Design and Analysis of Cross-Over Trials, Second Edition (Chapman & Hall CRC Monographs on Statistics & Applied Probability)

0412606402_01__SCLZZZZZZZ_......Page 1
The Design and Analysis of Cross-Over Trials, Second Edition......Page 2
Contents......Page 4
List of figures......Page 7
List of tables......Page 9
Preface to the second edition.......Page 17
Example 1.1......Page 20
Table of Contents......Page 0
1.2 With which sort of cross-over trial are we concerned?......Page 21
1.3 Why do cross-over trials need special consideration?......Page 22
1.4 A brief history......Page 24
1.5 Notation, Models and Analysis......Page 26
1.6 Aims of this book......Page 29
1.7 Structure of the book......Page 30
2.1 Introduction......Page 31
2.2 Plotting the data......Page 32
2.3 The analysis using t-tests......Page 39
2.4 Sample size calculations......Page 51
2.5 The analysis of variance......Page 56
2.6 Aliasing of e.ects......Page 61
2.7 Consequences of preliminary testing......Page 63
2.8 Analyzing the residuals......Page 71
2.9.1 Bayes using approximations......Page 74
2.9.2 Bayes using Gibbs sampling......Page 78
2.10 The use of baseline measurements......Page 81
2.11 The use of covariates......Page 91
2.12 Nonparametric analysis......Page 97
2.12.1 Testing λ1 = λ2......Page 99
2.12.2 Testing τ1 = τ2, given that λ1 = λ2......Page 103
2.12.4 Obtaining the exact version of the Wilcoxon rank-sum test using tables......Page 104
2.12.5 Point estimate and con.dence interval for d = t1 - t2......Page 105
2.12.6 A more general approach to nonparametric testing......Page 107
2.12.7 Nonparametric analysis of ordinal data......Page 114
2.12.8 Analysis of a multicentre trial......Page 117
2.12.9 Tests based on nonparametric measures of association......Page 118
2.13.1 Introduction......Page 129
2.13.2 McNemar’s test......Page 130
2.13.3 The Mainland-Gart test......Page 132
2.13.4 Fisher’s exact version of the Mainland-Gart test......Page 134
2.13.5 Prescott’s Test......Page 135
3.1 Introduction......Page 138
3.2 ‘Optimal’ designs......Page 139
3.3 Balaam’s design for two treatments......Page 141
3.4 The effect of preliminary testing in Balaam’s design......Page 143
3.5 Three-period designs with two sequences......Page 146
3.6 Three-period designs with four sequences......Page 151
3.8 Which three-period design to use?......Page 156
3.9 Four-period designs with two sequences......Page 158
3.10 Four-period designs with four sequences......Page 162
3.11 Four-period designs with six sequences......Page 163
3.12 Which four-period design to use?......Page 165
3.13 Which two-treatment design to use?......Page 166
4.1 Introduction......Page 170
Orthogonal Latin squares......Page 173
Designs obtained from orthogonal squares......Page 184
Extra-period designs......Page 189
Quenouille, Berenblut and Patterson designs......Page 191
Nearly strongly balanced designs......Page 194
4.3 Optimality results for cross-over designs......Page 196
4.4 Which variance balanced design to use?......Page 199
4.5 Partially balanced designs......Page 203
4.6 Comparing test treatments to a control......Page 209
4.7 Factorial treatment combinations......Page 212
4.8 Extending the simple model for carry-over e.ects......Page 216
4.9 Computer search algorithms......Page 217
Example 5.1 The INNOVO Trial: Dose-response Study......Page 224
5.2.1 Ignoring the baseline measurements......Page 225
5.2.2 Using the baseline measurements......Page 228
5.2.3 Adjusting for carry-over e.ects......Page 231
5.3.1 Random subject e.ects......Page 232
5.3.2 Recovery of between-subject information......Page 234
Example 5.2......Page 235
5.3.3 Small sample inference......Page 239
5.3.4 Missing values......Page 241
Example 5.3: Amantadine in Parkinsonism.......Page 244
5.4.2 Example 5.4: a two-treatment three-period design......Page 250
5.5 The general linear mixed model......Page 253
5.6 Analysis of repeated measurements within periods......Page 254
Example 5.5: Insulin Mixtures......Page 255
Example 5.6: Systolic Blood Pressures......Page 262
5.7.1 Allowing more general covariance structures......Page 265
Single dual pair designs......Page 267
Multiple dual pair designs......Page 269
Example 5.6......Page 276
5.8.1 Example 5.7: McNulty’s experiment......Page 282
5.8.3 Fixed e.ects analysis......Page 285
5.8.4 Random subject e.ects and covariance structure......Page 293
5.8.5 Modelling the period e.ects......Page 295
6.1.1 Modelling dependent categorical data......Page 300
Marginal models......Page 301
Subject-speci.c models......Page 303
The Mainland-Gart test......Page 305
The Mainland-Gart test in a logistic regression framework......Page 306
Small sample issues......Page 307
Conditional logistic regression......Page 310
Random subject e.ects......Page 312
Higher-order designs......Page 319
6.3.1 The marginal model......Page 322
6.4.1 Example 6.2: trial on patients with primary dysmenorrhea......Page 327
6.4.2 Types of model for categorical outcomes......Page 328
6.4.3 Subject e.ects models......Page 330
A generalized logit model......Page 332
A partial proportional odds model......Page 334
6.5.1 Count data......Page 337
6.5.3 Issues associated with scale......Page 339
7.1 What is bioequivalence......Page 341
7.2 Testing for average bioequivalence......Page 345
7.3 Power and sample size for ABE in the 2 × 2 design......Page 353
7.4 Individual bioequivalence......Page 357
7.5.1 PBE using a 2 × 2 design......Page 362
7.5.2 PBE using a replicate design......Page 363
7.6 ABE for a replicate design......Page 364
7.7 Kullback–Leibler divergence for evaluating bioequivalence......Page 365
7.8 Modelling pharmacokinetic data......Page 366
Appendix A: Least squares estimation......Page 369
Appendix B: SAS code for assessing ABE, IBE and PBE in replicate cross-over trials......Page 373
Bibliography......Page 377