Our aim, when evaluating many events, is to assess overall survival S(t) but also event free survival EFS(t). In addition, we often wish to estimate the respective contribution of each event involved in event free survival and to describe the distribution of the time of occurrence of each event by breaking down EFS into its different components. However, these different events are often dependent and/or exclusive. Appropriate statistic tools, named competing risk analyses are then required. The aim of this article is to define situations necessitating competing risk analyses as opposed to more simple alternatives which, although often used, are not always appropriate. First two examples are presented to illustrate the problems we face when studying many events. Statistical methods used to compute competing risks are then developed, as is the type of interpretation that can be given to these results.