Step-up procedures for non-inferiority tests with multiple experimental treatments

Area of Science:

• Clinical Trials
• Biostatistics
• Pharmaceutical Research

Background:

• Non-inferiority (NI) trials are increasingly utilized to evaluate new treatments against established standards.
• Valid NI trials require assay sensitivity, confirming the standard treatment's efficacy.
• Existing statistical methods for three-arm NI trials often focus on a single new treatment.

Purpose of the Study:

• To extend existing single-step procedures for NI trials with multiple new treatments.
• To introduce two novel step-up procedures for analyzing multi-treatment NI trials.
• To improve statistical power and flexibility in NI trial design.

Main Methods:

• Development of two new step-up statistical procedures.
• Extension of a previously developed single-step procedure.
• Comparative power analysis of the proposed procedures against the single-step method.

Main Results:

• Both proposed step-up procedures demonstrated significantly improved statistical power compared to the single-step procedure.
• One procedure offers flexibility in error rate allocation between sensitivity and NI hypotheses.
• This flexibility allows for potentially reduced placebo group sizes in trial design.

Conclusions:

• The developed step-up procedures offer enhanced statistical power for non-inferiority trials with multiple new treatments.
• These methods provide valuable alternatives to existing single-step procedures.
• The flexible procedure aids in optimizing patient allocation and trial design efficiency.