Hypothesis testing of hazard ratio parameters in marginal models for multivariate failure time data

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Modeling

Background:

• Marginal hazard models are crucial for analyzing multivariate failure time data.
• Existing likelihood-based hypothesis tests have limitations for these models.

Purpose of the Study:

• To extend and evaluate hypothesis test statistics for marginal hazard models.
• To address the inappropriateness of standard tests in this context.

Main Methods:

• Development of generalized Wald, score, and likelihood ratio tests.
• Examination of asymptotic distributions of the proposed statistics.
• Simulation studies to assess finite sample properties.

Main Results:

• Proposed generalized tests are more suitable for marginal hazard models.
• Asymptotic distributions and finite sample properties were analyzed.
• The method demonstrated applicability to real-world health survey data.

Conclusions:

• The generalized hypothesis tests offer a robust approach for marginal hazard models.
• This work provides improved statistical tools for multivariate survival data analysis.
• The findings have implications for epidemiological and health research.