经典的R语言 Meta分析教程.pdf

发布时间：2022-06-12 发布人：admin 分类：说明书资料大小：3.89M 资料格式：pdf 举报版权申诉

c92bfd6e-87a8-4397-a531-b81856802919.pdf-第1页.png

第1页 / 共256页

c92bfd6e-87a8-4397-a531-b81856802919.pdf-第2页.png

第2页 / 共256页

c92bfd6e-87a8-4397-a531-b81856802919.pdf-第3页.png

第3页 / 共256页

c92bfd6e-87a8-4397-a531-b81856802919.pdf-第4页.png

第4页 / 共256页

c92bfd6e-87a8-4397-a531-b81856802919.pdf-第5页.png

第5页 / 共256页

c92bfd6e-87a8-4397-a531-b81856802919.pdf-第6页.png

第6页 / 共256页

c92bfd6e-87a8-4397-a531-b81856802919.pdf-第7页.png

第7页 / 共256页

c92bfd6e-87a8-4397-a531-b81856802919.pdf-第8页.png

第8页 / 共256页

Preface

Acknowledgements

Contents

Part I Getting Started

1 An Introduction to Meta-Analysis in R

1.1 Getting Started with R

1.1.1 Quitting R

1.1.2 R as a Calculator

1.1.3 Getting Help

1.2 Loading, Saving and Restoring Data

1.2.1 Importing Data from Other Packages

1.3 Select Variables from an R Dataset

1.4 Running Scripts

1.5 Installing and Using Libraries of Additional Functions

1.6 A First Meta-Analysis with R

1.7 Summary

References

Part II Standard Methods

2 Fixed Effect and Random Effects Meta-Analysis

2.1 Effect Measures for Continuous Outcomes

2.1.1 Mean Difference

2.1.2 Standardised Mean Difference

2.2 Fixed Effect Model

2.3 Random Effects Model

2.3.1 Estimation of Between-Study Variance

2.3.2 Hartung–Knapp Adjustment

2.3.3 Prediction Intervals

2.4 Tests and Measures of Heterogeneity

2.5 Subgroup Analysis

2.6 Meta-Analysis of Other Outcomes

2.6.1 Meta-Analysis with Survival Outcomes

2.6.2 Meta-Analysis of Cross-Over Trials

2.6.3 Meta-Analysis of Adjusted Treatment Effects

2.7 Summary

References

3 Meta-Analysis with Binary Outcomes

3.1 Effect Measures for Binary Outcomes

3.1.1 Odds Ratio

3.1.2 Risk Ratio

3.1.3 Risk Difference

3.1.4 Arcsine Difference

3.2 Estimation in Sparse Data

3.2.1 Peto Odds Ratio

3.3 Fixed Effect Model

3.3.1 Inverse Variance Method

3.3.2 Mantel–Haenszel Method

Odds Ratio

Risk Ratio

Risk Difference

3.3.3 Peto Method

3.4 Random Effects Model

3.4.1 DerSimonian–Laird Method

3.5 Heterogeneity and Subgroup Analyses

3.6 Summary

References

4 Heterogeneity and Meta-Regression

4.1 Sources of Heterogeneity

4.2 Measures of Heterogeneity

4.3 Test for Subgroup Differences

4.3.1 Fixed Effect Model

4.3.2 Random Effects Model with Separate Estimates of τ2

Estimate Separate Between-Study Variances (DerSimonian–Laird Method)

4.3.3 Random Effects Model with Common Estimate of τ2

Estimate Common Between-Study Variance (DerSimonian–Laird Method)

4.4 Meta-Regression

4.4.1 Meta-Regression with a Categorical Covariate

4.4.2 Meta-Regression with a Continuous Covariate

4.5 Summary

References

Part III Advanced Topics

5 Small-Study Effects in Meta-Analysis

5.1 Graphical Illustration of Small-Study Effects

5.1.1 Funnel Plot

Details on funnel.meta Function

Contour-Enhanced Funnel Plot

5.1.2 Radial Plot

5.2 Statistical Tests for Small-Study Effects

5.2.1 Classical Tests by Begg and Egger

Begg and Mazumdar Test: Rank Correlation Test

Egger's Test: Linear Regression Test

Test by Thompson and Sharp

5.2.2 Modified Versions of Classical Tests for BinaryOutcomes

Harbord's Test: Score-Based Test

Macaskill's Test and Peters' Test

Schwarzer's Test

Rücker's Tests: Tests Based on Arcsine Difference

5.3 Adjusting for Small-Study Effects

5.3.1 Trim-and-Fill Method

5.3.2 Copas Selection Model

Funnel plot (Top Left)

Contour plot (Top Right)

Treatment effect plot (Bottom Left)

P-value plot (Bottom Right)

5.3.3 Adjustment by Regression

5.4 Summary

References

6 Missing Data in Meta-Analysis

6.1 Missing Outcome Data: Some Considerations

6.1.1 Study-Level Adjustment for Missing Data

6.1.2 Sensitivity Analysis Strategies

6.1.3 Strategy 1: Fixed Equal

6.1.4 Strategy 2: Fixed Opposite

6.1.5 Strategy 3: Random Equal

6.1.6 Strategy 4: Random Uncorrelated

6.1.7 Discussion of the Four Strategies

6.2 Missing Precision

6.2.1 Multiple Imputation Approach

Basic Idea of Multiple Imputation Algorithm

Further Details

6.2.2 Missing Participant Numbers

6.3 Summary

References

7 Multivariate Meta-Analysis

7.1 Fixed Effect Model

7.2 Dealing with Unbalanced Data

7.3 Random Effects Model

7.3.1 Fitting the Random Effects Model

7.4 Discussion

References

8 Network Meta-Analysis

8.1 Concepts and Challenges of Network Meta-Analysis

8.2 Model and Estimation in Network Meta-Analysis

8.2.1 Fixed Effect Model

Estimation of Treatment Effects

Variance Estimation

Multi-Arm Studies

I-Squared for Network Meta-Analysis

8.2.2 Random Effects Model

8.3 Using the R Package netmeta for Network Meta-Analysis

8.3.1 Basic Analysis and Network Plots

8.3.2 A First Network Plot

8.3.3 A More Detailed Look at the Output

8.3.4 Additional Network Plots

8.3.5 Forest Plots

8.4 Decomposition of the Heterogeneity Statistic

8.5 The Net Heat Plot

8.5.1 Bland–Altman Plot to Assess the Effect of Heterogeneity on Estimated Treatment Comparisons

8.6 Summary

References

9 Meta-Analysis of Diagnostic Test Accuracy Studies

9.1 Special Challenges in Meta-Analysis of Diagnostic Test Accuracy Studies

9.2 Analysis of Diagnostic Test Accuracy Studies

9.2.1 Definition of Sensitivity and Specificity

9.2.2 Additional Measures: Diagnostic Odds Ratio and Likelihood Ratios

9.2.3 Tests Based on a Continuous Marker

9.3 Scatterplot of Sensitivity and Specificity

9.4 Models for Meta-Analysis of Diagnostic Test Accuracy Studies

9.4.1 Hierarchical Model

9.4.2 Bivariate Model

9.5 Methods for Estimating a Summary ROC Curve

9.6 Summary

References

A Further Information on R

A.1 Installation of R

A.2 Importing Data into R

A.2.1 Import Text Files

A.2.2 Import Data from RevMan 5

A.3 R Packages for Meta-Analysis

A.3.1 General Purpose R Packages for Meta-Analysis

A.3.2 R Packages to Conduct Network Meta-Analysis

References

Index

UseR! GuidoSchwarzer JamesR.Carpenter GertaRücker Meta- Analysis with R

Use R! Series editors Robert Gentleman Kurt Hornik Giovanni Parmigiani

More information about this series at http://www.springer.com/series/6991

Guido Schwarzer • James R. Carpenter Gerta RRucker Meta-Analysis with R 123

Guido Schwarzer Institute for Medical Biometry and Statistics Medical Center – University of Freiburg Freiburg, Germany Gerta RRucker Institute for Medical Biometry and Statistics Medical Center – University of Freiburg Freiburg, Germany James R. Carpenter MRC Clinical Trials Unit, London and London School of Hygiene and Tropical Medicine London, United Kingdom ISSN 2197-5736 Use R! ISBN 978-3-319-21415-3 DOI 10.1007/978-3-319-21416-0 ISSN 2197-5744 (electronic) ISBN 978-3-319-21416-0 (eBook) Library of Congress Control Number: 2015949262 Springer Cham Heidelberg New York Dordrecht London © Springer International Publishing Switzerland 2015 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, speciﬁcally the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microﬁlms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a speciﬁc statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. Printed on acid-free paper Springer International Publishing AG Switzerland is part of Springer Science+Business Media (www.springer.com)

Preface Meta-analysis plays a key role in evidence synthesis in many research disciplines, not least the social sciences, medicine and economics. The aim of this book is to equip those involved in such work (who are often not trained statisticians) to use R for meta-analysis, and thus promote both the use of R and the latest statistical methods in this area. The attractions of R in this context (besides its free availability from http://www. r-project.org/) are its fast yet powerful and ﬂexible graphics and its well-established algorithmic base. The book assumes no prior knowledge of R, and takes readers through every step of the way from installing R, loading data from other packages, performing and interpreting the analyses. Parts I and II cover the essentials, while Part III considers more advanced topics, which remain the subject of active research. Throughout, the ideas are illustrated with examples, and all the codes necessary to repeat these examples (including creating all the plots in the book) are either in the text itself or the web-appendix http://meta-analysis-with-r.org/. In selecting the code to include in the main text, we have assumed readers are relatively new to R. More experienced users can easily skip over familiar material. Freiburg, Germany London, UK Freiburg, Germany December, 2014 Guido Schwarzer James R. Carpenter Gerta Rücker v

Acknowledgements This book was funded partly by the German Research Foundation (DFG), Research Unit 534, FOR Schw 821/2-2. Subsequently, Gerta Rücker was funded by the German Research Foundation (RU 1747/1-1). James Carpenter is grateful for fund- ing from the UK Medical Research Council’s London Hub for Trials Methodology. We are grateful to the members of the R Core Team who provide and maintain the statistical package R. We would also like to acknowledge our debt to the authors of R packages for meta-analysis, which we use in this book. Particular thanks go to Wolfgang Viechtbauer for the comprehensive R package metafor, Antonio Gasparrini for the versatile R package mvmeta for multivariate meta-analysis, and for advice on its use, and Philipp Doebler for the R package mada and his contribution to the chapter on meta-analysis of diagnostic test accuracy studies. Special thanks are also due to Ulrike Krahn for her advice and encouragement with the chapter on network meta-analysis. We are also grateful to Anna Wiksten, Jan Beyersmann, Karin Schiefele, Harriet Sommer and two anonymous referees for reading earlier versions of the manuscript and many helpful and detailed comments. These have substantially improved the book. We are also grateful to our families for their forbearance over the course of the project. Despite this encouragement and support, the text inevitably contains errors and shortcomings, for which we take full responsibility. Last but not least, we have found collaborating on this project both informative and fun; if readers feel the same after reading the book, we will be very satisﬁed! vii

分享到：

赞收藏

资料库

经典的R语言 Meta分析教程.pdf

相关推荐

课程资源

热门标签

最新资料