Sample Size Determination in Survival Analysis

This article outlines how the sample size for a study can be determined when the response variable is a survival time. This is a two-stage process in which the number of events required to meet prespecified power requirements is first determined. The occurrence of censoring generally means that not everyone recruited to a study will experience the end point, and so the next step is to calculate the number of individuals who need to be recruited to give the required number of events. This number will depend on parameters such as the anticipated recruitment rate and the follow-up period. The procedure is illustrated using a numerical example. Keywords: sample size; survival data; proportional hazards model