ABSTRACT
Incidence vs. Kaplan-Meier ......................................................................561 30.4 Example from Real Data: Squamous Cell Carcinoma ..............................563 30.5 Discussion ..................................................................................................564 References ..............................................................................................................564
In many fields of medical research, time-to-event end-points are used to assess the potential efficacy of a treatment. The outcome of interest associated with some of these end-points may be a particular type of failure, and it is often of interest to estimate the probability of this failure by a specified time among a particular population of patients. For example, the occurrence of end-stage renal failure is an outcome of interest among patients with insulin-dependent diabetes mellitus (IDDM). Given a sample drawn from the population of patients with IDDM, one may therefore wish to obtain an estimate of the probability of developing end-stage renal failure. As other examples, consider the probability of death due to prostate cancer among patients afflicted with this disease and the probability of the occurrence of CMV retinitis among patients with AIDS.
