A semiparametric Cox-Aalen transformation model with censored data

Area of Science:

• Biostatistics
• Survival Analysis
• Semiparametric Models

Background:

• Existing transformation models and Cox-Aalen models have limitations in capturing complex covariate effects.
• There is a need for versatile semiparametric models that can incorporate both additive and multiplicative covariate impacts on the hazard function.

Purpose of the Study:

• To propose a novel class of Cox-Aalen transformation models.
• To extend existing transformation and Cox-Aalen models by integrating additive and multiplicative covariate effects.
• To enhance statistical power for discovering covariate effects in survival data analysis.

Main Methods:

• Development of Cox-Aalen transformation models incorporating multiplicative and additive covariate effects.
• Proposal of an estimating equation approach with an expectation-solving (ES) algorithm for parameter estimation.
• Utilization of modern empirical process techniques to establish estimator consistency and asymptotic normality.

Main Results:

• The proposed Cox-Aalen transformation models are shown to be flexible and versatile.
• The expectation-solving (ES) algorithm provides a computationally efficient method for estimation and variance calculation.
• Simulation studies and real-world data applications demonstrate the models' performance and utility.

Conclusions:

• The proposed Cox-Aalen transformation models offer a powerful extension to existing survival analysis techniques.
• The developed methods are statistically sound, computationally efficient, and practically useful.
• These models can significantly enhance statistical power in identifying significant covariate effects, as evidenced in HIV prevention trials.