Title: Semiparametric methods for incomplete binary longitudinal data with dropouts
Authors: Sanjoy Sinha - Carleton University (Canada) [presenting]
Abstract: Some semiparametric methods are discussed for joint estimation of the regression parameters and association parameters in binary longitudinal models when the marginal mean response function is partially linear. We propose a spline regression method in the framework of the weighted generalized estimating equations (GEEs) for the simultaneous estimation of the unknown nonlinear function, regression and association parameters under the assumption of a missing at random (MAR) dropout mechanism. Empirical properties of the proposed estimators are investigated using Monte Carlo simulations. An application is also provided using actual longitudinal data from a clinical study, where the data show strong evidence of a nonlinear trend in the mean response function.