Time-to-event data often arise in clinical research, and in many cases represent the primary outcome of interest. These data generally represent the elapsed time between a reference time-point (e.g., treatment randomization) and an event of interest (e.g. death, relapse, etc.).
Whereas right censoring is a feature that is easily accommodated by most existing software, the same doesn’t strictly hold for another feature of survival data, left-truncation. In this post we’ll describe what left-truncation is, when it can arise and provide some SAS code that can be used to derive survival estimates and curves.