Statistics Applied to Clinical Trials: Self-Assessment BookSpringer Science & Business Media, 6. pro 2012. - Broj stranica: 226 The authors have taught statistics and given statistics workshops in France and the Netherlands for almost 4 years by now. Their material, mainly on power point, consists of 12 lectures that have been continuously changed and improved by interaction with various audiences. For the purpose of the current book simple English text has been added to the formulas and figures, and the power points sheets have been rewritten in the format given by Kluwer Academic Publishers. Cartoons have been removed, since this is not so relevant for the transmission of thought through a written text, and at the end of each lecture (chapter) a representative number of questions and exercises for self-assessment have been added. At the end of the book detailed answers to the questions and exercises per lecture are given. The book has been produced with the same size and frontpage as the textbook "Statistics Applied To Clinical Trials" by the same authors and edited by same publishers ( 2nd Edition, DordrechtiBostonlLondon, 2002), and can be applied together with the current self-assessment book or separately. The current self-assessment book is different from the texbook, because it focuses on the most important aspects rather than trying to be complete. So, it does not deal with all of the subjects assessed in the texbook. Instead, it repeats on and on the principle things that are needed for every analysis, and it gives many examples that are further explained by arrows in the figures. |
Sadržaj
1 | |
Historical controls 4 Factorial designs 5 Biology is full of variations | 2 |
Summarize the data | 3 |
Two Gaussian curves 8 Human brain | 4 |
Hypothesis | 8 |
Nullhypothesis | 9 |
Alpha the Type I error | 10 |
Ttable | 11 |
Sequential trials | 129 |
Efficacy Z | 130 |
Conclusion | 131 |
MULTIPLE TESTING 1 Two situations | 133 |
51 | 134 |
graphical display | 135 |
Two strategies | 136 |
Alternatives | 137 |
Reject the nullhypothesis | 12 |
Negative trial | 13 |
Borderline result | 14 |
Testing two means | 15 |
Testing paired samples | 16 |
Unpaired testing of paired samples | 17 |
Positive and negative correlations | 18 |
Unpaired analysis of variance ANOVA | 19 |
Paired analysis of variance ANOVA | 20 |
Nonparametric testing for skewed data | 21 |
Paired nonparametric test MannWhitney | 22 |
Unpaired nonparametric test Wilcoxon rank sum | 23 |
Summary | 24 |
Exercises to chapter 1 | 34 |
Alpha beta 1beta | 41 |
CHAPTER 2EQUIVALENCE TESTING V xiii XV 1 1 2 222 2 2 3 4 4 5 6 7 8 9 9 10 12 14 15 15 17 18 19 20 22 22 | 44 |
Use of Ttable to find power one more example | 47 |
Conclusions | 53 |
McNemars test for paired yesno observations | 65 |
PROPORTIONAL DATA ANALYSIS PART II | 72 |
2x2 table | 85 |
Sample size calculation | 87 |
Discussed so far | 88 |
1 sample two measurements | 89 |
A negative study equivalent study why so? 2 Summarize the data 3 Two Gaussian curves 4 Nullhypothesis | 90 |
McNemars test | 91 |
Cochrans test | 92 |
special techniques | 93 |
KaplanMeier curve | 94 |
definition | 95 |
3333 | 96 |
example | 97 |
logrank test | 98 |
software | 99 |
Questions to chapter 5 | 100 |
METAANALYSIS 1 Review of the past | 103 |
What is a metaanalysis | 104 |
Proportions standard errors of proportions odds odds ratios | 105 |
How to calculate 95 confidence intervals of an odds ratio | 106 |
Another summary of metaanalyses helpful to cardiologists for | 107 |
Important matters need few words | 108 |
Negative study 6 Equivalence testing 27 28 29 | 109 |
Strict inclusion criteria | 110 |
second pitfall heterogeneity | 111 |
How to test heterogeneity calculate and pool odds ratios of various studies | 112 |
What to do in case of heterogeneity | 113 |
third pitfall lack of robustness | 115 |
Criticizms of metaanalyses | 116 |
Example of published metaanalysis | 117 |
47 | 118 |
Additional exercises to chapter 6 | 121 |
looking at the data before closure | 123 |
50 | 124 |
Changing inclusion criteria | 125 |
Dangers random high | 126 |
Correction for increasing typeI error rate | 127 |
Some rules | 128 |
results | 138 |
another graphical display | 139 |
Multiple testing | 140 |
Correction | 141 |
graphical display | 142 |
graphical display | 143 |
Conclusion | 144 |
25 | 147 |
Paired data plotted first | 148 |
Correlation coefficient | 149 |
SPSS Statistical software to analyze data from paragraph 1 stool data | 150 |
Three columns of paired observations instead of two | 151 |
Another example of multiple linear regression model | 152 |
Multicollinearity testing of the above example | 153 |
Conclusions | 154 |
SUBGROUP ANALYSIS USING REGRESSION MODELING 1 Subgroup questions 159 | 158 |
Different regression models | 160 |
Linkfunctions | 161 |
Assumptions of the linear regression model | 162 |
Logistic and Cox regression model | 163 |
PROPORTIONAL DATA ANALYSIS PART I | 164 |
An example where the Cox model does not fit | 165 |
example | 166 |
example of a regression model | 167 |
graphical illustration | 168 |
Confounding | 169 |
warning | 170 |
example | 171 |
graphical illustration | 172 |
Questions to chapter 10 | 173 |
RELATIONSHIP AMONG STATISTICAL DISTRIBUTIONS 1 Variables to assess clinical data | 175 |
Creating a chisquare distribution | 176 |
1 x 2 table | 177 |
2 x 2 table | 178 |
Why not x for continuous data | 179 |
What is the advantage of a x²test compared to a z test normal test | 180 |
Examples | 181 |
More examples how to calculate 95 confidence intervals of an odds | 182 |
heterogeneity of trials in a metaanalysis | 184 |
Limitations of statistical tests as discussed and conclusions | 185 |
Questions and exercises to Chapter 11 | 186 |
STATISTICS IS NOT BLOODLESS ALGEBRA 1 Biological processes are full of variations | 189 |
Statistical principles improve quality of trial | 190 |
Statistics is not like algebra and requires biological thinking and just a bit | 191 |
Statistics for support rather than illumination? | 192 |
Limitations of statistics | 193 |
BIAS DUE TO CONFLICTS OF INTERESTS SOME GUIDELINES 1 Introduction 197 | 196 |
Need for circumspection recognized | 198 |
Flawed procedures jeopardizing current clinical trials | 199 |
The good news | 200 |
202 | |
STATISTICAL TABLES | 203 |
ANSWERS TO QUESTIONS AND EXERCISES | 211 |
220 | |
224 | |
Ostala izdanja - Prikaži sve
Uobičajeni izrazi i fraze
ACE-inhibitor alpha alternative is correct ANOVA assessed AUC left AUC right beta Bonferroni calculated censored data chance of finding chapter chi-squared distribution chi-squared test clinical trials compared confidence interval continuous data correlation covariate degrees of freedom discrete drug effect efficacy equivalence testing estimate EXAMPLE graph hazard heterogeneity interim analysis Kaplan-Meier curve linear regression logistic logrank test McNemar's test mean result meta-analyses method mmol/L multiple comparisons multiple testing negative normal distribution null null-hypothesis number of patients observed odds ratio p-value paired placebo pooled Pravastatin prior hypothesis procedure proportional data protocol publication bias quantitative rank numbers regression model rejected residual scientific SEMs distant side-effects significance level significantly different Sleepiness squared standard error statin statistical analysis statistical power strict inclusion criteria T-table t-test test-statistic time-point treatment groups trials similar type-I error rate unpaired variable variance vasodilator x-axis X₁ zero