American Statistical Association
In survival analysis, treatment effects are often described through contrasts in the hazard function, despite the fact that most investigators would prefer to know the effect of the treatment on mean lifetime. The majority of existing methods for estimating differences in mean lifetime only apply to settings where treatment is assigned at baseline (time 0). We develop semiparametric methods to estimate the effect on restricted mean survival time of a time-dependent treatment. In the data structure of interest, both an experimental and established form of treatment are available; pre- and post-treatment hazards are non-proportional; subjects may experience periods of treatment ineligibility; and treatment assignment is not randomized. The proposed methods involve weighting results from stratified proportional hazards models, which are fitted using a generalization of case-control sampling. Asymptotic properties of the proposed parameter estimators are derived and evaluated in finite samples through simulation. The proposed methods are applied to estimate the average life-years gained through expanded criteria donor kidney transplantation, using data from a national organ transplant registry.
|Date:||Wednesday, April 22, 2009|
|Time:||4:00 - 5:00 P.M.|
Memorial Sloan-Kettering Cancer Center
Department of Epidemiology and Biostatistics
307 East 63rd Street
(between First and Second Avenues)
3rd Floor Conference Room
New York, New York
Note: To gain access to the building, please follow the directions by the telephone in the foyer.