|
Abstract:
|
In late phase oncology trials, time–to-event endpoints such as overall survival (OS) are often used as the primary endpoint. The timing of interim and final analyses is triggered by a certain number of observed events. Therefore, accurate prediction of the event accrual is critical for a study team to plan activities appropriately and to facilitate decision making. Factors that need to be considered for prediction are data entry lag, observed and assumed event rate, enrollment status, and enrollment rate. In this research, various prediction models, such as exponential model, Weibull model, nonparametric model that combines Kaplan–Meier survival estimation with Bayesian bootstrap resampling, have been compared based on simulation analyses. There is no uniformly best method to obtain the most accurate prediction, but this research will help researchers better understand the pros and cons of each method and will show how to implement the methods in a Pharma industry setting.
|
