Description
Book SynopsisA comprehensive and practical resource for analyses of crossover designs
For ethical reasons, it isvital to keep the number of patients in a clinical trial aslow as possible.As evidenced by extensive research publications, crossover designcan bea useful and powerfultool to reduce the number of patients needed for a parallel group design in studying treatmentsfor non-curable chronic diseases.
This book introduces commonly-used and well-established statistical tests and estimators in epidemiology that can easily be applied to hypothesis testing and estimation of the relative treatment effect for various types of data scale in crossover designs. Models with distribution-free random effects are assumed and hence most approaches considered here are semi-parametric. The book provides clinicians and biostatisticians with the exact test procedures and exact interval estimators, which are applicable even when the number of patients in a crossover trial is small. Systemat
Table of Contents
About the author xi
Preface xii
About the companion website xiv
1 Crossover design – definitions, notes, and limitations 1
1.1 Unsuitability for acute or most infectious diseases 2
1.2 Inappropriateness for treatments with long-lasting effects 2
1.3 Loss of efficiency in the presence of carry-over effects 3
1.4 Concerns of treatment-by-period interaction 3
1.5 Flaw of the commonly used two-stage test procedure 4
1.6 Higher risk of dropping out or being lost to follow-up 4
1.7 More assumptions needed in use of a crossover design 5
1.8 General principle and conditional approach used in the book 5
2 AB/BA design in continuous data 7
2.1 Testing non-equality of treatments 10
2.2 Testing non-inferiority of an experimental treatment to an active control treatment 11
2.3 Testing equivalence between an experimental treatment and an active control treatment 12
2.4 Interval estimation of the mean difference 13
2.5 Sample size determination 16
2.5.1 Sample size for testing non-equality 16
2.5.2 Sample size for testing non-inferiority 17
2.5.3 Sample size for testing equivalence 18
2.6 Hypothesis testing and estimation for the period effect 19
2.7 Estimation of the relative treatment effect in the presence of differential carry-over effects 21
2.8 Examples of SAS programs and results 22
Exercises 27
3 AB/BA design in dichotomous data 30
3.1 Testing non-equality of treatments 34
3.2 Testing non-inferiority of an experimental treatment to an active control treatment 36
3.3 Testing equivalence between an experimental treatment and an active control treatment 39
3.4 Interval estimation of the odds ratio 40
3.5 Sample size determination 42
3.5.1 Sample size for testing non-equality 42
3.5.2 Sample size for testing non-inferiority 42
3.5.3 Sample size for testing equivalence 43
3.6 Hypothesis testing and estimation for the period effect 45
3.7 Testing and estimation for carry-over effects 47
3.8 SAS program codes and likelihood-based approach 48
Exercises 51
4 AB/BA design in ordinal data 57
4.1 Testing non-equality of treatments 62
4.2 Testing non-inferiority of an experimental treatment to an active control treatment 64
4.3 Testing equivalence between an experimental treatment and an active control treatment 65
4.4 Interval estimation of the generalized odds ratio 66
4.5 Sample size determination 67
4.5.1 Sample size for testing non-equality 67
4.5.2 Sample size for testing non-inferiority 68
4.5.3 Sample size for testing equivalence 68
4.6 Hypothesis testing and estimation for the period effect 70
4.7 SAS codes for the proportional odds model with normal random effects 72
Exercises 74
5 AB/BA design in frequency data 75
5.1 Testing non-equality of treatments 78
5.2 Testing non-inferiority of an experimental treatment to an active control treatment 81
5.3 Testing equivalence between an experimental treatment and an active control treatment 83
5.4 Interval estimation of the ratio of mean frequencies 84
5.5 Sample size determination 86
5.5.1 Sample size for testing non-equality 86
5.5.2 Sample size for testing non-inferiority 87
5.5.3 Sample size for testing equivalence 88
5.6 Hypothesis testing and estimation for the period effect 88
5.7 Estimation of the relative treatment effect in the presence of differential carry-over effects 90
Exercises 92
6 Three-treatment three-period crossover design in continuous data 95
6.1 Testing non-equality between treatments and placebo 102
6.2 Testing non-inferiority of an experimental treatment to an active control treatment 103
6.3 Testing equivalence between an experimental treatment and an active control treatment 104
6.4 Interval estimation of the mean difference 104
6.5 Hypothesis testing and estimation for period effects 105
6.6 Procedures for testing treatment-by-period interactions 107
6.7 SAS program codes and results for constant variance 109
Exercises 111
7 Three-treatment three-period crossover design in dichotomous data 115
7.1 Testing non-equality of treatments 121
7.1.1 Asymptotic test procedures 121
7.1.2 Exact test procedures 123
7.1.3 Procedures for simultaneously testing non-equality of two experimental treatments versus a placebo 124
7.2 Testing non-inferiority of an experimental treatment to an active control treatment 126
7.3 Testing equivalence between an experimental treatment and an active control treatment 127
7.4 Interval estimation of the odds ratio 129
7.5 Hypothesis testing and estimation for period effects 131
7.6 Procedures for testing treatment-by-period interactions 133
7.7 SAS program codes and results for a logistic regression model with normal random effects 136
Exercises 138
8 Three-treatment three-period crossover design in ordinal data 141
8.1 Testing non-equality of treatments 150
8.1.1 Asymptotic test procedures 150
8.1.2 Exact test procedure 152
8.2 Testing non-inferiority of an experimental treatment to an active control treatment 153
8.3 Testing equivalence between an experimental treatment and an active control treatment 153
8.4 Interval estimation of the GOR 154
8.5 Hypothesis testing and estimation for period effects 156
8.6 Procedures for testing treatment-by-period interactions 159
8.7 SAS program codes and results for the proportional odds model with normal random effects 160
Exercises 162
9 Three-treatment three-period crossover design in frequency data 164
9.1 Testing non-equality between treatments and placebo 170
9.2 Testing non-inferiority of an experimental treatment to an active control treatment 173
9.3 Testing equivalence between an experimental treatment and an active control treatment 174
9.4 Interval estimation of the ratio of mean frequencies 175
9.5 Hypothesis testing and estimation for period effects 178
9.6 Procedures for testing treatment-by-period interactions 179
Exercises 181
10 Three-treatment (incomplete block) crossover design in continuous and dichotomous data 183
10.1 Continuous data 185
10.1.1 Testing non-equality of treatments 188
10.1.2 Testing non-equality between experimental treatments (or non-nullity of dose effects) 189
10.1.3 Interval estimation of the mean difference 190
10.1.4 SAS codes for fixed effects and mixed effects models 192
10.2 Dichotomous data 194
10.2.1 Testing non-equality of treatments 197
10.2.2 Testing non-equality between experimental treatments (or non-nullity of dose effects) 199
10.2.3 Testing non-inferiority of either experimental treatment to an active control treatment 199
10.2.4 Interval estimation of the odds ratio 200
10.2.5 SAS codes for the likelihood-based approach 202
Exercises 203
References 208
Index 216
