Comparing Methods to Assess Treatment Effect Heterogeneity in General Parametric Regression Models

Area of Science:

• Biostatistics
• Statistical modeling
• Clinical trial analysis

Background:

• Assessing treatment effect heterogeneity is crucial for personalized medicine.
• Parametric regression models are commonly used but require robust methods for heterogeneity assessment.

Purpose of the Study:

• To review and compare methods for assessing treatment effect heterogeneity within parametric regression models.
• To emphasize and evaluate score-residual-based tests for treatment effect heterogeneity.

Main Methods:

• Comparison of standard likelihood ratio tests, bootstrap likelihood ratio tests, and Goeman's global test.
• Focus on score-residual-based tests for treatment effect, including variants.
• Simulation study and illustration in a time-to-event clinical trial.

Main Results:

• Score-residual-based methods demonstrate practical utility, flexibility, and reliability.
• These methods effectively explore treatment effect heterogeneity and modifiers.
• Guidance for decision-making regarding treatment effect heterogeneity is provided.

Conclusions:

• Score-residual-based methods are recommended for assessing treatment effect heterogeneity.
• These approaches offer valuable insights for clinical decision-making.
• The study validates the utility of these methods in real-world clinical settings.