4 - Non-Inferiority Trials** Flashcards
Define superiority trial
Trial designed to show that tx A is better than tx B
Define non-inferiority trial
Trial to test whether tx A is not worse than tx B, by a prespecified margin
Define equivalence trial
Trial to test whether tx A is equal to tx B
When should a non-inferiority trial be used?
When new tx has other advantages than primary outcomes (ex: route of administration, side effects, monitoring schedule)
What are the hypotheses for a superiority trial?
- Null = no difference in tx
- Alternative = difference in tx
What are the hypotheses for a non-inferiority trial?
- *Opposite to superiority
- Null = difference in tx
- Alternative = no difference in tx
What determines sample size?
- Power
- Type 1 error rate
- Effect size
- Underlying event rate
What is the difference between superiority and non-inferiority trials when determining statistical significance?
- Superiority compare to reference line (0 for comparing difference or 1 for comparing relative risk or hazard risk)
- Non-inferiority compare to NI margin
Should intent to treat or per protocol be used in non-inferiority trials?
- Both
- Results should agree or conclude inferior
Are failed superiority trials the same as non-inferiority trials?
No
Describe the non-inferior checklist
- Was study designed as non-inferiority trial?
- Has clinical effectiveness of standard tx been well established and consistent?
- Is non-inferiority margin defined and its derivation understood?
- Is sample size calculated based on alpha, beta, and non-inferiority margin?
- Was planned sample size achieved in trial?
- Was the result achieved in the active control arm as expected?
- Were ITT and per protocol analysis performed and do they give a consistent answer?
What is an acceptable NI margin?
30%
