Applied Missing Data Analysis in the Health Sciences E-Book

Xiao-Hua Zhou

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Serie: Statistics in Practice
• Sprache: Englisch

Seitenzahl: 377

Veröffentlichungsjahr: 2014

Beschreibung

Applied Missing Data Analysis in the Health Sciences

A modern and practical guide to the essential concepts and ideas for analyzing data with missing observations in the field of biostatistics

With an emphasis on hands-on applications, Applied Missing Data Analysis in the Health Sciences outlines the various statistical methods for the analysis of missing data. The authors acknowledge the limitations of established techniques and provide newly-developed methods with concrete applications in areas such as causal inference.

Organized by types of data, chapter coverage begins with an overall introduction to the existence and limitations of missing data and continues into techniques for missing data inference, including likelihood-based, weighted GEE, multiple imputation, and Bayesian methods. The book subsequently covers cross-sectional, longitudinal, hierarchical, survival data. In addition, Applied Missing Data Analysis in the Health Sciences features:

• Multiple data sets that can be replicated using SAS®, Stata®, R, and WinBUGS software packages
• Numerous examples of case studies to illustrate real-world scenarios and demonstrate applications of discussed methodologies
• Detailed appendices to guide readers through the use of the presented data in various software environments

Applied Missing Data Analysis in the Health Sciences is an excellent textbook for upper-undergraduate and graduate-level biostatistics courses as well as an ideal resource for health science researchers and applied statisticians.

Leseprobe

CONTENTS

Cover

Wiley Series in Statistics in Practice

Title Page

Copyright

Dedication

List of Figures

List of Tables

Preface

Chapter 1: Missing Data Concepts and Motivating Examples

1.1 Overview of the Missing Data Problem

1.2 Patterns and Mechanisms of Missing Data

1.3 Data Examples

Chapter 2: Overview of Methods for Dealing with Missing Data

2.1 Methods that Remove Observations

2.2 Methods that Utilize all Available Data

2.3 Methods that Impute Missing Values

2.4 Bayesian Methods

Chapter 3: Design Considerations in the Presence of Missing Data

3.1 Design Factors Related to Missing Data

3.2 Strategies for Limiting Missing Data in the Design of Clinical Trials

3.3 Strategies for Limiting Missing Data in the Conduct of Clinical Trials

3.4 Minimize the Impact of Missing Data

Chapter 4: Cross-sectional Data Methods

4.1 Overview of General Methods

4.2 Data Examples

4.3 Maximum Likelihood Approach

4.4 Bayesian Methods

4.5 Multiple Imputation

4.6 Imputing Estimating Equations

4.7 Inverse Probability Weighting

4.8 Doubly Robust Estimators

4.9 Code Used in This Chapter

Chapter 5: Longitudinal Data Methods

5.1 Overview

5.2 Examples

5.3 Longitudinal Regression Models for Complete Data

5.4 Missing Data Settings and Simple Methods

5.5 Likelihood Approach

5.6 Inverse Probability Weighted GEE with MAR Dropout

5.7 Extension to Nonmonotone Missingness

5.8 Multiple Imputation

5.9 Bayesian Inference

5.10 Other Approaches

Appendix 5.A: Technical Details of the Approximation Methods for GLMM and Computer Code for the Examples

Chapter 6: Survival Analysis under Ignorable Missingness

6.1 Overview

6.2 Introduction

6.3 Enhanced Complete-Case Analysis

6.4 Weighted Methods

6.5 Imputation Methods

6.6 Nonparametric Maximum Likelihood Estimation

6.7 Transformation Model

6.8 Data Example: Pathways Study

6.9 Concluding Remarks

Chapter 7: Nonignorable Missingness

7.1 Introduction

7.2 Cross-Sectional Data: Selection Model

7.3 Longitudinal Data with Dropout

7.4 Bayesian Analysis for Generalized Linear Models with Nonignorably Missing Covariates

7.5 Multiple Imputation

7.6 Inverse Probability Weighted Methods

Chapter 8: Analysis of Randomized Clinical Trials with Noncompliance

8.1 Overview

8.2 Examples

8.3 Some Common but Naive Methods

8.4 Notations, Assumptions, and Causal Definitions

8.5 Method of Instrumental Variables

8.6 Moment-based Method

8.7 Maximum Likelihood and Bayesian Methods

8.8 Noncompliance and Missing Outcome Data

8.9 Analysis of the Two Examples

8.10 Other Methods for Dealing with both Noncompliance and Missing Data

Appendix 8.A: Multivariate Delta Method

Bibliography

Index

End User License Agreement

List of Tables

Table 1.1

Table 1.2

Table 4.1

Table 4.2

Table 4.3

Table 4.4

Table 4.5

Table 4.6

Table 4.7

Table 5.1

Table 5.2

Table 5.3

Table 5.4

Table 5.5

Table 5.6

Table 5.7

Table 5.8

Table 5.9

Table 5.10

Table 5.11

Table 5.12

Table 5.13

Table 5.14

Table 5.15

Table 6.1

Table 8.1

Table 8.2

List of Illustrations

Figure 1.1

Figure 4.1

Figure 5.1

Figure 5.2

Figure 5.3

Guide

Cover

Table of Contents

Preface

Chapter 1

Pages

ii

iii

iv

v

xv

xvi

xvii

xviii

xix

xx

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

Applied Missing Data Analysis in Health Sciences

Xiao-Hua Zhou

University of Washington

Chuan Zhou

University of Washington

Danping Liu

National Institutes of Health

Xiaobo Ding

Chinese Academy of Sciences

Copyright 2014 by John Wiley & Sons, Inc. All rights reserved.

Published by John Wiley & Sons, Inc., Hoboken, New Jersey.

Published simultaneously in Canada.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate percopy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, [phone], fax [phone], or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, JohnWiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, [phone], fax [phone].

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herin may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services please contact our Customer Care Department with the U.S. at 877-762-2974, outside the U.S. at 317-572-3993 or fax [phone].

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print, however, may not be available in electronic format.

Library of Congress Cataloging-in-Publication Data:

Zhou, Xiao-Hua, author.

Applied missing data analysis in the health sciences / Xiao-Hua Zhou, Chuan Zhou, Danping Liu, Xiaobo Ding.

p.; cm.

Includes bibliographical references and index.

ISBN 978-0-470-52381-0 (cloth)

I. Zhou, Chuan, 1972-author. II. Liu, Danping, 1981-author. III. Ding, Xiaobo, author. IV. Title.

[DNLM: 1. Data Interpretation, Statistical. 2. Biomedical Research–methods. 3. Models, Statistical. 4. Research Design. WA 950]

R853.S7

610.72′4–dc23

2013047830

Dedication

To Yea-Jae, Yi,

Tingting, and Shuqin

List of Figures

1.1

Example 1.6 to explain the missing mechanisms, MCAR, MAR, and MNAR

4.1

Scatter plot of height versus weight

5.1

Forty individuals and mean trajectories

5.2

Diagnostic plot for the imputed

hachin

,

education

,

mmse2

, and

apoee4

5.3

Illustration of the imputation procedures

Preface

With a strong practical emphasis on health science applications, this book describes statistical methods and models for the analysis of data with missing values. We attempt to write so that researchers with experience in applied data analysis, but less technical knowledge than a statistician, should be able to understand and implement most of the methods described. For those with a stronger background in statistics, we provide more technical details as to not detract from the flow of rest of the chapter. We have also tried to choose examples that are relevant to most health science researchers who work in a variety of disciplines.

In all fields of study, missing data are a common problem since, for any data collection process, there are so many things that could go wrong that missing values are all too likely. Thus, when attempts are made to answer the scientific questions of interest, researchers ask the all-too-common question: what do we do with the missing data?

The statistical literature to answer this question is well developed, but overly technical and complicated for researchers who are not experts in statistics and methodology. Therefore, researchers may recognize the existence of missing data, but fail to respond for two reasons: first, they may not understand the consequences of ignoring missing data and how it can impact the validity of their results; second, there is a lack of understanding of the statistical methods for missing data and how to apply them in their own research. Therefore, the purpose of this book is to provide health science researchers with the means of understanding the importance of missing data in their own personal research and the ability to use these methods given the available software.

This book is organized into eight chapters. Chapter 1 introduces concepts on the missing data mechanism and some real-world examples. Chapter 2 gives an overview of methods for dealing with missing data. Chapter 3 describes some design strategies for minimizing the impact of missing data. Chapters 4 and 5 introduce methods for dealing with missing data problems in cross-sectional and longitudinal studies, respectively. Chapter 6 deals exclusively with missing data problems in survival analysis. Whereas Chapters 3–6 deal with ignorable missing data problems, Chapter 7 presents methods for dealing with nonignorable missing data problems. Finally, Chapter 8 discusses methods for dealing with missing data in causal inferences.

As we worked through examples in the book, we chose to provide software code in the text of the chapters as we want to encourage application of these methods after an understanding of the basic theory. We chose to include R code in the text as many of the methods can be implemented in R; in addition, R is also a publicly available software environment (see xix www.r-project.org). Since many researchers also use Stata in addition to R, we include code for some selected examples. All the analysis data sets, together with R and Stata codes used in this book, can be downloaded from http://faculty.washington.edu/azhou/.

X.H. Zhou, C. Zhou, D. Liu, and X. Ding

Seattle, Washington

March, 2014

Chapter 2Overview of Methods for Dealing with Missing Data

Regardless of the existence of missing data, the end result of any analysis is to make valid and efficient inferences about the population of interest. Neyman and Pearson 1933 established valid criteria for evaluating any statistical procedure. These criteria include having a small bias, where bias refers to the difference between the average sample estimate and its true value, and a small variance associated with the average sample estimate (efficiency). Bias and variance can be combined in a single measure called the mean square error (MSE) so that the bias, variance, and MSE describe the behavior of the estimate. Using these criteria, we discuss the various missing data methods that are available, each with its own strengths and limitations. This chapter provides a brief overview of the existing approaches and classifies them according to whether they remove observations with missing data, utilize all available data, or impute missing data. An excellent overview of missing data methods is provided by Schafer and Graham 2002.