Statistics Applied to Clinical Trials

TABLE OF CONTENTS......Page 5
PREFACE......Page 13
FOREWORD......Page 15
1. General considerations......Page 16
2. Two main hypotheses in drug trials: efficacy and safety......Page 17
3. Different types of data: continuous data......Page 18
4. Different types of data: proportions, percentages and contingency tables......Page 23
5. Different types of data: correlation coefficient......Page 26
6. Stratification issues......Page 28
7. Randomized versus historical controls......Page 29
9. Conclusions......Page 30
10. References......Page 31
1. Overview......Page 32
2. The principle of testing statistical significance......Page 33
3. The T-Value = standardized mean result of study......Page 36
4. Unpaired T-Test......Page 37
5. Null -hypothesis testing of 3 or more unpaired samples......Page 39
6. Three methods to test statistically a paired sample......Page 40
7. Null-hypothesis testing of 3 or more paired samples......Page 43
8. Paired data with a negative correlation......Page 45
9. Rank testing......Page 51
11. References......Page 54
1. Introduction, summary display......Page 56
2. Four methods to analyze two unpaired proportions......Page 57
3. Chi-square to analyze more than two unpaired proportions......Page 63
4. McNemar’s test for paired proportions......Page 66
5. Survival analysis......Page 67
6. Odds ratio method for analyzing two unpaired proportions......Page 69
8. Conclusions......Page 72
1. Introduction......Page 74
2. Overview of possibilities with equivalence testing......Page 76
3. Calculations......Page 77
5. Validity of equivalence trials......Page 78
6. Special point: level of correlation in paired equivalence studies......Page 79
7. Conclusions......Page 80
1. What is statistical power......Page 82
2. Emphasis on statistical power rather than null-hypothesis testing......Page 83
3. Power computations......Page 85
4. Example of power computation using the T-Table......Page 86
6. Calculations of required sample size, methods......Page 88
7. Testing not only superiority but also inferiority of a new treatment (the type III error)......Page 91
9. References......Page 93
2. Monitoring......Page 94
3. Interim analysis......Page 95
5. Continuous sequential statistical techniques......Page 98
7. References......Page 100
2. Multiple comparisons......Page 102
3. Multiple variables......Page 107
5. References......Page 110
1. Introduction......Page 112
2. Bonferroni test......Page 113
4. Other tests for adjusting the p-values......Page 114
6. No adjustments at all, and pragmatic solutions......Page 115
8. References......Page 116
2. Renewed attention to the interpretation of the p-values......Page 117
3. Standard interpretation of p-values......Page 118
5. Renewed interpretations of p-values, little difference between p = 0.06 and p = 0.04......Page 120
6. The real meaning of very large p-values like p>0.95......Page 121
7. P-values larger than 0.95, examples (Table 2)......Page 122
8. The real meaning of very small p-values like p<0.0001......Page 123
9. P-values smaller than 0.0001, examples (Table 3)......Page 124
11. Recommendations......Page 125
13. References......Page 127
1. Introduction......Page 131
2. Methods and results......Page 132
3. Discussion......Page 133
5. References......Page 136
1. Introduction......Page 139
2. More on paired observations......Page 140
3. Using statistical software for simple linear regression......Page 143
4. Multiple linear regression......Page 145
5. Multiple linear regression, example......Page 147
6. Purposes of linear regression analysis......Page 151
7. Another real data example of multiple linear regression (exploratory purpose)......Page 152
8. Conclusions......Page 154
2. Example......Page 155
3. Model......Page 156
4. (I.) Increased precision of efficacy......Page 158
5. (II.) Confounding......Page 159
6. (III.) Interaction and synergism......Page 160
7. Estimation, and hypothesis testing......Page 161
8. Goodness-of-fit......Page 162
10. Conclusion......Page 163
11. References......Page 164
1. Introduction......Page 165
2. Methods, statistical model......Page 166
3. Results......Page 168
4. Discussion......Page 174
6. References......Page 176
2. Linear regression......Page 179
3. Logistic regression......Page 183
4. Cox regression......Page 185
5. Markow models......Page 188
6. Discussion......Page 189
8. References......Page 191
2. Regression modeling for improved precision of clinical trials, the underlying mechanism......Page 193
3. Regression model for parallel-group trials with continuous efficacy data......Page 195
4. Regression model for parallel-group trials with proportions or odds as efficacy data......Page 196
5. Discussion......Page 197
7. References......Page 199
2. Examples......Page 201
3. Logistic regression equation......Page 204
5. References......Page 205
2. What exactly is interaction, a hypothesized example......Page 207
3. How to test the presence of interaction effects statistically, a real data example......Page 210
4. Additional real data examples of interaction effects......Page 212
6. Conclusions......Page 217
7. References......Page 218
1. Introduction......Page 219
2. Examples......Page 220
5. Strict inclusion criteria......Page 222
6. Uniform data analysis......Page 223
7. Discussion, where are we now?......Page 231
9. References......Page 232
1. Introduction......Page 233
2. Mathematical model......Page 234
3. Hypothesis testing......Page 235
4. Statistical power of testing......Page 237
5. Discussion......Page 240
6. Conclusion......Page 241
7. References......Page 242
1. Introduction......Page 243
2. Assessment of carryover and treatment effect......Page 244
3. Statistical model for testing treatment and carryover effects......Page 245
4. Results......Page 246
5. Examples......Page 248
6. Discussion......Page 249
8. References......Page 250
1. Introduction......Page 253
2. Examples from the literature in which cross-over trials are correctly used......Page 255
3. Examples from the literature in which cross-over trials should not have been used......Page 257
4. Estimate of the size of the problem by review of hypertension trials published......Page 259
5. Discussion......Page 260
6. Conclusions......Page 261
7. References......Page 262
2. Some terminology......Page 263
3. Defining QOL in a subjective or objective way......Page 265
4. The patients’ opinion is an important independent-contributor to QOL......Page 266
5. Lack of sensitivity of QOL-assessments......Page 267
6. Odds ratio analysis of effects of patient characteristics on QOL data provides increased precision......Page 268
7. Discussion......Page 271
9. References......Page 272
1. Introduction......Page 274
2. Some terminology......Page 275
3. Genetics, genomics, proteonomics, data mining......Page 277
4. Genomics......Page 278
6. References......Page 282
2. Variances......Page 283
3. The normal distribution......Page 284
4. Null-hypothesis testing with the normal or t-distribution......Page 286
5. Relationship between the normal-distribution and chi-square distribution, null-hypothesis testing with chi-square distribution......Page 288
6. Examples of data where variance is more important than mean......Page 290
7. Chi-square can be used for multiple samples of data......Page 291
8. Discussion......Page 294
10. References......Page 295
2. Individual data available......Page 296
3. Individual data not available......Page 302
4. Discussion......Page 304
5. Conclusions......Page 305
6. References......Page 306
2. Examples......Page 307
3. An index for variability in the data......Page 308
4. How to analyze variability, one sample......Page 309
5. How to analyze variability, two samples......Page 311
6. How to analyze variability, three or more samples......Page 312
7. Discussion......Page 314
9. References......Page 315
2. Testing reproducibility of quantitative data (continuous data)......Page 317
3. Testing reproducibility of qualitative data (proportions and scores)......Page 320
5. Additional real data examples......Page 322
7. Conclusions......Page 326
8. References......Page 327
2. Overall accuracy of a qualitative diagnostic test......Page 329
3. Overall accuracy of a quantitative diagnostic test......Page 331
4. Determining the most accurate threshold for positive quantitative tests......Page 333
5. Discussion......Page 337
7. References......Page 338
1. Introduction......Page 339
2. Type II ANOVA, random effects model......Page 340
3. Type III ANOVA, mixed models......Page 341
4. Repeated measures experiments......Page 343
5. Discussion......Page 345
7. References......Page 346
2. Statistics is fun because it proves your hypothesis was right......Page 347
4. Statistics can provide worthwhile extras to your research......Page 348
5. Statistics is not like algebra bloodless......Page 349
7. Statistics for support rather than illumination?......Page 350
9. Limitations of statistics......Page 351
10. Conclusions......Page 352
11. References......Page 353
2. The randomized controlled clinical trial as the gold standard......Page 354
4. The expanding commend of the pharmaceutical industry over clinical trials......Page 355
5. Flawed procedures jeopardizing current clinical trials......Page 356
7. Further solutions to the dilemma between sponsored research and the independence of science......Page 357
9. References......Page 359
APPENDIX......Page 362
D......Page 370
I......Page 371
P......Page 372
S......Page 373
U......Page 374
Z......Page 375