Generalized Linear and Nonlinear Models for Correlated Data: Theory and Applications Using SAS

Part IIIFurther Topics

6Missing Data in Longitudinal Clinical Trials

6.1 Background

6.2 Missing data mechanisms

6.3 Dropout mechanisms

6.4 Methods of analysis under MAR

6.5 Sensitivity analysis under MNAR

6.6 Missing data—case studies

6.7 Summary

In Chapters 2-5, a number of examples from different disciplines are presented illustrating the types of models and methods of analysis that are available in SAS for analyzing correlated response data. Particular emphasis is placed on applications requiring the use of generalized linear and nonlinear models, both marginal and mixed. Methods of estimation and inference for these models may be classified as either semiparametric or parametric in nature. The semiparametric methods are moments-based and ...