Statistical Thinking for Non-Statisticians in Drug Regulation

✍ Scribed by Richard Kay

• Year: 2007
• Tongue: English
• Leaves: 297
• Edition: 1
• Category: Library

✦ Synopsis

Written by a well-known lecturer and consultant to the pharmaceutical industry, this book focuses on the pharmaceutical non-statistician working within a very strict regulatory environment. Statistical Thinking for Clinical Trials in Drug Regulation presents the concepts and statistical thinking behind medical studies with a direct connection to the regulatory environment so that readers can be clear where the statistical methodology fits in with industry requirements. Pharmaceutical-related examples are used throughout to set the information in context. As a result, this book provides a detailed overview of the statistical aspects of the design, conduct, analysis and presentation of data from clinical trials within drug regulation.Statistical Thinking for Clinical Trials in Drug Regulation:Assists pharmaceutical personnel in communicating effectively with statisticians using statistical languageImproves the ability to read and understand statistical methodology in papers and reports and to critically appraise that methodologyHelps to understand the statistical aspects of the regulatory framework better quoting extensively from regulatory guidelines issued by the EMEA (European Medicines Evaluation Agency), ICH (International Committee on Harmonization and the FDA (Food and Drug Administration)

✦ Table of Contents

Statistical Thinking for Non-Statisticians in Drug Regulation......Page 4
Contents......Page 10
Preface......Page 16
Abbreviations......Page 20
1.1 Historical perspective......Page 22
1.2 Control groups......Page 23
1.3 Placebos and blinding......Page 24
1.4 Randomisation......Page 25
1.4.2 Block randomisation......Page 26
1.4.3 Unequal randomisation......Page 27
1.4.4 Stratified randomisation......Page 28
1.4.5 Central randomisation......Page 29
1.4.6 Dynamic allocation and minimisation......Page 30
1.4.7 Cluster randomisation......Page 31
1.5 Bias and precision......Page 32
1.6 Between- and within-patient designs......Page 33
1.7 Cross-over trials......Page 35
1.8.3 Signal-to-noise ratio......Page 36
1.9 Confirmatory and exploratory trials......Page 37
1.10 Superiority, equivalence and non-inferiority trials......Page 38
1.11 Data types......Page 39
1.12.1 Primary variables......Page 41
1.12.3 Surrogate variables......Page 42
1.12.4 Global assessment variables......Page 43
1.12.6 Categorisation......Page 44
2.1 Sample and population......Page 46
2.2.1 Sample and population distribution......Page 47
2.2.2 Median and mean......Page 48
2.2.3 Standard deviation......Page 49
2.3 The normal distribution......Page 50
2.4 Sampling and the standard error of the mean......Page 53
2.5.1 The standard error for the difference between two means......Page 56
2.5.3 The general setting......Page 59
3.1.1 The 95 per cent confidence interval......Page 60
3.1.3 Changing the multiplying constant......Page 62
3.1.4 The role of the standard error......Page 64
3.2.1 Difference between two means......Page 65
3.2.2 Confidence intervals for proportions......Page 66
3.2.3 General case......Page 67
3.3.1 Interpreting the p-value......Page 68
3.3.2 Calculating the p-value......Page 70
3.3.3 A common process......Page 73
3.3.5 One-tailed and two-tailed tests......Page 76
4.1 The unpaired t-test......Page 78
4.2 The paired t-test......Page 79
4.3 Interpreting the t-tests......Page 82
4.4.1 Pearson chi-square......Page 84
4.4.2 The link to a signal-to-noise ratio......Page 87
4.5.1 Odds ratio (OR)......Page 88
4.5.2 Relative risk (RR)......Page 89
4.5.4 Number needed to treat (NNT)......Page 90
4.5.5 Confidence intervals......Page 91
4.6 Fisher’s exact test......Page 92
4.7.1 Categorical data......Page 94
4.7.2 Ordered categorical (ordinal) data......Page 96
4.7.3 Measures of treatment benefit for categorical and ordinal data......Page 97
4.8.1 Between-patient designs and continuous data......Page 98
4.8.2 Within-patient designs and continuous data......Page 99
4.8.4 Dose ranging studies......Page 100
4.8.5 Further discussion......Page 101
5.1 Rationale for multi-centre trials......Page 102
5.2 Comparing treatments for continuous data......Page 103
5.3.1 Treatment-by-centre interactions......Page 105
5.3.2 Quantitative and qualitative interactions......Page 108
5.5 Combining centres......Page 109
6.1 Adjusting for baseline factors......Page 112
6.2 Simple linear regression......Page 113
∗6.3 Multiple regression......Page 115
6.4 Logistic regression......Page 117
6.5.1 Main effect of treatment......Page 118
6.5.2 Treatment-by-covariate interactions......Page 120
∗6.5.3 A single model......Page 122
6.5.5 Advantages of analysis of covariance......Page 123
6.6 Binary, categorical and ordinal data......Page 125
6.7 Regulatory aspects of the use of covariates......Page 127
6.9 Baseline testing......Page 130
7.1 The principle of intention-to-treat......Page 132
7.2.1 Full analysis set......Page 136
7.2.3 Sensitivity......Page 138
7.3.1 Introduction......Page 139
7.3.3 Last observation carried forward (LOCF)......Page 140
7.3.5 Worst case/best case imputation......Page 141
7.3.7 Avoidance of missing data......Page 142
7.4 Intention-to-treat and time-to-event data......Page 143
7.5 General questions and considerations......Page 145
8.1 Type I and type II errors......Page 148
8.2 Power......Page 149
8.3 Calculating sample size......Page 152
8.4.1 Standard deviation......Page 155
8.4.3 Clinically relevant difference......Page 156
8.5.1 Power > 80 per cent......Page 157
8.5.3 Sample size adjustment......Page 158
8.6 Reporting the sample size calculation......Page 159
9.1 Link between p-values and confidence intervals......Page 162
9.2 Confidence intervals for clinical importance......Page 164
9.3.1 Conclusions of similarity......Page 165
9.3.2 The problem with 0.05......Page 166
10.1 Inflation of the type I error......Page 168
10.3 Regulatory view......Page 169
10.4.2 Significance needed on all endpoints......Page 170
10.4.4 Variables ranked according to clinical importance......Page 171
10.5 Methods for adjustment......Page 173
10.6 Multiple comparisons......Page 174
10.7 Repeated evaluation over time......Page 175
10.8 Subgroup testing......Page 176
10.9.1 Using different statistical tests......Page 178
10.9.2 Different analysis sets......Page 179
11.1 Assumptions underlying the t-tests and their extensions......Page 180
11.3 The assumption of normality......Page 181
11.4 Transformations......Page 184
11.5.1 The Mann–Whitney U-test......Page 187
11.5.2 The Wilcoxon signed rank test......Page 189
11.6 Advantages and disadvantages of non-parametric methods......Page 190
11.7 Outliers......Page 191
12.1 Demonstrating similarity......Page 194
12.2 Confidence intervals for equivalence......Page 196
12.3 Confidence intervals for non-inferiority......Page 197
12.4 A p-value approach......Page 199
12.5 Assay sensitivity......Page 201
12.7 The choice of Δ......Page 203
12.7.2 Therapeutic equivalence......Page 204
12.7.3 Non-inferiority......Page 205
12.7.4 The 10 per cent rule for cure rates......Page 206
12.7.5 Biocreep and constancy......Page 207
12.8 Sample size calculations......Page 208
12.9 Switching between non-inferiority and superiority......Page 210
13.1 Time-to-event data and censoring......Page 214
13.2.1 Plotting KM curves......Page 216
13.2.3 Median event times......Page 217
13.3 Treatment comparisons......Page 218
13.4.1 The hazard rate......Page 221
13.4.3 Non-constant hazard ratio......Page 222
13.4.4 Link to survival curves......Page 223
∗13.4.5 Calculating KM curves......Page 224
13.5.2 Proportional hazards regression......Page 225
13.5.3 Accelerated failure time model......Page 228
13.6 Independent censoring......Page 229
13.7 Sample size calculations......Page 230
14.1 Stopping rules for interim analysis......Page 234
14.2.1 Efficacy......Page 235
14.2.2 Futility and conditional power......Page 236
14.2.3 Some practical issues......Page 237
14.2.4 Analyses following completion of recruitment......Page 238
14.3 Monitoring safety......Page 239
14.4.1 Introduction and responsibilities......Page 240
14.4.2 Structure......Page 241
14.4.3 Meetings and recommendations......Page 243
14.5.1 Sample size re-evaluation......Page 244
14.5.2 Flexible designs......Page 245
15.1 Definition......Page 250
15.2 Objectives......Page 252
15.3.1 Methods for combination......Page 253
15.3.2 Confidence Intervals......Page 254
15.3.4 Graphical methods......Page 255
15.3.6 Robustness......Page 257
15.4.1 Planning......Page 258
15.4.2 Publication bias and funnel plots......Page 259
15.5.1 Retrospective analyses......Page 261
15.5.2 One pivotal study......Page 262
16.1 The importance of statistical thinking at the design stage......Page 266
16.2 Regulatory guidelines......Page 268
16.3 The statistics process......Page 270
16.3.2 The statistical analysis plan......Page 271
16.3.4 The blind review......Page 272
16.3.6 Reporting the analysis......Page 273
16.3.7 Pre-planning......Page 274
16.3.8 Sensitivity and robustness......Page 276
16.4 The regulatory submission......Page 277
16.5 Publications and presentations......Page 278
References......Page 282
Index......Page 288