Handbook of Statistics Vol 27 - Epidemiology and Medical Statistics

Table of contents
Preface xiii
Contributors xv
Ch. 1. Statistical Methods and Challenges in Epidemiology and
Biomedical Research 1
Ross L. Prentice
1. Introduction 1
2. Characterizing the study cohort 3
3. Observational study methods and challenges 6
4. Randomized controlled trials 12
5. Intermediate, surrogate, and auxiliary outcomes 17
6. Multiple testing issues and high-dimensional biomarkers 18
7. Further discussion and the Women’s Health Initiative example 21
References 22
Ch. 2. Statistical Inference for Causal Effects, With Emphasis on
Applications in Epidemiology and Medical Statistics 28
Donald B. Rubin
1. Causal inference primitives 28
2. The assignment mechanism 36
3. Assignment-based modes of causal inference 41
4. Posterior predictive causal inference 47
5. Complications 55
References 58
Ch. 3. Epidemiologic Study Designs 64
Kenneth J. Rothman, Sander Greenland and Timothy L. Lash
1. Introduction 64
2. Experimental studies 65
3. Nonexperimental studies 73
4. Cohort studies 73
v
5. Case-control studies 84
6. Variants of the case-control design 97
7. Conclusion 104
References 104
Ch. 4. Statistical Methods for Assessing Biomarkers and Analyzing
Biomarker Data 109
Stephen W. Looney and Joseph L. Hagan
1. Introduction 109
2. Statistical methods for assessing biomarkers 110
3. Statistical methods for analyzing biomarker data 126
4. Concluding remarks 143
References 144
Ch. 5. Linear and Non-Linear Regression Methods in Epidemiology and
Biostatistics 148
Eric Vittinghoff, Charles E. McCulloch, David V. Glidden and Stephen
C. Shiboski
1. Introduction 148
2. Linear models 151
3. Non-linear models 167
4. Special topics 176
References 182
Ch. 6. Logistic Regression 187
Edward L. Spitznagel Jr.
1. Introduction 187
2. Estimation of a simple logistic regression model 188
3. Two measures of model fit 191
4. Multiple logistic regression 192
5. Testing for interaction 194
6. Testing goodness of fit: Two measures for lack of fit 195
7. Exact logistic regression 196
8. Ordinal logistic regression 201
9. Multinomial logistic regression 204
10. Probit regression 206
11. Logistic regression in case–control studies 207
References 209
Ch. 7. Count Response Regression Models 210
Joseph M. Hilbe and William H. Greene
1. Introduction 210
2. The Poisson regression model 212
3. Heterogeneity and overdispersion 224
4. Important extensions of the models for counts 230
Table of contents vi
5. Software 247
6. Summary and conclusions 250
References 251
Ch. 8. Mixed Models 253
Matthew J. Gurka and Lloyd J. Edwards
1. Introduction 253
2. Estimation for the linear mixed model 259
3. Inference for the mixed model 261
4. Selecting the best mixed model 264
5. Diagnostics for the mixed model 268
6. Outliers 270
7. Missing data 270
8. Power and sample size 272
9. Generalized linear mixed models 273
10. Nonlinear mixed models 274
11. Mixed models for survival data 275
12. Software 276
13. Conclusions 276
References 277
Ch. 9. Survival Analysis 281
John P. Klein and Mei-Jie Zhang
1. Introduction 281
2. Univariate analysis 282
3. Hypothesis testing 288
4. Regression models 295
5. Regression models for competing risks 310
References 317
Ch. 10. A Review of Statistical Analyses for Competing Risks 321
Melvin L. Moeschberger, Kevin P. Tordoff and Nidhi Kochar
1. Introduction 321
2. Approaches to the statistical analysis of competing risks 324
3. Example 327
4. Conclusion 339
References 340
Ch. 11. Cluster Analysis 342
William D. Shannon
1. Introduction 342
2. Proximity measures 344
3. Hierarchical clustering 350
4. Partitioning 355
5. Ordination (scaling) 358
Table of contents vii
6. How many clusters? 361
7. Applications in medicine 364
8. Conclusion 364
References 365
Ch. 12. Factor Analysis and Related Methods 367
Carol M. Woods and Michael C. Edwards
1. Introduction 367
2. Exploratory factor analysis (EFA) 368
3. Principle components analysis (PCA) 375
4. Confirmatory factor analysis (CFA) 375
5. FA with non-normal continuous variables 379
6. FA with categorical variables 380
7. Sample size in FA 382
8. Examples of EFA and CFA 383
9. Additional resources 389
Appendix A 391
Appendix B 391
References 391
Ch. 13. Structural Equation Modeling 395
Kentaro Hayashi, Peter M. Bentler and Ke-Hai Yuan
1. Models and identification 395
2. Estimation and evaluation 399
3. Extensions of SEM 410
4. Some practical issues 415
References 418
Ch. 14. Statistical Modeling in Biomedical Research: Longitudinal
Data Analysis 429
Chengjie Xiong, Kejun Zhu, Kai Yu and J. Philip Miller
1. Introduction 429
2. Analysis of longitudinal data 431
3. Design issues of a longitudinal study 456
References 460
Ch. 15. Design and Analysis of Cross-Over Trials 464
Michael G. Kenward and Byron Jones
1. Introduction 464
2. The two-period two-treatment cross-over trial 467
3. Higher-order designs 476
4. Analysis with non-normal data 482
5. Other application areas 485
6. Computer software 488
References 489
Table of contents viii
Ch. 16. Sequential and Group Sequential Designs in Clinical Trials:
Guidelines for Practitioners 491
Madhu Mazumdar and Heejung Bang
1. Introduction 492
2. Historical background of sequential procedures 493
3. Group sequential procedures for randomized trials 494
4. Steps for GSD design and analysis 507
5. Discussion 508
References 509
Ch. 17. Early Phase Clinical Trials: Phases I and II 513
Feng Gao, Kathryn Trinkaus and J. Philip Miller
1. Introduction 513
2. Phase I designs 514
3. Phase II designs 526
4. Summary 539
References 541
Ch. 18. Definitive Phase III and Phase IV Clinical Trials 546
Barry R. Davis and Sarah Baraniuk
1. Introduction 546
2. Questions 548
3. Randomization 550
4. Recruitment 551
5. Adherence/sample size/power 552
6. Data analysis 554
7. Data quality and control/data management 558
8. Data monitoring 558
9. Phase IV trials 563
10. Dissemination – trial reporting and beyond 564
11. Conclusions 565
References 565
Ch. 19. Incomplete Data in Epidemiology and Medical Statistics 569
Susanne Ra¨ssler, Donald B. Rubin and Elizabeth R. Zell
1. Introduction 569
2. Missing-data mechanisms and ignorability 571
3. Simple approaches to handling missing data 573
4. Single imputation 574
5. Multiple imputation 578
6. Direct analysis using model-based procedures 581
7. Examples 584
8. Literature review for epidemiology and medical studies 586
9. Summary and discussion 587
Appendix A 588
Appendix B 592
References 598
Table of contents ix
Ch. 20. Meta-Analysis 602
Edward L. Spitznagel Jr.
1. Introduction 602
2. History 603
3. The Cochran–Mantel–Haenszel test 604
4. Glass’s proposal for meta-analysis 606
5. Random effects models 607
6. The forest plot 609
7. Publication bias 610
8. The Cochrane Collaboration 614
References 614
Ch. 21. The Multiple Comparison Issue in Health Care Research 616
Lemuel A. Moye′
1. Introduction 616
2. Concerns for significance testing 617
3. Appropriate use of significance testing 618
4. Definition of multiple comparisons 619
5. Rational for multiple comparisons 620
6. Multiple comparisons and analysis triage 621
7. Significance testing and multiple comparisons 623
8. Familywise error rate 625
9. The Bonferroni inequality 626
10. Alternative approaches 629
11. Dependent testing 631
12. Multiple comparisons and combined endpoints 635
13. Multiple comparisons and subgroup analyses 641
14. Data dredging 651
References 651
Ch. 22. Power: Establishing the Optimum Sample Size 656
Richard A. Zeller and Yan Yan
1. Introduction 656
2. Illustrating power 658
3. Comparing simulation and software approaches to power 663
4. Using power to decrease sample size 672
5. Discussion 677
References 677
Ch. 23. Statistical Learning in Medical Data Analysis 679
Grace Wahba
1. Introduction 679
2. Risk factor estimation: penalized likelihood estimates 681
3. Risk factor estimation: likelihood basis pursuit and the LASSO 690
Table of contents x
4. Classification: support vector machines and related estimates 693
5. Dissimilarity data and kernel estimates 700
6. Tuning methods 704
7. Regularization, empirical Bayes, Gaussian processes priors, and
reproducing kernels 707
References 708
Ch. 24. Evidence Based Medicine and Medical Decision Making 712
Dan Mayer
1. The definition and history of evidence based medicine 712
2. Sources and levels of evidence 715
3. The five stage process of EBM 717
4. The hierarchy of evidence: study design and minimizing bias 718
5. Assessing the significance or impact of study results: Statistical significance and
confidence intervals 721
6. Meta-analysis and systematic reviews 722
7. The value of clinical information and assessing the usefulness
of a diagnostic test 722
8. Expected values decision making and the threshold approach
to diagnostic testing 726
9. Summary 727
10. Basic principles 727
References 728
Ch. 25. Estimation of Marginal Regression Models with Multiple
Source Predictors 730
Heather J. Litman, Nicholas J. Horton, Bernardo Herna′ndez and
Nan M. Laird
1. Introduction 730
2. Review of the generalized estimating equations approach 732
3. Maximum likelihood estimation 735
4. Simulations 737
5. Efficiency calculations 740
6. Illustration 741
7. Conclusion 743
References 745
Ch. 26. Difference Equations with Public Health Applications 747
Asha Seth Kapadia and Lemuel A. Moye′
1. Introduction 747
2. Generating functions 748
3. Second-order nonhomogeneous equations and generating functions 750
4. Example in rhythm disturbances 752
5. Follow-up losses in clinical trials 758
6. Applications in epidemiology 765
References 773
Table of contents xi
Ch. 27. The Bayesian Approach to Experimental Data Analysis 775
Bruno Lecoutre
Preamble: and if you were a Bayesian without knowing it? 775
1. Introduction 776
2. Frequentist and Bayesian inference 778
3. An illustrative example 783
4. Other examples of inferences about proportions 795
5. Concluding remarks and some further topics 803
References 808
Subject Index 813
Handbook of Statistics Contents of Previous Volumes 823
Table of contents xii