Analysis of Multiple Endpoints—Nonparametric and Semi-Parametric Methods

Abstract

In the past 15 years, various methods have been developed for the analysis of multivariate observations. In this talk, we will concentrate on the non-parametric and semi-parametric methods. For the two sample problem, we will discuss Wei-Johnson (Biometrika, 1985) procedure and its extension to the case with non-randomly missing data (Glidden and Wei, JASA, 1999; Yao, Wei and Hogan, 1998, Biometrika). For regression problems, we will discuss the mixed- effects models and marginal models (generalized estimating equations techniques, Liang and Zeger, 1986). For possibly censored event time data, we will discuss cases for recurrent events, clustered failure times and multiple endpoints of different types (Wei, Lin and Weissfeld, JASA, 1989; Lee, Wei and Amato, 1992; Lin and Wei, JASA, 1992; Lee, Wei and Ying, JASA, 1993; Lin, Wei and Ying, Biometrika, 1998).