Geert Molenberghs

Parametric models for incomplete continuous and categorical longitudinal data

This paper reviews models for incomplete continuous and categorical longitudinal data. In terms of Rubins classification of missing value processes we are specifically concerned with the problem of nonrandom missingness. A distinction is drawn between the classes of selection and pattern-mixture models and, using several examples, these approaches are compared and contrasted. The central roles of identifiability and sensitivity are emphasized throughout.

An Introduction to (Generalized (Non)Linear Mixed Models

In applied sciences, one is often confronted with the collection of correlated data or otherwise hierarchical data. This generic term embraces a multitude of data structures, such as multivariate observations, clustered data, repeated measurements (called ‘repeated observations’ in this volume), longitudinal data, and spatially correlated data. In particular, studies are often designed to investigate changes in a specific parameter which is measured repeatedly over time in the participating persons. This is in contrast to cross-sectional studies where the response of interest is measured only once for each individual. Longitudinal studies are conceived for the investigation of such changes, together with the evolution of relevant covariates.

A Meta-analytic Validation Framework for Continuous Outcomes

A meta-analytic approach was called for by several authors, e.g., Albert et al. (1998). A first formal proposal, using a Bayesian approach, was given by Daniels and Hughes (1997). Buyse et al. (2000a) extended these ideas using the theory of linear mixed-effects models. Gail et al. (2000) extended it further using generalized estimating equations methodology. In what follows, we describe the approach as proposed by Buyse et al. (2000a).