Retrieved-Dropout-Based multiple imputation for time-to-event data in cardiovascular outcome trials

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Survival Analysis

Background:

• Multiple imputation is crucial for handling missing data in clinical trials.
• Retrieved-dropout multiple imputation has been applied to continuous endpoints for treatment policy estimands.
• Extending this method to time-to-event data is a significant methodological advancement.

Purpose of the Study:

• To extend retrieved-dropout-based multiple imputation to time-to-event endpoints.
• To provide a practical implementation guide for this advanced statistical method.
• To compare the performance of this method against established techniques like Cox proportional hazards models.

Main Methods:

• Application of retrieved-dropout multiple imputation to time-to-event data.
• Utilizing data from study completers and dropouts within the same treatment arm.
• Comparison with Cox proportional hazard and reference-based multiple imputation using a cardiovascular outcome trial dataset.

Main Results:

• The extended retrieved-dropout multiple imputation method is feasible for time-to-event endpoints.
• Demonstration of practical implementation using real-world clinical trial data.
• Comparative analysis highlights the utility and potential advantages over existing methods.

Conclusions:

• Retrieved-dropout multiple imputation offers a viable approach for analyzing time-to-event data with treatment discontinuation.
• This methodology enhances the ability to address the treatment policy estimand in survival analysis.
• The study provides valuable insights for statisticians and researchers in clinical trial design and analysis.