Data analysis with competing risks and intermediate states / Ronald B. Geskus.

Format:

Book

Author/Creator:

Geskus, Ronald Bertus, 1962-

Contributor:

Tom Farinholt and Blair Edlow Memorial Book Fund.

Series:

Chapman & Hall/CRC biostatistics series

Language:

English

Subjects (All):

Random operators.

Stochastic analysis.

Physical Description:

xxix, 247 pages : illustrations ; 25 cm.

Place of Publication:

Boca Raton : Taylor & Francis, [2016]

Summary:

Data Analysis with Competing Risks and Intermediate States explains when and how to use models and techniques for the analysis of competing risks and intermediate states. This practical and thorough book covers the most recent insights on estimation techniques and discusses in detail how to interpret the obtained results. It extends your knowledge and skills in the analysis of standard time-to-event data to the competing risks and multi-state settings. Although the book uses examples from the biomedical and epidemiological fields, you can easily translate individuals and diseases to units and phenomena from your own research areas. Features, Emphasizes the translation of research questions into modeling choices and the interpretation of results, Describes interpretation in relation to the type of hazard and cumulative probability, enabling you to better understand results of the analysis, Shows how to use standard software for estimating the subdistribution hazard, Illustrates the use of a stacked data set that combines all transitions, which allows for more power and flexibility in model choice with rare event types or transitions, Summarizes which type of quantity is best suited to answer a specific research question and under what assumptions it can be estimated, Includes computer practicals that encourage you to employ the methods in practice, Offers answers to all theoretical exercises as well as hints and detailed answers with R code for the computer practicals on the book's website Book jacket.

Contents:

1 Basic Concepts 1

1.1 Introduction 1

1.2 Examples 2

1.2.1 Infection during a hospital stay 2

1.2.2 HIV infection 3

1.2.3 Bone marrow transplantation 6

1.3 Data structure 7

1.3.1 Time scales 7

1.3.2 Right censored data 8

1.3.3 Left truncated data 10

1.4 On rates and risks 13

1.5 Non-informative observation schemes? 15

1.5.1 Some possible solutions 20

1.6 The examples revisited 23

1.6.1 Infection during a hospital stay 23

1.6.2 HIV infection 24

1.6.3 Bone marrow transplantation 28

1.7 Notation 28

1.8 Basic techniques from survival analysis 30

1.8.1 Maui concepts and theoretical relations 30

1.8.2 The Kaplan-Meier product-limit estimator 32

1.8.2.1 Confidence intervals 34

1.8.3 Nonparametric group comparisons 36

1.8.4 Cox proportional hazards model 37

1.8.5 Counting process format 40

1.9 Summary and preview 43

1.10 Exercises 44

1.11 R code for classical survival analysis 50

1.11.1 The aidssi data set 50

1.11.2 Define time and status information 51

1.11.3 Perform calculations 51

1.11.4 Summary of outcome 52

1.11.5 Log-rank test 55

1.12 Computer practicals 56

2 Competing Risks; Nonparametric Estimation 59

2.1 Introduction 59

2.2 Theoretical relations 60

2.2.1 The multi-state approach; cause-specific hazards 60

2.2.2 The sub distribution approach 62

2.3 Estimation based on cause-specific hazard 63

2.4 Estimation: the sub distribution approach 68

2.4.1 Estimation with complete follow-up 70

2.4.2 A special choice for Γ and Φ 71

2.4.3 The ECDF and PL forms 73

2.4.3.1 The ECDF form 74

2.4.3.2 The PL form 74

2.4.4 Interpretation of the weighted estimators 76

2.5 Standard errors and confidence intervals 79

2.6 Log-rank tests and other subgroup comparisons 83

2.7 Summary; three principles of interpretability 85

2.8 Exercises 88

2.9 Software 92

2.9.1 Nonparametric estimation of F_k 92

2.9.1.1 The Aalen-Johansen form 92

2.9.1.2 The weighted product-limit form 96

2.9.2 Log-rank tests 100

2.10 Computer practicals 101

3 Intermediate Events; Nonparametric Estimation 105

3.1 Introduction: multi-state models 105

3.2 Main concepts and theoretical relations 107

3.2.1 Basic framework and definitions 107

3.3 Estimation 111

3.3.1 Data representation 111

3.3.2 Nelson-Aalen and Aalen-Johansen estimator 111

3.4 Example: HIV, SI, AIDS and death 115

3.4.1 The data 116

3.4.2 Analyses 117

3.5 Summary; some alternative approaches 125

3.6 Exercises 126

3.7 Software 127

3.7.1 The etm package 130

3.7.2 The msSurv package 134

3.7.3 The instate package 138

3.8 Computer practicals 141

4 Regression; Cause-Specific/Transition Hazard 143

4.1 Introduction 143

4.2 Regression on cause-specific hazard: basic structure 144

4.3 Combined analysis and type-specific covariablcs 146

4.3.1 Same results in one analysis 147

4.3.2 Type-specific covariables 149

4.3.3 Effects equal over causes 151

4.3.4 Proportional baseline hazards 152

4.4 Why does the stacked approach work? 153

4.4.1 Cause as stratum variable 153

4.4.2 Effects equal over causes 155

4.4.3 Proportional baseline hazards 155

4.5 Multi-state regression models for transition hazards 156

4.5.1 Combined analyses: assume effects to be equal 159

4.5.2 Proportional baseline hazards 160

4.5.3 Dual role of intermediate states 164

4.5.4 Beyond the Markov model: effect of transition tune 165

4.5.5 Standard error 167

4.6 Example: causes of death in HIV infected individuals 167

4.6.1 Analysis using well-defined contrasts 172

4.7 Summary 176

4.8 Exercises 177

4.9 Software 179

4.10 Computer practicals 180

5 Regression; Translation to Cumulative Scale 183

5.1 Introduction 183

5.2 Prom cause-specific/transition hazard to probability 184

5.2.1 Competing risks 184

5.2.2 Multi-state models 187

5.3 Regression on subdistribution hazard 188

5.3.1 Choice of weight function 191

5.3.2 Estimation of standard error 192

5.3.3 Time-varying covariables 192

5.3.4 Examples 195

5.4 Multinomial regression 199

5.5 Summary 200

5.6 Exercises 202

5.7 Software 202

5.7.1 From cause-specific/transition hazard to probability 202

5.7.2 Regression on subdistribution hazard 206

5.7.3 Proportional odds model 208

5.8 Computer practicals 209

5.8.1 Multi-state analysis 209

6 Epilogue 213

6.1 Which type of quantity to choose? 213

6.2 Exercises 217.

Notes:

"A CRC title."

Includes bibliographical references and index.

Local Notes:

Acquired for the Penn Libraries with assistance from the Tom Farinholt and Blair Edlow Memorial Book Fund.

ISBN:

9781466570351

1466570350

OCLC:

918940718

Publisher Number:

99968460826