1. Edition - March 2007
83.90 Euro
2007. 526 Pages, Hardcover
ISBN-10: 0-470-84981-9
ISBN-13: 978-0-470-84981-1 - John Wiley & Sons






Sample Chapter

Short description
Since the early 70s, research into statistical analysis of data with missing values has increased considerably. This is in someway due to advances in computing power, which have lead to previously arduous numerical calculations becoming very simple. Statistical techniques that can handle missing data are particularly important in analyzing clinical trials, as missing values are so prevalent. The first book to concentrate on analyzing data from clinical trials, Missing Data in Clinical Studiestakes a practical approach to the subject. Featuring examples and case studies, the book includes a chapter on software, with discussion of implementation of the techniques in various different statistical software packages.

From the contents
Preface.

Acknowledgements.

I Preliminaries.

1 Introduction.

2 Key Examples.

3 Terminology and Framework.

II Classical Techniques and the Need for Modelling.

4 A Perspective on Simple Methods.

5 Analysis of the Orthodontic Growth Data.

6 Analysis of the Depression Trials.

III Missing at Random and Ignorability.

7 The Direct Likelihood Method.

8 The Expectation-Maximization Algorithm.

9 Multiple Imputation.

10 Weighted Estimating Equations.

11 Combining GEE and MI.

12 Likelihood-Based Frequentist Inference.

13 Analysis of the Age-Related Macular Degeneration Trial.

14 Incomplete Data and SAS.

IV Missing Not at Random.

15 Selection Models.

16 Pattern-Mixture Models.

17 Shared-Parameter Models.

18 Protective Estimation.

V Sensitivity Analysis.

19 MNAR, MAR, and the Nature of Sensitivity.

20 Sensitivity Happens.

21 Regions of Ignorance and Uncertainty.

22 Local and Global Influence Methods.

23 The Nature of Local Influence.

24 A Latent-Class Mixture Model for Incomplete Longitudinal Gaussian Data.

VI Case Studies.

25 The Age-Related Macular Degeneration Trial.

26 The Vorozole Study.

References.

Index.