ABSTRACT

The two main reasons to perform a statistical analysis in medicine and epidemiology are to unravel disease etiology and to obtain accurate predictions. The effect of a covariable is often quantified via the hazard or rate, which is an instantaneous quantity. For prediction, which concerns the future, the cumulative probability or risk is the quantity of interest. If there is only a single event type of interest, there is only one hazard, and effects on the hazard translate to qualitatively similar effects on the cumulative probability.