Gamma frailty transformation models for multivariate survival times

Area of Science:

• Statistics
• Biostatistics
• Survival Analysis

Background:

• Multivariate failure time data analysis is crucial in many fields.
• Existing models like the gamma frailty model have limitations.
• There is a need for flexible models that handle marginal distributions effectively.

Purpose of the Study:

• To propose a novel class of transformation models for multivariate failure times.
• To generalize existing frailty models.
• To provide a framework for marginally linear transformation models for each failure time.

Main Methods:

• Development of a new class of transformation models.
• Application of nonparametric maximum likelihood estimation for inference.
• Theoretical analysis of consistency and asymptotic normality of estimators.

Main Results:

• The proposed models generalize the standard gamma frailty model.
• Maximum likelihood estimators are consistent and asymptotically normal.
• Asymptotic variances achieve the semiparametric efficiency bound.
• Simulation studies confirm asymptotic efficiency and good small-sample performance.

Conclusions:

• The proposed transformation models offer a powerful tool for multivariate failure time analysis.
• The estimation procedure is statistically efficient and robust.
• The method is applicable to real-world data, as demonstrated in a cardiovascular study.