Multiple Testing Problems in Pharmaceutical Statistics (Chapman & Hall CRC Biostatistics Series)

Front matter......Page 1
Multiple Testing Problems in Pharmaceutical Statistics......Page 4
Contents......Page 6
Outline of the book......Page 11
Acknowledgments......Page 13
List of Contributors......Page 14
1.1 Introduction......Page 16
1.2 Common multiplicity problems in clinical trials......Page 18
1.2.1 Clinical win criteria with primary endpoints......Page 19
1.2.2 Consistency-ensured analysis of multiple endpoints......Page 22
1.2.3 Co-primary endpoints......Page 23
1.2.4 Secondary endpoints......Page 26
1.2.5 Dose-response and dose-control comparisons......Page 27
1.2.6 Subgroup analyses......Page 29
1.2.7 Combination drug trials......Page 31
1.2.8 Three-arm trials with an active control......Page 32
1.2.9 Advanced multiplicity problems......Page 35
Example: Gastric ulcer trial......Page 38
Example: A dose-finding hypertension trial......Page 39
1.3.2 Composite endpoints......Page 40
1.4.1 Use of multiple subsets of patient data......Page 43
1.4.3 Analysis of covariance......Page 44
1.5 Multiplicity in the analysis of safety endpoints......Page 45
1.6 Concluding remarks......Page 47
2.1 Introduction......Page 49
2.2.1 Comparisonwise and familywise error rates......Page 50
Generalized familywise error rate......Page 52
2.2.3 Role of Type II errors......Page 53
2.3.1 Union-intersection testing......Page 54
2.3.3 Closure principle......Page 55
Motivating example......Page 56
General case......Page 57
α-exhaustive procedures......Page 58
Motivating example......Page 59
2.4.1 Adjusted significance levels and p-values......Page 60
2.4.2 Simultaneous confidence intervals......Page 61
Single-step and stepwise procedures......Page 62
Distributional assumptions......Page 63
2.5.3 Dose-finding trial example......Page 64
2.6.1 Bonferroni procedure......Page 66
2.6.2 Holm procedure and its extensions......Page 67
Shaffer procedure......Page 69
2.6.3 Fixed-sequence procedure......Page 70
2.6.4 Fallback procedure......Page 71
2.6.5 Bonferroni-based closed testing procedures......Page 72
Sequentially rejective Bonferroni-based closed testing procedures......Page 73
Simultaneous confidence intervals......Page 74
Graphical visualization......Page 75
2.6.6 Reverse fixed-sequence procedure......Page 76
2.6.7 Simes global test......Page 77
2.6.8 Hommel procedure......Page 79
2.6.9 Hochberg procedure......Page 80
Bonferroni procedure......Page 81
Dose-finding trial example......Page 82
Holm procedure......Page 84
Fallback procedure......Page 85
Dose-finding trial example......Page 86
2.7 Parametric multiple testing procedures......Page 87
2.7.1 Single-step Dunnett procedure......Page 88
Step-down Dunnett procedure......Page 89
Step-up Dunnett procedure......Page 90
Step-down Dunnett procedure......Page 92
Step-up Dunnett procedure......Page 93
Dose-finding trial example......Page 94
2.7.5 Multiple comparisons in general linear models......Page 95
2.8 Resampling-based multiple testing procedures......Page 96
2.8.1 Closed resampling-based procedures......Page 97
2.8.2 Step-down Dunnett procedures based on parametric and nonparametric resampling......Page 99
Nonparametric resampling......Page 100
2.8.3 Step-down resampling-based procedures for multivariate linear models......Page 102
2.8.4 Permutation-based procedures......Page 104
2.9.1 Multiple comparison procedures in SAS......Page 108
2.9.2 Multiple comparison procedures in R......Page 109
Acknowledgements......Page 112
3.1 Introduction......Page 113
3.2.1 Notation......Page 115
3.2.2 Contrast tests......Page 116
3.2.3 Multiple contrast tests for general linear models......Page 118
3.2.4 Thalidomide study example......Page 119
3.3 Target dose estimation using multiple hypothesis testing......Page 123
3.3.1 Estimation of the minimum effective dose......Page 125
SD1PC Procedure......Page 126
Simultaneous confidence intervals......Page 127
Estimation of MINED in a major depressive disorder trial......Page 128
3.3.2 Estimation of the maximum safe dose......Page 129
Estimation of MAXSD in a toxicology study......Page 130
3.4 Power and sample size calculation for target dose estimation......Page 131
3.5 Hybrid approaches combining multiple testing and modeling......Page 132
3.5.1 Modeling techniques......Page 133
3.5.2 MCP-Mod: An integrated dose-finding approach combining multiple comparisons and modeling......Page 135
Step 1......Page 136
Step 2......Page 137
Step 4......Page 138
3.5.3 Asthma dose-finding trial......Page 139
Step 2: Optimal model contrasts......Page 140
Step 4: Model selection......Page 142
Acknowledgements......Page 143
4.1 Introduction......Page 145
4.2.1 At-least-one procedures......Page 147
4.2.2 Global procedures......Page 148
4.2.4 Superiority-noninferiority procedures......Page 149
4.3.1 Procedures based on univariate p-values......Page 150
Bonferroni and related procedures......Page 151
Fixed-sequence procedure......Page 152
Fallback procedure......Page 153
Comparison of the fixed-sequence and fallback procedures......Page 154
Adaptive alpha allocation approach......Page 156
Bonferroni-type parametric procedure......Page 157
Fallback-type parametric procedure......Page 158
Resampling-based procedures......Page 160
4.4.1 Normal theory model......Page 161
OLS and GLS test statistics......Page 162
Power calculations for OLS and GLS procedures......Page 164
Nonparametric global procedures......Page 166
Exact likelihood ratio procedures......Page 167
Lauter exact procedure......Page 168
4.4.4 Procedures for individual endpoints......Page 169
Osteoarthritis trial example......Page 170
4.5 All-or-none procedures......Page 171
4.6 Superiority-noninferiority procedures......Page 173
Tamhane-Logan superiority-noninferiority procedure......Page 174
Alzheimer’s disease trial example......Page 175
4.7 Software implementation......Page 177
Acknowledgements......Page 178
5.1 Introduction......Page 179
5.2 Motivating examples......Page 180
5.2.1 Clinical trials with serial gatekeepers......Page 181
5.2.2 Clinical trials with parallel gatekeepers......Page 182
5.2.3 Clinical trials with tree-structured gatekeepers......Page 184
5.3 Serial gatekeeping procedures......Page 185
5.3.1 General serial gatekeeping framework......Page 186
Alzheimer’s disease clinical trial example......Page 187
Type II diabetes clinical trial example......Page 189
5.3.4 Other serial gatekeeping procedures......Page 190
5.4.1 General parallel gatekeeping framework......Page 191
5.4.2 Multistage parallel gatekeeping procedures......Page 192
Truncated multiple tests......Page 193
Multistage testing algorithm......Page 194
Computation of adjusted p-values......Page 195
5.4.3 Cardiovascular clinical trial example......Page 196
5.5.1 General tree gatekeeping framework......Page 200
Closure-based tree gatekeeping procedures......Page 201
5.5.2 Combination-therapy clinical trial example......Page 202
5.6 Software implementation......Page 203
Weight assignment algorithm for Bonferroni tree gatekeeping procedures......Page 204
6.1 Introduction......Page 206
6.2.2 Causes of multiplicity and bias in adaptive designs......Page 209
6.2.3 Repeated hypothesis testing at interim analyses in group sequential designs......Page 210
Stopping boundaries......Page 211
Sample size adjustment for nuisance parameters......Page 213
Unblinded sample size adjustment......Page 214
Blinded sample size adjustment......Page 215
Sample size adjustment based on the observed effect size......Page 216
6.2.5 Multiple hypothesis selection and testing in adaptive designs......Page 217
Combining repeated and multiple testing......Page 218
Comparison of sample size needed for adaptive vs. separate trial strategies......Page 221
6.3 Principles of adaptive testing procedures......Page 223
6.3.1 Combination test principle......Page 224
6.3.2 Conditional error principle......Page 225
Tests based on conditional error rates of standard testing procedures......Page 226
6.4.1 Adaptive closed tests......Page 227
Closed combination tests......Page 228
Closed conditional error tests......Page 229
6.4.2 Estimation......Page 231
6.5.1 Treatment selection design......Page 233
Data analysis: confirmatory hypothesis testing......Page 239
Decision making at the interim analysis......Page 240
Comparison to the conditional error rate approach......Page 241
6.5.2 Subpopulation selection design......Page 242
Adaptive subpopulation selection design: a discussion......Page 247
6.6 Discussion......Page 248
7.1 Potential uses of biomarkers......Page 251
7.2 Clinical uses of genetic profiling......Page 252
7.3 Two stages of pharmacogenomic development......Page 253
7.4 Multiplicity in pharmacogenomics......Page 255
Two different definitions of Type I error rate......Page 256
Popular error rates to control in bioinformatics......Page 257
Conditional versus unconditional inference......Page 258
Taking dependence into account......Page 259
7.4.2 Multiplicity of subgroups......Page 260
7.5.2 Design to discover group differences......Page 262
Replicate to estimate and remove variabilities......Page 263
Block to avoid confounding......Page 264
7.5.3 Permutation tests may not control error rates......Page 266
7.6 Analyzing microarray data by modeling......Page 267
7.7 A proof of concept experiment......Page 269
Normalization using an external reference sample......Page 270
Multiple testing for differential expressions......Page 273
7.8.1 SAS procedures......Page 274
7.8.2 R Functions......Page 275
Acknowledgements......Page 276
Bibliography......Page 277