Some Issues of Sample Size Calculation for Time-to-Event Endpoints Using the Freedman and Schoenfeld Formulas.

This article deals with seven special issues related to the assumptions, applicability, and practical use of formulas for calculating power or sample size, respectively, for comparative clinical trials with time-to-event endpoints, with particular focus on the well-known Freedman and Schoenfeld methods. All problems addressed are illustrated by numerical examples, and recommendations are given on how to deal with them in the planning of clinical trials.

Sample sizes for time-to-event endpoints: should you insure against chance variations in accrual?

This note addresses the questions of whether one should safeguard against a potential loss of power due to random variations of the accrual time, how this "insurance" can be formulated, and how much the sample size needs to be increased to obtain it.