MARGINAL MODELS FOR THE ANALYSIS OF LONGITUDINAL MEASUREMENTS WITH NONIGNORABLE NONMONOTONE MISSING DATA

We propose methods for the analysis of continuous responses subject to nonignorable non-monotone missing data. We form a pseudolikelihood by naively assuming independence over time and using a product of margin al likelihoods at each time point, and we obtain consistent and asympt otically normal estimators of the mean and missingness parameters. Our primary interest is in estimating the parameters of the marginal mode l at each time point, and we make no assumption about the correlation structure. An approximate variance is found for the parameter estimate s using a sandwich estimate. We provide an example, and present a simu lation study that assesses the performance of the model.