Estimating a Treatment Effect in Residual Time Quantiles under the Additive Hazards Model

Area of Science:

• Biostatistics
• Survival Analysis
• Clinical Trials

Background:

• Traditional time-to-event analyses in clinical trials often rely on the Cox proportional hazards model.
• The proportional hazards assumption may not always hold, necessitating alternative statistical approaches.
• The additive hazards model offers an alternative but its hazard difference is often clinically opaque.

Purpose of the Study:

• To introduce and study residual time quantiles within the additive hazards model framework.
• To translate the hazard difference into a more interpretable difference in residual time quantiles.
• To provide a clinically meaningful interpretation of treatment effects in censored time-to-event data.

Main Methods:

• Estimation of residual time quantiles for a covariate's conditional survival function.
• Utilizing the additive hazards model for survival data analysis.
• Asymptotic properties determination and Monte-Carlo simulations for performance assessment.

Main Results:

• Successful estimation of residual time quantiles.
• Demonstration of translating hazard differences into interpretable quantile differences.
• Application of the method to two real-world randomized clinical trials.

Conclusions:

• Residual time quantiles provide a more intuitive clinical interpretation of treatment effects in additive hazards models.
• This approach enhances the clinical utility of survival analysis, aiding patient counseling and resource planning.
• The method is robust and applicable to real-world randomized clinical trial data.