Introduction to Randomized Controlled Clinical Trials, Second Edition

by Matthews; John N.S.

• ISBN13: 9781584886242
• ISBN10: 1584886242
• Edition: 2nd
• Format: Nonspecific Binding
• Copyright: 2006-06-26
• Publisher: Chapman & Hall/

Summary

Evidence from randomized controlled clinical trials is widely accepted as the only sound basis for assessing the efficacy of new medical treatments. Statistical methods play a key role in all stages of these trials, including their justification, design, and analysis. This second edition of Introduction to Randomized Controlled Clinical Trials provides a concise presentation of the principles applied in this area. It details the concepts behind randomization and methods for designing and analyzing trials and also includes information on meta-analysis and specialized designs, such as cross-over trials, cluster-randomized designs, and equivalence studies.This latest edition features new and revised references, examples, exercises, and a new chapter dedicated to binary outcomes and survival analysis. It also presents numerous examples taken from the medical literature, contains exercises at the end of each chapter, and offers solutions in an appendix. The author uses Minitab and R software throughout the text for implementing the methods that are presented.Comprehensive and accessible, Introduction to Randomized Controlled Clinical Trials is well-suited for those familiar with elementary statistical ideas and methods who want to further their knowledge of the subject.

Table of Contents

1 What Is a Randomized Controlled Trial?	1	(14)
1.1 Definition and Key Features	1	(1)
1.2 Historical Context and the Nature of RCTs	2	(3)
1.2.1 Historical Background	2	(1)
1.2.2 Impact of RCTs and Importance of Statistics	3	(1)
1.2.3 Ethical Issues	4	(1)
1.3 Structure and Justification of RCTs	5	(7)
1.3.1 Eligible Patients	5	(3)
1.3.2 The Need for Concurrent Controls	8	(2)
1.3.3 Allocation at Random	10	(6)
1.3.3.1 Comparable Groups	10	(1)
1.3.3.2 Allocation Unknown at Entry to the RCT	11	(1)
1.3.3.3 Valid Estimate of Error	11	(1)
Exercises	12	(3)
2 Bias	15	(10)
2.1 What Is Meant by Bias in RCTs?	15	(1)
2.2 Types of Bias	16	(6)
2.2.1 Selection Bias	17	(3)
2.2.2 Allocation Bias	20	(1)
2.2.3 Assessment Bias	21	(1)
2.2.4 Publication Bias	21	(1)
2.2.5 Stopping Rules	22	(1)
Exercises	22	(3)
3 How Many Patients Do I Need?	25	(18)
3.1 Criteria for Sample Size Calculations	25	(2)
3.1.1 The Primary Question	26	(1)
3.1.2 Adequate Evidence	27	(1)
3.2 Hypothesis Tests	27	(3)
3.2.1 General Remarks	27	(1)
3.2.2 Implication for RCTs	28	(1)
3.2.3 One-Sided vs. Two-Sided Tests	29	(1)
3.3 Sample Size for a Normally Distributed Variable	30	(4)
3.4 Sample Size for a Binary Variable	34	(4)
3.4.1 Formulation	34	(1)
3.4.2 The Delta Method	35	(1)
3.4.3 Sample Size Formula	36	(2)
3.5 General Remarks about Sample Size Calculations	38	(1)
Exercises	39	(4)
4 Methods of Allocation	43	(22)
4.1 Simple Randomization	43	(3)
4.1.1 Chance Imbalance of Group Sizes	43	(2)
4.1.2 Distribution of	45	(1)
4.1.3 Practical Consequences	45	(1)
4.2 Random Permuted Blocks	46	(3)
4.2.1 Random Permuted Blocks of Fixed Length	46	(1)
4.2.2 RPBs with Random Block Length	47	(2)
4.3 Biased Coin Designs and Urn Schemes	49	(4)
4.3.1 Biased Coin Designs	49	(1)
4.3.2 Urn Models	50	(3)
4.4 Unequal Randomization	53	(1)
4.5 Stratification	54	(4)
4.5.1 The Problem of Allocation Bias	54	(3)
4.5.2 Using Stratification	57	(1)
4.6 Minimization	58	(3)
4.6.1 Minimization for Two-Treatment Trials	58	(3)
Exercises	61	(4)
5 Assessment, Blinding, and Placebos	65	(8)
5.1 Double and Single Blindness	65	(2)
5.1.1 General Principles	65	(2)
5.2 Placebos	67	(4)
5.2.1 The Placebo Effect	67	(2)
5 2.2 The Role of Placebos Blinding	69	(2)
5.3 Practical Considerations
Exercises	71	(2)
6 Analysis Dif Results	73	(22)
6.1 Example	73
6.2 Use of Confidence Intervals	71	(6)
6.3 Baselines: Uses and Abuses
6.3.1 Comparison of Baseline Values	77	(1)
6.3.2. Analyzing Change from Baseline	77	(2)
6.3.3 An Erroneous Analysis Based on Change from Baseline	79	(1)
6.4 Analysis of Covariance	80	(11)
6.4.1 JBaseline Bias Leads to Outcome Bias	80	(1)
6.4.2 Interlude: Bivariate Normal Variables	81	(2)
6.4.3 Allowing for Baseline Imbalance: The Theory	83	(2)
6.4.4 Allowing for Baseline Imbalance: The Practice	85	(3)
6.4.5 Conditional vs. Unconditional Inference	88	(1)
6.4.6 Some General Remarks on Covariates	89	(7)
6.4.6.1 Choice of Analysis	89	(1)
6.4.6.2 Other Types of Covariate	89	(1)
6.4.6.3 Which Covariates Should Be Used?	89	(2)
Exercises	91	(4)
7 Further Analysis: Binary and Survival Data	95	(54)
7.1 Binary Data: An Example and a Statistical Model	95	(1)
7.2 Point Estimates and Hypothesis Tests	96	(5)
7.2.1 Methods Based on Informal Derivations	96	(3)
7.2.2 Methods Based on More Formal Derivations	99	(2)
7.3 Interval Estimates for the Binary Case	101	(9)
7.3.1 Different Measures of Difference between Treatments	101	(1)
7.3.2 Interval Estimates or Confidence Intervals	102	(8)
7.3.2.1 Interval Estimates for ARD and NNT	102	(2)
7.3.2.2 Problems with the Confidence Interval for ARD and a Simple Solution	104	(3)
7.3.2.3 Interval Estimates for RR and OR	107	(2)
7.3.2.4 Summary of Results	109	(1)
7.4 Adjusting Binary Outcomes for Baseline Observations	110	(13)
7.4.1 Using a Logistic Model	110	(2)
7.4.2 Fitting a Logistic Regression	112	(2)
7.4.3 Adjusting the OR in the PUVA vs. TL-01 Trial	114	(3)
7.4.4 An Alternative Approach	117	(6)
7.4.4.1 The Hypergeometric Distribution	117	(2)
7.4.4.2 The Mantel—Haenszel Method	119	(4)
7.5 Survival Analysis	123	(15)
7.5.1 The Hazard Function, the Survival Curve, and Some Estimators	124	(6)
7.5.1.1 The Kaplan—Meier Estimator	126	(1)
7.5.1.2 Parametric Estimators	126	(4)
7.5.2 Comparing Survival Curves	130	(3)
7.5.3 Adjusting Survival Analyses for Baseline Values	133	(5)
7.6 Analyses Using Randomization Models	138	(9)
7.6.1 Randomization Models: Simple Randomization	138	(4)
7.6.2 Randomization Models: Stratified Randomization	142	(9)
7.6.2.1 Stratified Randomization: Exact Analysis	146	(1)
7.6.2.2 Blocked Allocations	147	(1)
Exercises	147	(2)
8 Monitoring Accumulating Data	149	(16)
8.1 Motivation and Problems with Repeated Analysis of Data	149	(2)
8.2 Sequential and Group Sequential Methods	151	(7)
8.2.1 Using Repeated Significance Tests	151	(2)
8.2.2 Some Theory for Group Sequential Designs with Normal Outcomes	153	(2)
8.2.3 Other Forms of Stopping Rule	155	(3)
8.3 Other Approaches to Accumulating Data	158	(4)
8.3.1 SPRT and Triangular Test	158	(2)
8.3.2 Bayesian and Likelihood Approaches	160	(1)
8.3.3 Adaptive Allocation	161	(1)
8.4 Data Monitoring Committees	162	(2)
Exercises	164	(1)
9 Subgroups and Multiple Outcomes	165	(18)
9.1 The Role of Subgroups in Randomized Clinical Trials	165	(1)
9.2 Methods for Comparing Subgroups	166	(4)
9.2.1 Approximate Theory	166	(1)
9.2.2 Practice: Continuous Outcomes	167	(2)
9.2.2.1 An Erroneous Analysis	168	(1)
9.2.3 Practice: Binary Outcomes	169	(1)
9.3 Methods of Selecting Subgroups	170	(2)
9.4 Qualitative Interactions	172	(2)
9.5 Multiple Outcomes	174	(2)
9.6 Correction of P-Values	176	(1)
9.6.1 Boole's Inequality	176	(1)
9.6.2 Application to RCTs	176	(1)
9.7 Some Alternative Methods for Multiple Outcomes	177	(3)
9.7.1 Extension to Bonferroni Inequality	178	(1)
9.7.2 Multivariate Methods	178	(2)
9.7.3 O'Brien's Rank-Based Method	180	(1)
Exercises	180	(3)
10 Protocols and Protocol Deviations	183	(10)
10.1 Protocols: Their Nature and Role	183	(1)
10.2 Protocol Deviation	184	(4)
10.2.1 Ineligible Patients	185	(1)
10.2.2 Administration of Treatment Other than That Allocated	186	(2)
10.3 Analysis by Intention-to-Treat	188	(3)
10.3.1 Informal Description	188	(1)
10.3.2 Theoretical Description	189	(2)
Exercises	191	(2)
11 Some Special Designs: Crossovers, Equivalence, and Clusters	193	(18)
11.1 Crossover Trials	193	(1)
11.2 The AB/BA Design	194	(1)
11.3 Analysis of AB/BA Design for Continuous Outcomes	194	(3)
11.3.1 The Theory	194	(2)
11.3.2 An Application	196	(1)
11.4 The Issue of Carryover	197	(2)
11.5 Equivalence Trials	199	(3)
11.5.1 General Remarks	199	(1)
11.5.2 Sample Sizes for Equivalence Trials with Normally Distributed Outcomes	200	(2)
11.5.3 Comparison of Conventional and Equivalence Trials	202	(1)
11.6 Cluster Randomized Trials	202	(6)
11.6.1 Introduction and Rationale	202	(3)
11.6.2 Methods of Analysis for Cluster-Randomized Trials	205	(1)
11.6.3 Sample Size Estimation for Continuous Outcomes from Cluster-Randomized Trials	206	(1)
11.6.4 General Remarks about Cluster-Randomized Trials	207	(1)
Exercises	208	(3)
12 Meta-Analyses of Clinical Trials	211	(12)
12.1 What Are Meta-Analyses, and Why Are They Needed?	211	(2)
12.2 Some Methodology for Meta-Analysis	213	(4)
12.2.1 Data Collection	213	(1)
12.2.2 Estimating the Overall Effect Using a Fixed-Effects Model	213	(2)
12.2.3 Estimating the Overall Effect Using a Random-Effects Model	215	(2)
12.3 Some Graphical Methods for Meta-Analysis	217	(4)
12.3.1 Meta-Analysis Diagrams	217	(1)
12.3.2 Sensitivity Plots for Random-Effects Analysis	218	(1)
12.3.3 Funnel or Radial Plots	219	(2)
12.4 Some General Issues in Meta-Analysis	221	(1)
12.4.1 Random-Effects and Intertrial Variation	221	(1)
12.4.2 Position of Meta-Analysis	221	(1)
Exercises	222	(1)
Further Reading	223	(4)
Books on Clinical Trials in General	223	(1)
Books on More Specialized Topics	224	(3)
Solutions to Exercises	227	(42)
References	269	(4)
Index	273