Statistical Issues in Drug Development

Preface to the Third Edition

Preface to the Second Edition xiii

Preface to the First Edition xvii

Acknowledgements xxi

1 Introduction 1

• 1.1 Drug development 1
• 1.2 The role of statistics in drug development 2
• 1.3 The object of this book 3
• 1.4 The author’s knowledge of statistics in drug development 4
• 1.5 The reader and his or her knowledge of statistics 4
• 1.6 How to use the book 5
• References 6

Part 1 Four Views of Statistics in Drug Development: Historical, Methodological, Technical and Professional 9

2 A Brief and Superficial History of Statistics for Drug Developers 11

• 2.1 Introduction 11
• 2.2 Early Probabilists 12
• 2.3 James Bernoulli (1654–1705) 13
• 2.4 John Arbuthnott (1667–1753) 14
• 2.5 The mathematics of probability in the late 17th, the 18th and early 19th centuries 14
• 2.6 Thomas Bayes (1701–1761) 15
• 2.7 Adolphe Quetelet (1796–1874) 16
• 2.8 George Biddell Airy (1801-1892)
• 2.9 Francis Galton (1822–1911) 16
• 2.10 Karl Pearson (1857–1936) 17
• 2.11 ‘Student’ (1876–1937) 17
• 2.12 R.A. Fisher (1890–1962) 17
• 2.13 Modern mathematical statistics 18
• 2.14 Medical statistics 19
• 2.15 Statistics in clinical trials today 20
• 2.16 The current debate 21
• 2.17 A living science 21
• 2.18 Further reading 23
• References 23

3 Design and Interpretation of Clinical Trials as Seen by a Statistician 27

• 3.1 Prefatory warning 27
• 3.2 Introduction 27
• 3.3 Defining effects 28
• 3.4 Practical problems in using the counterfactual argument 28
• 3.5 Regression to the mean 29
• 3.6 Control in clinical trials 33
• 3.7 Randomization 34
• 3.8 Blinding 36
• 3.9 Using concomitant observations 37
• 3.10 Measuring treatment effects 38
• 3.11 Data generation models 39
• 3.12 In conclusion 41
• 3.13 Further reading 41
• References 41

4 Probability, Bayes, P-values, Tests of Hypotheses and Confidence Intervals 43

• 4.1 Introduction 43
• 4.2 An example 44
• 4.3 Odds and sods 44
• 4.4 The Bayesian solution to the example 45
• 4.5 Why don’t we regularly use the Bayesian approach in clinical trials? 46
• 4.6 A frequentist approach 47
• 4.7 Hypothesis testing in controlled clinical trials 48
• 4.8 Significance tests and P-values 49
• 4.9 Confidence intervals and limits and credible intervals 50
• 4.10 Some Bayesian criticism of the frequentist approach 51
• 4.11 Decision theory 51
• 4.12 Conclusion 52
• 4.13 Further reading 52
• References 53

5 The Work of the Pharmaceutical Statistician 55

• 5.1 Prefatory remarks 55
• 5.2 Introduction 56
• 5.3 In the beginning 57
• 5.4 The trial protocol 57
• 5.5 The statistician’s role in planning the protocol 58
• 5.6 Sample size determination 59
• 5.7 Other important design issues 60
• 5.8 Randomization 60
• 5.9 Data collection preview 61
• 5.10 Performing the trial 61
• 5.11 Data analysis preview 61
• 5.12 Analysis and reporting 62
• 5.13 Other activities 63
• 5.14 Statistical research 63
• 5.15 Further reading 64
• References 65

Part 2 Statistical Issues: Debatable and Controversial Topics in Drug Development 67

6 Allocating Treatments to Patients in Clinical Trials 69

• 6.1 Background 69
• 6.2 Issues 71
• References 87
• 6.A Technical appendix 88

7 Baselines and Covariate Information 95

• 7.1 Background 95
• 7.2 Issues 98
• References 108
• 7.A Technical appendix 109

8 The Measurement of Treatment Effects 113

• 8.1 Background 113
• 8.2 Issues 114
• References 129
• 8.A Technical appendix 130

9 Demographic Subgroups: Representation and Analysis 133

• 9.1 Background 133
• 9.2 Issues 134
• References 144
• 9.A Technical appendix 145

10 Multiplicity 149

• 10.1 Background 149
• 10.2 Issues 150
• References 161
• 10.A Technical appendix 162

11 Intention to Treat, Missing Data and Related Matters 165

• 11.1 Background 165
• 11.2 Issues 167
• References 178
• 11.A Technical appendix 180

12 One-sided and Two-sided Tests and other Issues to Do with Significance and P-values 183

• 12.1 Background 183
• 12.2 Issues 184
• References 192

13 Determining the Sample Size 195

• 13.1 Background 195
• 13.2 Issues 198
• References 211

14 Multicentre Trials 213

• 14.1 Background 213
• 14.2 Issues 213
• References 230
• 14.A Technical appendix 231

15 Active Control Equivalence Studies 235

• 15.1 Background 235
• 15.2 Issues 237
• References 247
• 15.A Technical appendix 249

16 Meta-Analysis 251

• 16.1 Background 251
• 16.2 Issues 253
• References 268
• 16.A Technical appendix 270

17 Cross-over Trials 273

• 17.1 Background 273
• 17.2 Issues 275
• References 284

18 n-of-1 Trials 287

• 18.1 Background 287
• 18.2 Issues 289
• References 293

19 Sequential Trials 295

• 19.1 Background 295
• 19.2 Issues 302
• References 313

20 Dose-finding 317

• 20.1 Background 317
• 20.2 Issues 319
• References 334

21 Concerning Pharmacokinetics and Pharmacodynamics 337

• 21.1 Background 337
• 21.2 Issues 343
• References 358

22 Bioequivalence Studies 361

• 22.1 Background 361
• 22.2 Issues 362
• References 379

23 Safety Data, Harms, Drug Monitoring and Pharmaco-epidemiology 383

• 23.1 Background 383
• 23.2 Issues 388
• References 403

24 Pharmaco-economics and Portfolio Management 405

• 24.1 Background 405
• 24.2 Issues 407
• References 429

25 Concerning Pharmacogenetics, Pharmacogenomics and Related Matters 433

• 25.1 Background 433
• 25.2 Issues 437
• References 450
• 25.A Technical appendix 451

Glossary 453

Index 483

This will be the third edition of Statistical Issues in Drug Development, and will be fully revised and updated to include information on the latest industry standards and guidelines. Both the first (1997) and second (2007) editions were very well received and the book has become a standard.

This book is unique in providing a thorough and critical discussion of the most important and controversial issues encountered by statisticians and their life scientist colleagues on both sides of the regulatory divide in drug development. The primary purpose of the book is to encourage and facilitate discussion between statisticians and their colleagues of the many complex statistical issues that arise in drug development. The book will be suitable as a course of self-instruction for statisticians who are new to the pharmaceutical industry, either because of recent graduation or change of career. It will also act as an authoritative reference for those working in drug development, and provide possible topics for discussion in journal forums.