Individual Participant Data Meta-Analysis E-Book
67,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Serie: Statistics in Practice
- Sprache: Englisch
Individual Participant Data Meta-Analysis: A Handbook for Healthcare Research provides a comprehensive introduction to the fundamental principles and methods that healthcare researchers need when considering, conducting or using individual participant data (IPD) meta-analysis projects. Written and edited by researchers with substantial experience in the field, the book details key concepts and practical guidance for each stage of an IPD meta-analysis project, alongside illustrated examples and summary learning points.
Split into five parts, the book chapters take the reader through the journey from initiating and planning IPD projects to obtaining, checking, and meta-analysing IPD, and appraising and reporting findings. The book initially focuses on the synthesis of IPD from randomised trials to evaluate treatment effects, including the evaluation of participant-level effect modifiers (treatment-covariate interactions). Detailed extension is then made to specialist topics such as diagnostic test accuracy, prognostic factors, risk prediction models, and advanced statistical topics such as multivariate and network meta-analysis, power calculations, and missing data.
Intended for a broad audience, the book will enable the reader to:
- Understand the advantages of the IPD approach and decide when it is needed over a conventional systematic review
- Recognise the scope, resources and challenges of IPD meta-analysis projects
- Appreciate the importance of a multi-disciplinary project team and close collaboration with the original study investigators
- Understand how to obtain, check, manage and harmonise IPD from multiple studies
- Examine risk of bias (quality) of IPD and minimise potential biases throughout the project
- Understand fundamental statistical methods for IPD meta-analysis, including two-stage and one-stage approaches (and their differences), and statistical software to implement them
- Clearly report and disseminate IPD meta-analyses to inform policy, practice and future research
- Critically appraise existing IPD meta-analysis projects
- Address specialist topics such as effect modification, multiple correlated outcomes, multiple treatment comparisons, non-linear relationships, test accuracy at multiple thresholds, multiple imputation, and developing and validating clinical prediction models
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 1299
Veröffentlichungsjahr: 2021
Ähnliche
Table of Contents
Cover
Series Page
Title Page
Copyright Page
Dedication Page
Acknowledgements
1 Individual Participant Data Meta‐Analysis for Healthcare Research
1.1 Introduction
1.2 What Is IPD and How Does It Differ from Aggregate Data?
1.3 IPD Meta‐Analysis: A New Era for Evidence Synthesis
1.4 Scope of This Book and Intended Audience
Part I: Rationale, Planning, and Conduct
2 Rationale for Embarking on an IPD Meta‐Analysis Project
2.1 Introduction
2.2 How Does the Research Process Differ for IPD and Aggregate Data Meta‐Analysis Projects?
2.3 What Are the Potential Advantages of an IPD Meta‐Analysis Project?
2.4 What Are the Potential Challenges of an IPD Meta‐Analysis Project?
2.5 Empirical Evidence of Differences Between Results of IPD and Aggregate Data Meta‐Analysis Projects
2.6 Guidance for Deciding When IPD Meta‐Analysis Projects Are Needed to Evaluate Treatment Effects from Randomised Trials
2.7 Concluding Remarks
3 Planning and Initiating an IPD Meta‐Analysis Project
3.1 Introduction
3.2 Organisational Approach
3.3 Developing a Project Scope
3.4 Assessing Feasibility and ‘In Principle’ Support and Collaboration
3.5 Establishing a Team with the Right Skills
3.6 Advisory and Governance Functions
3.7 Estimating How Long the Project Will Take
3.8 Estimating the Resources Required
3.9 Obtaining Funding
3.10 Obtaining Ethical Approval
3.11 Data‐sharing Agreement
3.12 Additional Planning for Prospective Meta‐Analysis Projects
3.13 Concluding Remarks
4 Running an IPD Meta‐Analysis Project
4.1 Introduction
4.2 Preparing to Collect IPD
4.3 Initiating and Maintaining Collaboration
4.4 Obtaining IPD
4.5 Checking and Harmonising Incoming IPD
4.6 Checking the IPD to Inform Risk of Bias Assessments
4.7 Assessing and Presenting the Overall Quality of a Trial
4.8 Verification of Finalised Trial IPD
4.9 Merging IPD Ready for Meta‐Analysis
4.10 Concluding Remarks
Part I: References
Part II: Fundamental Statistical Methods and Principles
5 The Two‐stage Approach to IPD Meta‐Analysis
5.1 Introduction
5.2 First Stage of a Two‐stage IPD Meta‐Analysis
5.3 Second Stage of a Two‐stage IPD Meta‐Analysis
5.4 Meta‐regression and Subgroup Analyses
5.5 The
ipdmetan
Software Package
5.6 Combining IPD with Aggregate Data from non‐IPD Trials
5.7 Concluding Remarks
6 The One‐stage Approach to IPD Meta‐Analysis
6.1 Introduction
6.2 One‐stage IPD Meta‐Analysis Models Using Generalised Linear Mixed Models
6.3 One‐stage Models for Time‐to‐event Outcomes
6.4 One‐stage Models Combining Different Sources of Evidence
6.5 Reporting of One‐stage Models in Protocols and Publications
6.6 Concluding Remarks
7 Using IPD Meta‐Analysis to Examine Interactions between Treatment Effect and Participant‐level Covariates
7.1 Introduction
7.2 Meta‐regression and Its Limitations
7.3 Two‐stage IPD Meta‐Analysis to Estimate Treatment‐covariate Interactions
7.4 The One‐stage Approach
7.5 Combining IPD and non‐IPD Trials
7.6 Handling of Continuous Covariates
7.7 Handling of Categorical or Ordinal Covariates
7.8 Misconceptions and Cautions
7.9 Is My Identified Treatment‐covariate Interaction Genuine?
7.10 Reporting of Analyses of Treatment‐covariate Interactions
7.11 Can We Predict a New Patient’s Treatment Effect?
7.12 Concluding Remarks
8 One‐stage versus Two‐stage Approach to IPD Meta‐Analysis: Differences and Recommendations
8.1 Introduction
8.2 One‐stage and Two‐stage Approaches Usually Give Similar Results
8.3 Ten Key Reasons Why One‐stage and Two‐stage Approaches May Give Different Results
8.4 Recommendations and Guidance
8.5 Concluding Remarks
Part II: References
Part III: Critical Appraisal and Dissemination
9 Examining the Potential for Bias in IPD Meta‐Analysis Results
9.1 Introduction
9.2 Publication and Reporting Biases of Trials
9.3 Biased Availability of the IPD from Trials
9.4 Trial Quality (risk of bias)
9.5 Other Potential Biases Affecting IPD Meta‐Analysis Results
9.6 Concluding Remarks
10 Reporting and Dissemination of IPD Meta‐Analyses
10.1 Introduction
10.2 Reporting IPD Meta‐Analysis Projects in Academic Reports
10.3 Additional Means of Disseminating Findings
10.4 Concluding Remarks
11 A Tool for the Critical Appraisal of IPD Meta‐Analysis Projects (CheckMAP)
11.1 Introduction
11.2 The CheckMAP Tool
11.3 Was the IPD Meta‐Analysis Project Done within a Systematic Review Framework?
11.4 Were the IPD Meta‐Analysis Project Methods Pre‐specified in a Publicly Available Protocol?
11.5 Did the IPD Meta‐Analysis Project Have a Clear Research Question Qualified by Explicit Eligibility Criteria?
11.6 Did the IPD Meta‐Analysis Project Have a Systematic and Comprehensive Search Strategy?
11.7 Was the Approach to Data Collection Consistent and Thorough?
11.8 Were IPD Obtained from Most Eligible Trials and Their Participants?
11.9 Was the Validity of the IPD Checked for Each Trial?
11.10 Was the Risk of Bias Assessed for Each Trial and Its Associated IPD?
11.11 Were the Methods of Meta‐Analysis Appropriate?
11.12 Concluding Remarks
Part III: References
Part IV: Special Topics in Statistics
12 Power Calculations for Planning an IPD Meta‐Analysis
12.1 Introduction
12.2 Motivating Example: Power of a Planned IPD Meta‐Analysis of Trials of Interventions to Reduce Weight Gain in Pregnant Women
12.3 Power of an IPD Meta‐Analysis to Detect a Treatment‐covariate Interaction for a Continuous Outcome
12.4 The Contribution of Individual Trials Toward Power
12.5 The Impact of Model Assumptions on Power
12.6 Extensions
12.7 Concluding Remarks
13 Multivariate Meta‐Analysis Using IPD
13.1 Introduction
13.2 General Two‐stage Approach for Multivariate IPD Meta‐Analysis
13.3 Application to an IPD Meta‐Analysis of Anti‐hypertensive Trials
13.4 Extension to Multivariate Meta‐regression
13.5 Potential Limitations of Multivariate Meta‐Analysis
13.6 One‐stage Multivariate IPD Meta‐Analysis Applications
13.7 Special Applications of Multivariate Meta‐Analysis
13.8 Concluding Remarks
14 Network Meta‐Analysis Using IPD
14.1 Introduction
14.2 Rationale and Assumptions for Network Meta‐Analysis
14.3 Network Meta‐Analysis Models Assuming Consistency
14.4 Ranking Treatments
14.5 How Do We Examine Inconsistency between Direct and Indirect Evidence?
14.6 Benefits of IPD for Network Meta‐Analysis
14.7 Combining IPD and Aggregate Data in Network Meta‐Analysis
14.8 Further Topics
14.9 Concluding Remarks
Part IV: References
Part V: Diagnosis, Prognosis and Prediction
15 IPD Meta‐Analysis for Test Accuracy Research
15.1 Introduction
15.2 Motivating Example: Diagnosis of Fever in Children Using Ear Temperature
15.3 Key Steps Involved in an IPD Meta‐Analysis of Test Accuracy Studies
15.4 IPD Meta‐Analysis of Test Accuracy at Multiple Thresholds
15.5 IPD Meta‐Analysis for Examining a Test’s Clinical Utility
15.6 Comparing Tests
15.7 Concluding Remarks
16 IPD Meta‐Analysis for Prognostic Factor Research
16.1 Introduction
16.2 Potential Advantages of an IPD Meta‐Analysis
16.3 Key Steps Involved in an IPD Meta‐Analysis of Prognostic Factor Studies
16.4 Software
16.5 Concluding Remarks
17 IPD Meta‐Analysis for Clinical Prediction Model Research
17.1 Introduction
17.2 IPD Meta‐Analysis for Prediction Model Research
17.3 External Validation of an Existing Prediction Model Using IPD Meta‐Analysis
17.4 Updating and Tailoring of a Prediction Model Using IPD Meta‐Analysis
17.5 Comparison of Multiple Existing Prediction Models Using IPD Meta‐Analysis
17.6 Using IPD Meta‐Analysis to Examine the Added Value of a New Predictor to an Existing Prediction Model
17.7 Developing a New Prediction Model Using IPD Meta‐Analysis
17.8 Examining the Utility of a Prediction Model Using IPD Meta‐Analysis
17.9 Software
17.10 Reporting
17.11 Concluding Remarks
18 Dealing with Missing Data in an IPD Meta‐Analysis
18.1 Introduction
18.2 Motivating Example: IPD Meta‐Analysis Validating Prediction Models for Risk of Pre‐eclampsia in Pregnancy
18.3 Types of Missing Data in an IPD Meta‐Analysis
18.4 Recovering Actual Values of Missing Data within IPD
18.5 Mechanisms and Patterns of Missing Data in an IPD Meta‐Analysis
18.6 Multiple Imputation to Deal with Missing Data in a Single Study
18.7 Ensuring Congeniality of Imputation and Analysis Models
18.8 Dealing with Sporadically Missing Data in an IPD Meta‐Analysis by Applying Multiple Imputation for Each Study Separately
18.9 Dealing with Systematically Missing Data in an IPD Meta‐Analysis Using a Bivariate Meta‐Analysis of Partially and Fully Adjusted Results
18.10 Dealing with Both Sporadically and Systematically Missing Data in an IPD Meta‐Analysis Using Multilevel Modelling
18.11 Comparison of Methods and Recommendations
18.12 Software
18.13 Concluding Remarks
Part V: References
Index
End User License Agreement
List of Tables
Chapter 2
Table 2.1 Key potential advantages of an IPD meta‐analysis project compared w...
Table 2.2 Signalling questions to help decide when aggregate data are insuffi...
Chapter 3
Table 3.1 Consent sought to collaborate in an IPD analysis of predictive fact...
Chapter 4
Table 4.1 Excerpt from a data dictionary developed for an IPD meta‐analysis o...
Table 4.2 Excerpt from a data dictionary developed for an IPD meta‐analysis p...
Table 4.3 Example of items to include in a data transfer guide when requestin...
Table 4.4 Domains in the Risk of Bias 2 tool
91
(RoB 2) of particular relevanc...
Table 4.5 Alleviating potential bias in trials that stopped early for perceiv...
Table 4.6 Excerpt of a RoB2 table for an IPD meta‐analysis of adjuvant chemot...
Chapter 5
Table 5.1 Example of hypothetical IPD for one trial similar to those received...
Table 5.2 Basic format of the regression model to be fitted separately within...
Table 5.3 Example of aggregate data calculated for each trial in the first st...
Chapter 6
Table 6.1 Basic format of the GLMM for one‐stage IPD meta‐analysis models of ...
Table 6.2 Software for fitting one‐stage IPD meta‐analysis via the GLMM or su...
Chapter 8
Table 8.1 Comparison of summary results from ML estimation of one‐stage and t...
Table 8.2 Summary results from an IPD meta‐analysis of 10 randomised trials e...
Table 8.3 Summary treatment effect results for the hypertension data to illus...
Table 8.4 One‐stage and two‐stage REML results for the prognostic effect of s...
Table 8.5 Summary results from one‐stage and two‐stage IPD meta‐analyses of 1...
Chapter 10
Table 10.1 Title and Introduction sections of the PRISMA‐IPD Checklist as rep...
Table 10.2 Methods section of the PRISMA‐IPD Checklist as reported by Stewart...
Table 10.3 Results and discussion sections of the PRISMA‐IPD Checklist as rep...
Chapter 11
Table 11.1 Summary of completion of the CheckMAP tool for an IPD meta‐analysi...
Chapter 12
Table 12.1 Trial characteristics used in the power calculations of an IPD met...
Chapter 13
Table 13.1 Summary results for the 10 trials included in the meta‐analysis of...
Table 13.2 Bivariate and multivariate results for the IPD meta‐analysis of 10...
Table 13.3 Results from two‐stage and one‐stage bivariate meta‐analysis of SB...
Table 13.4 Results from one‐stage bivariate IPD meta‐analysis of 10 randomise...
Table 13.5 Summary results following REML estimation of a two‐stage multivari...
Chapter 14
Table 14.1 Trials in the thrombolytic network meta‐analysis summarised in ter...
Table 14.2 Summary results from the thrombolytics network meta‐analysis compa...
Table 14.3 Results reported by Donegan et al.
191
after estimation of network ...
Table 14.4 Baseline covariate summaries from the UNCOVER and FIXTURE trials. ...
Chapter 15
Table 15.1 Cross classification of index test results and reference standard ...
Table 15.2 Typical statistical measures of test accuracy in a single study
i
...
Table 15.3 Summary of 23 studies used by Riley et al. in an IPD meta‐analysis...
Chapter 16
Table 16.1 Selected items of the CHARMS‐PF
*
checklist to be extracted when id...
Chapter 17
Table 17.1 Domains and signalling questions within the first three domains of...
Table 17.2 Selected items from the CHARMS checklist to be extracted when iden...
Table 17.3 Relevant statistics to be estimated in the first stage of a two‐st...
Table 17.4 Coefficients of eight prediction models for diagnosing DVT in pati...
Table 17.5 Trivariate random‐effects meta‐analysis results for calibration (a...
Table 17.6 Model parameter estimates for the fitted DVT model obtained in eac...
Table 17.7 Trivariate meta‐analysis results
*
for the calibration and discrimi...
Table 17.8 Preliminary version of the TRIPOD‐CLUSTER statement extended to si...
Chapter 18
Table 18.1 Summary of five IPD studies from the IPPIC network that were used ...
Table 18.2 Example of sporadically missing predictors (previous pre‐eclampsia...
Table 18.3 Summary statistics for participants with complete data, and for pa...
Table 18.4 A summary of the results at each stage of the multiple imputation ...
Table 18.5 Description of variables and missing data (NA) in the simulated IP...
Table 18.6 Key advantages and limitations of the multilevel joint modelling a...
List of Illustrations
Chapter 1
Figure 1.1 Number of published IPD meta‐analysis articles over time, based o...
Chapter 2
Figure 2.1 Key differences between the process for a IPD meta‐analysis proje...
Chapter 3
Figure 3.1 Typical phases of an IPD meta‐analysis project.
Chapter 4
Figure 4.1 Overview of key steps involved in obtaining, managing and checkin...
Figure 4.2 PICOS example: objective and eligibility criteria for an IPD meta...
Figure 4.3 Excerpt from a trial‐level data collection form for the STOPCAP M...
Figure 4.4 Excerpts from the website https://www.york.ac.uk/crd/research/epp...
Figure 4.5 Median follow‐up based on published aggregate data compared to up...
Figure 4.6 Summary of the data validity, range and consistency checks on IPD...
Figure 4.7 The cumulative number of participants randomised to the intervent...
Figure 4.8 The cumulative number of participants allocated to chemotherapy o...
Figure 4.9 Date (shown by year‐month) participants were allocated to treatme...
Figure 4.10 Days of the week participants were allocated to treatment and co...
Figure 4.11 Percentage of participants excluded from the original analyses o...
Figure 4.12 ‘Reverse’ Kaplan‐Meier analysis of participants who are event‐fr...
Figure 4.13 Example of items to include in summary of finalised trial IPD fo...
Chapter 5
Figure 5.1 Posterior distributions after applying a Bayesian random treatmen...
Figure 5.2 Forest plots of two hypothetical meta‐analyses that give the same...
Chapter 6
Figure 6.1 Results of the simulation study of Abo‐Zaid et al.,
3
comparing (a...
Figure 6.2 Simulation results for binary outcomes from Riley et al.
181
for s...
Figure 6.3 Baseline hazard functions from ML estimation of a one‐stage IPD m...
Figure 6.4 Estimated baseline hazard rate for one of the 10 trials included ...
Chapter 7
Figure 7.1 Scatter plot (on logit scale) of
p
‐values from 31 re‐analyses of ...
Figure 7.2 A two‐stage IPD meta‐analysis of treatment‐sex interactions, summ...
Figure 7.3 Is blood pressure–lowering treatment more effective amongst women...
Figure 7.4 Bland‐Altman plot showing level of agreement between treatment‐co...
Figure 7.5 PORT meta‐analysis results for the interaction between treatment ...
Figure 7.6 Representations of how the effect of an early supported hospital ...
Figure 7.7 Evidence for a potential non‐linear interaction between age and t...
Figure 7.8 Evidence of a potential non‐linear interaction between baseline S...
Figure 7.9 The predicted effect of anti‐hypertensive treatment on SBP condit...
Chapter 8
Figure 8.1 One‐stage and two‐stage IPD meta‐analysis summary results from 10...
Figure 8.2 Forest plot showing the one‐stage and two‐stage IPD meta‐analysis...
Figure 8.3 Forest plot showing one‐stage and two‐stage IPD meta‐analysis res...
Figure 8.4 Summary treatment effect estimates and 95% confidence intervals f...
Chapter 9
Figure 9.1 Examination of small‐study effects in an IPD meta‐analysis to eva...
Figure 9.2 Forest plot of the IPD meta‐analysis
*
of Martineau et al.
24
, with...
Figure 9.3 Example of combining IPD and non‐IPD trials in an IPD meta‐analys...
Figure 9.4 Contour‐enhanced funnel plots for the IPD meta‐analysis of Rogozi...
Chapter 10
Figure 10.1 Example of the PRISMA‐IPD flowchart applied to an IPD meta‐analy...
Figure 10.2 Example of a figure displaying the extent and pattern of missing...
Chapter 12
Figure 12.1 Simulation‐based power estimates (based on 10,000 replications) ...
Figure 12.2 Simulation‐based power estimates (based on 10,000 replications) ...
Figure 12.3 Simulation‐based power estimates for the planned two‐stage IPD m...
Figure 12.4 Predictive distribution for the true treatment effect in a new P...
Chapter 13
Figure 13.1 Scatterplot of 10,000 pairs of treatment effect estimates on sys...
Figure 13.2 Relationship between the observed treatment effect estimates on ...
Figure 13.3 Forest plot comparing the summary treatment effect results and p...
Figure 13.4 Relationship between the observed treatment effect estimates on ...
Figure 13.5 Comparison of summary meta‐analysis results for outcome 1 in the...
Chapter 14
Figure 14.1 Visual representation of direct and indirect evidence toward the...
Figure 14.2 Network map of the direct comparisons available in the 28 trials...
Figure 14.3 Extended forest plot showing the thrombolytics network meta‐anal...
Figure 14.4 Plots of the ranking probability for each treatment considered i...
Figure 14.5 Ranking of antimanic drugs for response and acceptability, based...
Figure 14.6 The UNCOVER
228,229
and FIXTURE
227
trials form a network of six t...
Figure 14.7 Average treatment effect estimates at the population level, for ...
Chapter 15
Figure 15.1 Summary receiver operating characteristic (ROC) curves for the d...
Figure 15.2 Confidence and prediction regions following application of model...
Figure 15.3 Predictive (posterior) distributions for the true sensitivity an...
Figure 15.4 Results of an IPD meta‐analysis of 29 studies (6725 participants...
Figure 15.5 Histograms showing the observed distributions of ear temperature...
Figure 15.6 Summary normal distribution for the fever and non‐fever groups b...
Figure 15.7 Summary ROC curves when applying the conventional bivariate mode...
Chapter 16
Figure 16.1 Forest plot showing the study estimates and IPD meta‐analysis re...
Figure 16.2 Non‐linear prognostic relationship between temperature and relat...
Figure 16.3 IPD meta‐analysis to examine whether glucose is a prognostic fac...
Figure 16.4 Summary of the relationship between MVD and mortality rate, acco...
Figure 16.5 Summary estimate and 95% confidence interval (CI) for the relati...
Figure 16.6 Prognostic effect of BMI for subsequent cardiovascular disease, ...
Figure 16.7 Contour‐enhanced funnel plot of the 31 studies included in the m...
Chapter 17
Figure 17.1 Examples of the calibration performance of a logistic regression...
Figure 17.2 Forest plot showing study‐specific and IPD meta‐analysis results...
Figure 17.3 Meta‐regression showing a strong association between the observe...
Figure 17.4 Summary estimates (and 95% CIs) of standardised and unstandardis...
Figure 17.5 Overall calibration performance of a prognostic model for breast...
Figure 17.6 Overall
*
calibration performance of a prediction model for morta...
Figure 17.7 Calibration of QRISK2 and the Framingham risk score in women age...
Figure 17.8 Forest plot of summary calibration slopes for different predicti...
Figure 17.9 Calibration plots for models predicting late‐onset pre‐eclampsia...
Figure 17.10 Top panel: Summary and predicted decision curves for diagnosing...
Chapter 18
Figure 18.1 Options to analysing multiple imputation datasets in an IPD meta...
Figure 18.2 IPD meta‐analysis of 10 studies examining the prognostic effect ...
Figure 18.3 Summary results from a one‐stage IPD meta‐analysis of the GREAT ...
Figure 18.4 MCMC sampling chain for the beta corresponding to history of dia...
Guide
Cover Page
Series Page
Title Page
Copyright Page
Dedication Page
Acknowledgements
Table of Contents
Begin Reading
Index
Wiley End User License Agreement
Pages
ii
iii
iv
v
ix
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
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
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
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
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
237
239
240
241
242
243
244
245
246
247
248
249
250
251
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
285
286
287
288
289
291
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
375
376
377
378
379
380
381
382
383
384
385
386
387
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
543
544
545
546
547
548
549
550
551
WILEY SERIES IN STATISTICS IN PRACTICE
Advisory Editor, Marian Scott, University of Glasgow, Scotland, UK
Founding Editor, Vic Barnett, Nottingham Trent University, UK
Statistics in Practice is an important international series of texts which provide detailed coverage of statistical concepts, methods, and worked case studies in specific fields of investigation and study.
With sound motivation and many worked practical examples, the books show in down-to-earth terms how to select and use an appropriate range of statistical techniques in a particular practical field within each title's special topic area.
The books provide statistical support for professionals and research workers across a range of employment fields and research environments. Subject areas covered include medicine and pharmaceutics; industry, finance, and commerce; public services; the earth and environmental sciences, and so on.
The books also provide support to students studying statistical courses applied to the above areas. The demand for graduates to be equipped for the work environment has led to such courses becoming increasingly prevalent at universities and colleges.
It is our aim to present judiciously chosen and well-written workbooks to meet everyday practical needs. Feedback of views from readers will be most valuable to monitor the success of this aim.
A complete list of titles in this series appears at the end of the volume.
Human and Biological Sciences
Brown and Prescott · Applied Mixed Models in Medicine
Ellenberg, Fleming, and DeMets · Data Monitoring Committees in Clinical Trials: A Practical Perspective
Lawson, Browne, and Vidal Rodeiro · Disease Mapping With WinBUGS and MLwiN
Lui · Statistical Estimation of Epidemiological Risk
Marubini and Valsecchi · Analysing Survival Data from Clinical Trials and Observation Studies
Parmigiani · Modeling in Medical Decision Making: A Bayesian Approach
Senn · Cross-over Trials in Clinical Research, Second Edition
Senn · Statistical Issues in Drug Development
Spiegelhalter, Abrams, and Myles · Bayesian Approaches to Clinical Trials and Health-Care Evaluation
Turner · New Drug Development: Design, Methodology, and Analysis
Whitehead · Design and Analysis of Sequential Clinical Trials, Revised Second Edition
Whitehead · Meta-Analysis of Controlled Clinical Trials
Zhou, Zhou, Liu, and Ding · Applied Missing Data Analysis in the Health Sciences
Earth and Environmental Sciences
Buck, Cavanagh and Litton · Bayesian Approach to Interpreting Archaeological Data
Cooke · Uncertainty Modeling in Dose Response: Bench Testing Environmental Toxicity
Gibbons, Bhaumik, and Aryal · Statistical Methods for Groundwater Monitoring, Second Edition
Glasbey and Horgan · Image Analysis in the Biological Sciences
Helsel · Nondetects and Data Analysis: Statistics for Censored Environmental Data
Helsel · Statistics for Censored Environmental Data Using Minitab®and R, Second Edition
McBride · Using Statistical Methods for Water Quality Management: Issues, Problems and Solutions
Ofungwu · Statistical Applications for Environmental Analysis and Risk Assessment
Webster and Oliver · Geostatistics for Environmental Scientists
Industry, Commerce and Finance
Aitken and Taroni · Statistics and the Evaluation of Evidence for Forensic Scientists, Second Edition
Brandimarte · Numerical Methods in Finance and Economics: A MATLAB-Based Introduction, Second Edition
Brandimarte and Zotteri · Introduction to Distribution Logistics
Chan and Wong · Simulation Techniques in Financial Risk Management, Second Edition
Jank · Statistical Methods in eCommerce Research
Jank and Shmueli · Modeling Online Auctions
Lehtonen and Pahkinen · Practical Methods for Design and Analysis of Complex Surveys, Second Edition
Lloyd · Data Driven Business Decisions
Ohser and Mücklich · Statistical Analysis of Microstructures in Materials Science
Rausand · Risk Assessment: Theory, Methods, and Applications
Individual Participant Data Meta‐Analysis
A Handbook for Healthcare Research
Edited by
Richard D. Riley
Keele University
Keele, UK
Jayne F. Tierney
MRC Clinical Trials Unit at UCL
London, UK
Lesley A. Stewart
University of York
York, UK
This edition first published 2021© 2021 John Wiley & Sons Ltd
All rights reserved. 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 or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.
The right of Richard D. Riley, Jayne F. Tierney, and Lesley A. Stewart to be identified as the authors of the editorial material in this work has been asserted in accordance with law.
Registered OfficesJohn Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USAJohn Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, UK
Editorial Office9600 Garsington Road, Oxford, OX4 2DQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of WarrantyThe contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging‐in‐Publication Data
Names: Riley, Richard D., editor. | Tierney, Jayne F., editor. | Stewart, Lesley A., editor.Title: Individual participant data meta-analysis : a handbook for healthcare research / edited by Richard D. Riley, Jayne F. Tierney, Lesley A. Stewart.Other titles: Statistics in practice. Description: Hoboken, NJ : Wiley, 2021. | Series: Wiley series in statistics in practice | Includes bibliographical references and index. | Contents: Individual Participant Data Meta‐Analysis for Healthcare Research / Richard D. Riley, Lesley A. Stewart, Jayne F. Tierney – Rationale for Embarking on an IPD Meta-Analysis Project / Jayne F. Tierney, Richard D. Riley, Catrin Tudur Smith, Mike Clarke, and Lesley A. Stewart – Planning and Initiating an IPD Meta-Analysis Project / Lesley A. Stewart, Richard D. Riley, and Jayne F. Tierney – Running an IPD Meta-Analysis Project : From Developing the Protocol to Preparing Data for Metaanalysis / Jayne F. Tierney, Richard D. Riley, Larysa H.M. Rydzewska, and Lesley A. Stewart – The Two-stage Approach to IPD Meta-Analysis / Richard D. Riley, Thomas P.A. Debray, Tim P. Morris, and Dan Jackson – The One-stage Approach to IPD Meta-Analysis / Richard D. Riley and Thomas P.A. Debray – Using IPD Meta-Analysis to Examine Interactions between Treatment Effect and Participant-level Covariates / Richard D. Riley and David J. Fisher – One-stage versus Two-stage Approach to IPD Meta-Analysis : Differences and Recommendations / Richard D. Riley, Danielle L. Burke, and Tim Morris – Examining the Potential for Bias in IPD Meta-Analysis Results / Richard D. Riley, Jayne F. Tierney, and Lesley A. Stewart – Reporting and Dissemination of IPD Meta-Analyses / Lesley A. Stewart, Richard D. Riley, and Jayne F. Tierney – A Tool for the Critical Appraisal of IPD Meta-Analysis Projects (CheckMAP) / Jayne F. Tierney, Lesley A. Stewart, Claire L. Vale, and Richard D. Riley – Power Calculations for Planning an IPD Meta-Analysis / Richard D. Riley and Joie Ensor – Multivariate Meta-Analysis Using IPD / Richard D. Riley, Dan Jackson, and Ian R. White – Network Meta-Analysis Using IPD / Richard D. Riley, David M Phillippo, and Sofia Dias – IPD Meta-Analysis for Test Accuracy Research / Richard D. Riley, Brooke Levis, and Yemisi Takwoingi – IPD Meta-Analysis for Prognostic Factor Research / Richard D. Riley, Karel G.M. Moons, and Thomas P.A. Debray – IPD Meta-Analysis for Clinical Prediction Model Research / Richard D. Riley, Kym I.E. Snell, Laure Wynants, Valentijn M.T. de Jong, Karel G.M. Moons, and Thomas P.A. Debray – Dealing with Missing Data in an IPD Meta-Analysis / Thomas P.A. Debray, Kym I.E. Snell, Matteo Quartagno, Shahab Jolani, Karel G.M. Moons, and Richard D. Riley.Identifiers: LCCN 2021000638 (print) | LCCN 2021000639 (ebook) | ISBN 9781119333722 (cloth) | ISBN 9781119333760 (adobe pdf) | ISBN 9781119333753 (epub) | ISBN 9781119333784 (oBook)Subjects: MESH: Meta-Analysis as Topic | Data Interpretation, Statistical | Models, Statistical | Randomized Controlled Trials as TopicClassification: LCC R853.S7 (print) | LCC R853.S7 (ebook) | NLM WA 950 | DDC 610.72/7–dc23LC record available at https://lccn.loc.gov/2021000638LC ebook record available at https://lccn.loc.gov/2021000639
Cover Design: WileyCover Image: © Zaie/Shutterstock
To Lorna, Sebastian and ImogenTo Phil and EllieTo Simon, Catriona and Kirstin
Acknowledgements
The Editors would like to acknowledge the support of their colleagues in their host institutions and departments: the School of Medicine, Keele University; the Centre for Reviews and Dissemination, University of York; and the MRC Clinical Trials Unit at University College London. Particular thanks to Lucinda Archer, John Allotey, Sarah Burdett, Thomas Debray, Miriam Hattle, Mel Holden, Carl Moons, Max Parmar, Bob Phillips, Larysa Rydewska, Mark Simmonds, Kym Snell, Shakila Thangaratinam, and Claire Vale, who have worked with us on many of our applied IPD meta‐analysis projects, and to Danielle Burke and Joie Ensor for organising the Statistical Methods for IPD Meta‐Analysis course at Keele. We are particularly grateful to Mike Clarke who has shared the IPD journey from the start and has contributed so much to the field. We also recognise the contributions by the convenors and other members of the Cochrane IPD Meta‐Analysis Methods Group over many years, and acknowledge stimulating discussions with participants at our various workshops and training courses. We thank our colleagues and research collaborators on the various applied and methodological IPD meta‐analysis projects we have been involved in, many of which formed motivating examples and case studies in the book chapters. In particular, we are indebted to the participants in the various trials and studies, and the associated investigators, without whom IPD meta‐analysis projects would not be possible.
Part IRationale, Planning, and Conduct
2Rationale for Embarking on an IPD Meta‐Analysis Project
Jayne F. Tierney, Richard D. Riley, Catrin Tudur Smith, Mike Clarke, and Lesley A. Stewart
Summary Points
Many of the principles, methods and processes of IPD meta‐analysis projects are similar to those of a conventional systematic review and meta‐analysis of aggregate data. The most substantial differences relate to the collection, checking and analysis of data at the participant level, and collaboration with the investigators responsible for the existing trials.
Compared to using aggregate data, IPD projects can potentially provide substantial improvements to the extent and quality of data available, and give greater scope and flexibility in the analyses, for example to examine participant‐level associations.
Important differences can occur between IPD and aggregate data meta‐analysis results. This depends on many aspects including the availability of IPD, whether IPD leads to improvements in the completeness and quantity of information, and how analysis methods planned by the IPD researchers compare with those done by original trial investigators.
Given the additional resource requirements, it is important to consider carefully whether an IPD project is needed instead of a conventional systematic review using aggregate data. The decision will depend on the particular research question, and whether IPD would produce a more reliable and comprehensive answer than using the aggregate data already available for eligible trials.
2.1 Introduction
In this chapter, we overview those elements of an IPD meta‐analysis project that differ from a conventional meta‐analysis of aggregate data (Section 2.2), describe the advantages (Section 2.3) and challenges of the IPD approach (Section 2.4), and summarise empirical evidence comparing results of IPD and aggregate data meta‐analyses (Section 2.5). Although IPD projects almost always provide advantages, sometimes a standard aggregate data meta‐analysis may be sufficient to answer a particular research question. Hence, researchers should only embark on an IPD project after careful consideration, especially as it requires additional time, resources and skills. We provide guidance to help researchers decide when the use of IPD is likely to provide more robust conclusions than using available aggregate data alone (Section 2.6). We focus on the synthesis of evidence from randomised trials evaluating treatment effects, but most of what is presented also applies to other study types and to other types of research questions, such as those for diagnosis and prognosis (Part 5).
2.2 How Does the Research Process Differ for IPD and Aggregate Data Meta‐Analysis Projects?
IPD meta‐analysis projects follow many of the same principles and research processes as conventional systematic reviews and meta‐analyses of aggregate data. However, there are also important differences, as now described.
2.2.1 The Research Aims
A first and fundamental step of all research projects is to define their aims. As for systematic reviews and meta‐analyses based on aggregate data, the aims of an IPD meta‐analysis project should be defined in relation to key components such as the participants, interventions, comparators or controls, outcomes and study designs of interest, aided by a framework such as PICOS (an example is given in Section 3.3).42 Most reviews based on aggregate data focus on summarising the overall treatment effect, and often IPD meta‐analysis projects also have this objective. However, IPD additionally allows participant‐level information to be examined and analysed, and so most IPD projects are specifically set up to utilise this. In particular, they may aim to summarise treatment effects conditional on prognostic factors (Chapters 5 and 6); to assess whether the treatment effect varies according to participant‐level characteristics (Chapter 7), or to evaluate treatment effects at multiple time‐points during follow‐up (Chapter 13). Indeed, the potential research questions that can be addressed by an IPD project are broad, and a wide variety of applications are demonstrated throughout this book.
2.2.2 The Process and Methods
Figure 2.1 provides key differences in the process of conducting IPD meta‐analysis projects compared to conventional reviews based on aggregate data.7,43 Best practice is to publish and adhere to a protocol, regardless of whether aggregate data or IPD are being used, although protocols for IPD projects will usually be more detailed (Section 4.2.2). Methods for identifying trials are very similar in the two approaches, but in an IPD project searches may be conducted prior to or in tandem with protocol development, in order to generate a preliminary list of trials (Section 4.2.3), and to identify the associated investigators from whom IPD will be sought (Section 3.2).
Prior to data collection, an IPD project may require ethical approval (Section 3.10) and development of formal data‐sharing agreements (Section 3.11), as well as the preparation of a detailed data dictionary (Section 4.2.7). These are rarely required for an aggregate data review. Furthermore, the subsequent data collection, checking and analytical aspects of an IPD project are much more exacting than those for aggregate data reviews. They may include data entry, data re‐coding and harmonisation, together with checking, querying and subsequent validation of IPD with original trial investigators (Chapter 4),7,43,44 as well as advanced statistical methods for meta‐analysis (Part 2).
Unlike aggregate data reviews, IPD projects usually involve and benefit from establishing partnerships with trial investigators who, in addition to providing their IPD, play an active role throughout the process, from identifying relevant trials through to helping interpret and disseminate IPD meta‐analysis results. This may include establishing a collaborative group that authors the main project publication, with all those involved being listed as co‐authors, and holding a meeting of this Group where preliminary results are presented and discussed (Section 3.8).7,43 Recently, a range of clinical study data repositories and platforms have been established, offering another source of IPD from existing trials, but there are both advantages and disadvantages of obtaining for IPD meta‐analysis projects in this way (Sections 3.2.2 and 4.4.5).45
Figure 2.1 Key differences between the process for a IPD meta‐analysis project and a conventional systematic review and meta‐analysis of aggregate data.
Source: Jayne Tierney.
Given these differences, IPD meta‐analysis projects require a greater range of skills (Section 3.5), generally take longer (Section 3.7), and need more resources (Section 3.8) than traditional systematic reviews and meta‐analyses based on aggregate data.
2.3 What Are the Potential Advantages of an IPD Meta‐Analysis Project?
Provided it is conducted appropriately, an IPD meta‐analysis project offers many advantages over the conventional aggregate data approach (Table 2.1).7,9,43,44 A key benefit is the potential to improve the quantity and quality of data, because there is no need to be limited by what has been published. For example, IPD from unpublished trials can be included (Section 4.2.3), as can any outcomes that were not reported for published trials, or even participants who were inappropriately excluded from the original trial analyses.7,9,43 As well as helping to circumvent potential reporting biases,46 this can increase the quantity of information available for analysis and, therefore, boost the statistical power to detect genuine effects.47 In addition, there is greater ability to standardise outcome and covariate definitions across trials (Section 4.5), which not only facilitates the conduct of meta‐analysis, but also aids the interpretation of findings. Detailed data checking helps to ensure the completeness, validity and internal consistency of data items for each trial, further enhancing data quality (Section 4.5.4),7,9,43,44 as well as providing independent scrutiny of the trial data.
Table 2.1 Key potential advantages of an IPD meta‐analysis project compared with a conventional systematic review and meta‐analysis of aggregate data focusing on the synthesis of randomised trials to evaluate treatment effects, adapting those shown by Tierney et al.9
Source: Adapted from Tierney et al.,9 with permission, © 2015 Tierney et al. (CC BY 4.0).
Aspect of systematic review or meta‐analysis
Advantages of an IPD meta‐analysis project
Trial identification and inclusion
Ask collaborative group (trial investigators and other experts in the clinical field) to help identify eligible trials (particularly those that are unpublished or ongoing)
*
Clarify a trial’s eligibility with the trial’s investigators
*
Data completeness and uniformity
Include data from trials that are unpublished or not reported in full
*
Include unreported data (e.g. unpublished subgroups, outcomes and time‐points), more complete information on outcomes, and data on participants excluded from original trial analyses
*
Check each trial’s IPD for completeness, validity and consistency, and resolve any queries with trial investigators
Derive new or standardised outcome definitions across trials or translate different definitions to a common scale
Derive new or standardised classifications of participant‐level characteristics, or translate different definitions to a common scale
Update follow‐up of time‐to‐event or other time‐related outcomes beyond those reported
*
Risk of bias assessment
Clarify trial design, conduct and analysis methods with trial investigators
*
Resolve unclear risk of bias assessments (i.e. based on trial reports) through direct contact with investigators
*
Examine trial IPD directly for evidence of potential bias in trial design and conduct, and resolve any queries with trial investigators
Obtain extra data where necessary to alleviate or mitigate against potential biases
*
Analyses
Apply a consistent method of analysis for each trial (independent of original trial analyses)
Analyse all important outcomes irrespective of whether published
*
Explore validity of analytical assumptions e.g. normality of residuals in a linear regression analysis
Derive outcomes and measures of effect directly from IPD (independent of trial reporting), potentially at multiple time‐points of interest
Use a consistent unit of analysis for each trial (e.g. consistently analyse preterm birth events per mother rather a mix of per mother and per baby in trials that include twin pregnancies)
Account for complexities in each trial in the analysis, such as cluster randomised trials or multi‐centre trials
Analyse continuous outcomes on their continuous scale and adjust for baseline value
Adjust for a pre‐defined set of prognostic factors
Apply consistent definitions for categorised data (e.g. stage of cancer)
Conduct more detailed and appropriate analysis of time‐to‐event outcomes (e.g. handling of censored observations, generating Kaplan Meier curves, examination of non‐proportional hazards)
Achieve greater power for assessing interactions between effects of interventions and participant‐level characteristics
Model associations at the participant level, including potential non‐linear relationships
Use appropriate but non‐standard models (e.g. that account for repeated measurements or correlation between multiple outcomes) or measures of effect
Explain potential heterogeneity and inconsistency in network meta‐analysis
Address additional important questions over and above efficacy, or not considered by original trials e.g. to explore the natural history of disease, prognostic factors or surrogate outcomes
Interpretation
Discuss implications for clinical practice and research with a multi‐disciplinary group of collaborators including trial investigators who supplied data, and patient research partners
*
Dissemination
Achieve more widespread dissemination though collaborative group networks and patient groups
*These advantages accrue from direct contact with trial investigators (rather than the IPD per se), so potentially could be achieved for conventional systematic reviews if more active communication with trial investigators were adopted. This is seldom done in practice.
In general, having access to IPD also supports more flexible and sophisticated analyses than are possible with only existing aggregate data. IPD are vital for a thorough investigation of participant‐level associations, for example to identify treatment effect modifiers (Chapter 7).7,9,43
