General Departmental Seminar Series
Regression Modeling of Semi-Competing Risks Data
Limin Peng, Graduate Student
Department of Statistics, UW-Madison
Friday, April 2, 2004, 12-1 p.m.
132 WARF Building, 610 Walnut St.
Semi-competing risks data are often encountered in clinical trials with intermediate endpoints subject to dependent censoring from informative dropout. Unlike with competing risks data, dropout may nor be dependently censored by the intermediate event. There has recently been increased attention to this data, in particular, inferences without covariates. In this work we incorporate covariates and formulate their effects via a functional regression model. To accomodate nformative censoring, a time-dependent copula model is proposed. New parameter estimators for the marginal and dependence models are derived from estimating equations and are shown to be uniformly consistent and to converger weakly to Gaussian process. Hypothesis tests are developed accordingly, as are inferences for parametric submodels for he time-varying covariate effects and copula parameters. Simulations and an AIDS data analysis demonstrate that the medthods perform well with realistic sample sizes.