Generalised linear models for correlated pseudo‐observations, with applications to multi‐state models

摘要: will be in a given state at some time, is complex nonlinear function of the intensity regression coefficients. We present technique which models probabilities directly. This method based on pseudo-values from jackknife statistic constructed simple summary estimates probabilities. These are then used generalised estimating equation to obtain model parameters. illustrate how this works by studying examples common problems. apply acute graft-versus-host disease bone marrow transplants.