Parametric versus non-parametric methods for estimating cure rates based on censored survival data.

If a patient's failure time is incorrectly recorded as being too early, the correction will lower the plateau of the Kaplan-Meier curve and, hence, the associated estimated cure rate. Implications of this counter-intuitive observation are discussed. In addition, a parametric approach, based on the Gompertz distribution, to the problem of cure rate estimation is presented.