On the Repeated Estimation of Survival Time Quantiles for a Time-to-Event Study in Follow-Up Stage

• Activity Number: 360
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Sunhee Kwon Ro*+ and Yu-Lin Chang
• Companies: Onyx Pharmaceuticals and Onyx Pharmaceuticals
• Keywords: survival analysis ; Kaplan-Meier estimate ; median survival time ; minimum censored time ; on-going follow-up

Abstract

For a time-to-event study with closed enrollment, but with on-going follow-up, it is common that the Kaplan-Meier estimates of survival curve and survival time quantiles (mostly median) are generated multiple times as the data gets more mature with increasing number of uncensored events. Often this repeatition of estimation is performed assuming that more events will produce better estimates. However, it is not clear whether this objective is always feasible by repeating the estimation with more mature data. If the Kaplan-Meier estimate of a survival time quantile from current analysis is less or equal to the minimum of the censored survival times, the estimate for the survival time quantile at analyses performed later will not change any more. We present numerical illustration of this. Application on data are presented also.