ipcwswitch

ipcwswitch implements inverse probability of censoring weighting (IPCW) to adjust time-to-event analyses in randomized clinical trials with treatment crossover, enabling estimation of treatment effects such as overall survival under a hypothetical scenario of adherence to initial randomization.

Key Features:

• Handling Crossover Events: Manages treatment switches in randomized clinical trials to reduce bias in estimates of overall survival and other time-to-event outcomes.
• IPCW Weighting: Implements the inverse probability of censoring weighting method by assigning weights based on the inverse probability of remaining uncensored due to crossover.
• Time-dependent Covariate Support: Accommodates time-dependent covariates with varying measurement times across patients within the IPCW framework.
• Survival Estimate Recalibration: Recalibrates survival estimates to reflect a hypothetical scenario where participants remain on their randomized treatments.
• R Package Implementation: Provides the IPCW computational implementation within an R package.

Scientific Applications:

• Survival Analysis: Adjusts survival analyses in randomized clinical trials affected by treatment crossover to derive less biased effect estimates.
• Time-to-Event Outcomes: Handles complex time-to-event data structures, including analyses involving time-dependent covariates.
• Molecular-Profiling Trials (SHIVA01): Applies to trials such as SHIVA01 that evaluated therapy benefits using tumor molecular profiling in advanced cancers.

Methodology:

Uses the inverse probability of censoring weighting (IPCW) approach by weighting individuals with the inverse probability of remaining uncensored due to crossover and applying those weights to recalibrate survival estimates, with support for models including time-dependent covariates with varying measurement times.