An alternative summary measure for time-to-event data: residual life?
View Presentation View Presentation
*Jong-Hyeon Jeong, University of Pittsburgh 

Keywords: censoring, competing risks, median, quantile residual life, survival data

The concept of residual life has been used for centuries under the name of life expectancy, so that it is natural in summarizing time-to-event data. The use of residual life may also facilitate more efficient or straightforward communication between patients and physicians in discussing an efficacy of a drug. In this talk, recent development in the theory and application of residual life function in survival analysis will be reviewed. Specifically considered are estimation of quantile residual lifetime, two-sample comparison of quantile residual lifetimes, and regression on quantile residual life function to control for confounding factors. An extension to competing risks data will be also discussed. The proposed methods will be applied to breast cancer data with long-term follow-up.