that is too wide might allow approval of a less effective anticoagulant, potentially leading to increased VTE events
- VTE following hip surgery carries significant morbidity and mortality risks, with DVT occurring in up to 37% of hip fracture patients without prophylaxis 1
- Previous noninferiority trials of anticoagulants (like dabigatran) used carefully justified margins to ensure clinical relevance 2
Recommendations for Improvement
Provide clear justification for the 12% noninferiority margin based on:
- Historical data on the active comparator's effect versus placebo
- Clinical judgment about what constitutes a clinically meaningful difference
- Statistical considerations to preserve a substantial portion of the active control effect
Consider complementary analyses:
- Include both ITT and per-protocol analyses to provide robust evidence
- Consider sensitivity analyses with different approaches to missing data
Document the clinical rationale:
- Explain why the chosen margin would preserve sufficient clinical benefit
- Address how the margin relates to patient-important outcomes like symptomatic VTE, major bleeding, and mortality
Common Pitfalls in Noninferiority Trials
- Selecting arbitrary noninferiority margins without clinical or statistical justification
- Using only ITT or per-protocol analysis rather than both
- Inadequate reporting of the rationale for key design decisions
- Insufficient attention to the quality of the active control implementation
By providing proper justification for the noninferiority margin, researchers ensure that any conclusions about the new anticoagulant's efficacy will be clinically meaningful and protect patients from receiving inferior treatments for VTE prevention after hip surgery.