Testing for interaction in studies of noninferiority

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Pharmacoeconomics

Background:

• Active-controlled clinical trials aim to establish noninferiority of new treatments versus standard ones.
• Stratification of subjects (e.g., by site, demographics) is common, leading to interest in stratum-specific noninferiority.

Purpose of the Study:

• To evaluate interaction tests for assessing heterogeneity of treatment effects in noninferiority trials.
• To present and compare five distinct analysis strategies for testing treatment effect differences across strata.

Main Methods:

• The study proposes five analysis strategies, adapted from those used for superiority trials.
• Simulation studies are conducted to quantify power and Type I error rates under various scenarios.
• An illustrative example demonstrates the application of the proposed tests.

Main Results:

• The power of interaction tests in noninferiority trials is generally low.
• No single analysis strategy is universally optimal across all parameter configurations.
• The proposed tests are most suitable for descriptive or exploratory purposes.

Conclusions:

• Interaction tests can be adapted for noninferiority trials to explore treatment effect heterogeneity.
• The choice of strategy depends on specific trial parameters and objectives.
• Extensions to equivalence testing are also considered, highlighting the nuanced application of these statistical tools.