Assessing consistency in clinical trials with two subgroups and binary endpoints: A new test within the logistic...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacometrics

Background:

• Late-stage drug development requires robust evidence of therapeutic efficacy in diverse populations.
• Subgroup analyses are crucial for marketing authorization, traditionally using interaction tests.
• Interaction tests have limitations, including low power and oversensitivity, potentially leading to clinically irrelevant findings.

Purpose of the Study:

• To propose a novel consistency test for subgroup analyses in drug development.
• To overcome the limitations of conventional interaction tests for binary endpoints.
• To confirm consistency of subgroup-specific treatment effects while controlling Type I error.

Main Methods:

• Development of a consistency test based on the interval inclusion principle for binary endpoints.
• The homogeneity test measures deviation between overall and subgroup-specific effects on the odds ratio scale.
• Comparison with an equivalence test based on the ratio of subgroup-specific effects; assessment via simulation studies and application to relative risk regression.

Main Results:

• The proposed homogeneity test demonstrates sufficient power in realistic scenarios with small interactions.
• Power decreases with unbalanced subgroups, lower sample sizes, and narrower equivalence margins.
• Severe interactions are more likely to be rejected when stronger.

Conclusions:

• The interval inclusion principle provides a robust method for confirming subgroup treatment effect consistency.
• This approach offers advantages over traditional interaction tests in drug development.
• The proposed test is effective in controlling Type I error and identifying relevant treatment effects across subgroups.