Yale University
Department of Statistics

Monday, April 29, 1996

Susan Murray

Harvard School of Public Health
Dana Farber Cancer Institute

Nonparametric Two-Sample Tests for Survival Data Incorporating Longitudinal Covariates

One of the primary problems facing statisticians who work with survival data is the loss of information that occurs with right censored data. This research considers trying to recover some of this endpoint information through the use of a prognostic covariate which is measured on each individual. Survival estimates which nonparametrically incorporate prognostic covariate information are employed in constructing a test statistic used for making inferences in the two sample censored data problem. This statistic, which focuses on the integrated differences between two survival curves, detects significant differences in years of life saved between two treatments of interest. A special design advantage of test statistics that detect years of life saved is the lack of restrictions of the shapes of the survival curves being compared. Pepe and Fleming studied the performance of this type of statistic when the survival estimates incorporated were estimated using the Kaplan-Meier method. By substituting the Kaplan-Meier survival estimate with the improved survival estimate, the resulting test statistic remains valid when informative censoring is captured by the incorporated covariate.

Seminar to be held in Room 107, 24 Hillhouse@ 4:15 pm