Abstract
Clinical studies in which a major objective is to produce Kaplan–Meier estimates of survival probabilities should be designed to produce those estimates with a desired prespecified precision as measured by their standard errors. By considering the Peto and Greenwood formulae for the estimated standard error of the Kaplan–Meier estimate and replacing their constituents with expected values based on the study's design parameters, formulae for projected standard errors can be produced. These formulae are shown, through simulations, to be quite accurate. Copyright © 2001 John Wiley & Sons, Ltd.