ABSTRACT
There are many aspects of the design of a medical study that need to be considered, most of which are common to research programs in many areas of application. In this chapter, methods for determining the number of individuals needed when the response variable is the time to an event, such as the death of a patient, are summarised. The emphasis is on evaluating the sample size required for comparing two groups of survival times using a log-rank test. The occurrence of censored observation times means that it will not usually be possible to determine the event time of every individual recruited to a study. The first step is therefore to determine the number of events that are required for the study to have sufficient power to identify an observed difference in the survival times between two groups as significant at some chosen level. It is then shown how the probability of death over the duration of a study can be used to identify the number of individuals needed to give the required number of events.
