Design and Analysis of Experiments in the Health Sciences

GERALD VAN BELLE, PhD, is Professor Emeritus in the Departments of Biostatistics and Environmental and Occupational Health Sciences at the University of Washington. A Fellow of the American Statistical Association and the American Association for the Advancement of Science, he has published more than 140 articles in the areas of experimental design and data characterization as well as analysis with application to neurodegenerative diseases, effects of air pollution on health and toxicology, and clinical trials in resuscitation outcomes research.KATHLEEN F. KERR, PhD, is Associate Professor of Biostatistics at the University of Washington. A former Burroughs Wellcome postdoctoral fellow in mathematics and molecular biology, Dr. Kerr currently serves as associate editor of PLoS Genetics and Statistical Applications in Genetics and Molecular Biology. Her research interests include gene expression microarrays, statistical genetics, experimental design, and biomarker research.

“Overall, Design and Analysis of Experiments in the Health Sciencesis a balanced and approachable text suitable for a graduate level experimental design course, and will prove particularly useful to practitioners in the health sciences.”  (Journal of Biopharmaceutical Statistics, 1 January 2013)“The book will be a valuable resource for researchers in medicine, dentistry, and the public health sciences.  The authors are faculty members in the Department of Biostatistics at the University of Washington in Seattle.”  (Journal of Clinical Research Best Practices, 1 September 2012)

• Preface xiii1 The Basics 11.1 Four Basic Questions 11.2 Variation 41.3 Principles of Design and Analysis 51.4 Experiments and Observational Studies 91.5 Illustrative Applications of Principles 111.6 Experiments in the Health Sciences 121.7 Adaptive Allocation 151.7.1 Equidistribution 151.7.2 Adaptive Allocation Techniques 161.8 Sample Size Calculations 181.9 Statistical Models for the Data 201.10 Analysis and Presentation 221.10.1 Graph the Data in Several Ways 221.10.2 Assess Assumptions of the Statistical Model 221.10.3 Confirmatory and Exploratory Analysis 231.10.4 Missing Data Need Careful Accounting 231.10.5 Statistical Software 241.11 Notes 241.11.1 Characterization Studies 241.11.2 Additional Comments on Balance 251.11.3 Linear and Nonlinear Models 251.11.4 Analysis of Variance versus Regression Analysis 261.12 Summary 261.13 Problems 262 Completely Randomized Designs 312.1 Randomization 312.2 Hypotheses and Sample Size 322.3 Estimation and Analysis 322.4 Example 342.5 Discussion and Extensions 362.5.1 Preparing Data for Computer Analysis 362.5.2 Treatment Assignment in this Example 372.5.3 Check on Randomization 372.5.4 Partitioning the Treatment Sum of Squares 372.5.5 Alternative Endpoints 382.5.6 Dummy Variables 382.5.7 Contrasts 392.6 Randomization 412.7 Hypotheses and Sample Size 412.8 Estimation and Analysis 412.9 Example 422.10 Discussion and Extensions 442.10.1 Two Roles for ANCOVA 442.10.2 Partitioning of Sums of Squares 452.10.3 Assumption of Parallelism 462.11 Notes 472.11.1 Constrained Randomization 472.11.2 Assumptions of the Analysis of Variance and Covariance 482.11.3 When the Assumptions Don’t Hold 492.11.4 Alternative Graphical Displays 502.11.5 Sample Sizes for More Than Two Levels 512.11.6 Limitations of Computer Output 512.11.7 Unequal Sample Sizes 512.11.8 Design Implications of the CRD 512.11.9 Power and Alternative Hypotheses 522.11.10 Regression or Analysis of Variance? 522.11.11 Bioassay 522.12 Summary 532.13 Problems 533 Randomized Block Designs 633.1 Randomization 643.2 Hypotheses and Sample Size 643.3 Estimation and Analysis 643.4 Example 653.5 Discussion and Extensions 673.5.1 Evaluating Model Assumptions 673.5.2 Multiple Comparisons 693.5.3 Number of Treatments and Block Size 713.5.4 Missing Data 713.5.5 Does It Always Pay to Block? 713.5.6 Concomitant Variables 723.5.7 Imbalance 743.6 Randomization 773.7 Hypotheses and Sample Size 773.8 Estimation and Analysis 773.9 Example 773.10 Discussion and Extensions 793.10.1 Implications of the Model 793.10.2 Number of Latin Squares 793.11 Randomization 803.12 Hypotheses and Sample Size 813.13 Estimation and Analysis 823.14 Example 823.15 Discussion and Extensions 853.15.1 Partially Balanced Incomplete Block Designs 853.16 Notes 863.16.1 Analysis Follows Design 863.16.2 Relative Efficiency 863.16.3 Additivity of the Model 873.17 Summary 883.18 Problems 884 Factorial Designs 934.1 Randomization 954.2 Hypotheses and Sample Size 954.3 Estimation and Analysis 964.4 Example 1 974.5 Example 2 1004.6 Notes 1034.6.1 Regression Analysis Approaches 1034.6.2 Almost Factorial 1054.6.3 Design Structure and Factor Structure 1054.6.4 Effect and Interaction Tables 1054.6.5 Balanced Design 1054.6.6 Missing Data 1064.6.7 Fixed, Random, and Mixed Effects Models 1064.6.8 Fractional Factorials 1084.7 Summary 1094.8 Problems 1105 Multilevel Designs 1175.1 Randomization 1185.2 Hypotheses and Sample Size 1185.3 Estimation and Analysis 1195.4 Example 1215.5 Discussion and Extensions 1275.5.1 Whole-Plot and Split-Plot Variability 1275.5.2 Getting the Computer to Do the Right Analysis 1285.6 Notes 1295.6.1 Fractional Factorials—Example 1295.6.2 Missing Data 1295.7 Summary 1305.8 Problems 1306 Repeated Measures Designs 1356.1 Randomization 1366.2 Hypotheses and Sample Size 1366.3 Estimation and Analysis 1376.4 Example 1396.5 Discussion and Extensions 1426.6 Notes 1436.6.1 RBD and RMD 1436.6.2 Missing Data: The Fundamental Challenge in RMD 1436.6.3 Correlation Structure 1446.6.4 Derived Variable Analysis 1446.7 Summary 1446.8 Problems 1457 Randomized Clinical Trials 1497.1 Endpoints 1517.2 Randomization 1527.3 Hypotheses and Sample Size 1537.4 Follow-Up 1547.5 Estimation and Analysis 1547.6 Examples 1557.7 Discussion and Extensions 1597.7.1 Statistical Significance and Clinical Importance 1597.7.2 Ethics 1617.7.3 Reporting 1627.8 Notes 1637.8.1 Multicenter Trials 1637.8.2 International Harmonization 1677.8.3 Data Safety Monitoring 1677.8.4 Ancillary Studies 1687.8.5 Subgroup Analysis and Data Mining 1687.8.6 Meta-Analysis 1697.8.7 Authorship and Recognition 1697.8.8 Communication 1697.8.9 Data Sharing 1707.8.10 N-of-1 Trials 1707.9 Resources 1717.10 Summary 1717.11 Problems 1718 Microarrays 1798.1 Introduction 1798.2 Genes, Gene Expression, and Microarrays 1798.2.1 Genes and Gene Expression 1798.2.2 Gene Expression Microarrays 1808.3 Examples of Microarray Studies 1868.4 Replication and Sample Size 1888.5 Blocking and Microarrays 1898.6 Randomization and Microarrays 1908.7 Microarray Data Analysis Issues 1918.7.1 Image Analysis 1918.7.2 Data Preprocessing 1938.7.3 Identifying Differentially Expressed Genes 1968.7.4 Multiple Testing 1968.7.5 Gene Set Analysis 1988.7.6 The Class Prediction Problem 1988.8 Data Analysis Example 2008.9 Notes 2028.9.1 Sample Size 2028.9.2 FDR Estimation 2028.9.3 Evaluation of Data Preprocessing Methods 2038.10 Summary 2038.11 Problems 203Bibliography 207Author Index 217Subject Index 223