Available:*
Library | Item Barcode | Call Number | Material Type | Item Category 1 | Status |
---|---|---|---|---|---|
Searching... | 30000004589978 | QA278 M64 2005 | Open Access Book | Book | Searching... |
On Order
Summary
Summary
The linear mixed model has become the main parametric tool for the analysis of continuous longitudinal data, as the authors discussed in their 2000 book.
Without putting too much emphasis on software, the book shows how the different approaches can be implemented within the SAS software package.
The authors received the American Statistical Association's Excellence in Continuing Education Award based on short courses on longitudinal and incomplete data at the Joint Statistical Meetings of 2002 and 2004.
Author Notes
Geert Molenberghs is Professor of Biostatistics at the Universiteit Hasselt in Belgium and has published methodological work on surrogate markers in clinical trials, categorical data, longitudinal data analysis, and the analysis of non-response in clinical and epidemiological studies
Geert Verbeke is Professor of Biostatistics at the Biostatistical Centre of the Katholieke Universiteit Leuven in Belgium
Table of Contents
Preface | p. vii |
Acknowledgments | p. ix |
I Introductory Material | p. 1 |
1 Introduction | p. 3 |
2 Motivating Studies | p. 7 |
2.1 Introduction | p. 7 |
2.2 The Analgesic Trial | p. 8 |
2.3 The Toenail Data | p. 8 |
2.4 The Fluvoxamine Trial | p. 12 |
2.5 The Epilepsy Data | p. 14 |
2.6 The Project on Preterm and Small for Gestational Age Infants (POPS) Study | p. 14 |
2.7 National Toxicology Program Data | p. 17 |
2.8 The Sports Injuries Trial | p. 23 |
2.9 Age Related Macular Degeneration Trial | p. 24 |
3 Generalized Linear Models | p. 27 |
3.1 Introduction | p. 27 |
3.2 The Exponential Family | p. 27 |
3.3 The Generalized Linear Model (GLM) | p. 28 |
3.4 Examples | p. 29 |
3.5 Maximum Likelihood Estimation and Inference | p. 30 |
3.6 Logistic Regression for the Toenail Data | p. 31 |
3.7 Poisson Regression for the Epilepsy Data | p. 32 |
4 Linear Mixed Models for Gaussian Longitudinal Data | p. 35 |
4.1 Introduction | p. 35 |
4.2 Marginal Multivariate Model | p. 36 |
4.3 The Linear Mixed Model | p. 36 |
4.4 Estimation and Inference for the Marginal Model | p. 39 |
4.5 Inference for the Random Effects | p. 41 |
5 Model Families | p. 45 |
5.1 Introduction | p. 45 |
5.2 The Gaussian Case | p. 46 |
5.3 Model Families in General | p. 47 |
5.4 Inferential Paradigms | p. 52 |
II Marginal Models | p. 53 |
6 The Strength of Marginal Models | p. 55 |
6.1 Introduction | p. 55 |
6.2 Marginal Models in Contingency Tables | p. 56 |
6.3 British Occupational Status Study | p. 62 |
6.4 The Caithness Data | p. 62 |
6.5 Analysis of the Fluvoxamine Trial | p. 64 |
6.6 Extensions | p. 68 |
6.7 Relation to Latent Continuous Densities | p. 79 |
6.8 Conclusions and Perspective | p. 80 |
7 Likelihood-based Marginal Models | p. 83 |
7.1 Notation | p. 84 |
7.2 The Bahadur Model | p. 86 |
7.3 A General Framework for Fully Specified Marginal Models | p. 93 |
7.4 Maximum Likelihood Estimation | p. 99 |
7.5 An Influenza Study | p. 99 |
7.6 The Multivariate Probit Model | p. 102 |
7.7 The Dale Model | p. 113 |
7.8 Hybrid Marginal-conditional Specification | p. 122 |
7.9 A Cross-over Trial: An Example in Primary Dysmenorrhoea | p. 127 |
7.10 Multivariate Analysis of the POPS Data | p. 131 |
7.11 Longitudinal Analysis of the Fluvoxamine Study | p. 134 |
7.12 Appendix: Maximum Likelihood Estimation | p. 136 |
7.13 Appendix: The Multivariate Plackett Distribution | p. 142 |
7.14 Appendix: Maximum Likelihood Estimation for the Dale Model | p. 147 |
8 Generalized Estimating Equations | p. 151 |
8.1 Introduction | p. 151 |
8.2 Standard GEE Theory | p. 153 |
8.3 Alternative GEE Methods | p. 161 |
8.4 Prentice's GEE Method | p. 162 |
8.5 Second-order Generalized Estimating Equations (GEE2) | p. 164 |
8.6 GEE with Odds Ratios and Alternating Logistic Regression | p. 165 |
8.7 GEE2 Based on a Hybrid Marginal-conditional Model | p. 168 |
8.8 A Method Based on Linearization | p. 169 |
8.9 Analysis of the NTP Data | p. 170 |
8.10 The Heatshock Study | p. 174 |
8.11 The Sports Injuries Trial | p. 181 |
9 Pseudo-Likelihood | p. 189 |
9.1 Introduction | p. 189 |
9.2 Pseudo-Likelihood: Definition and Asymptotic Properties | p. 190 |
9.3 Pseudo-Likelihood Inference | p. 192 |
9.4 Marginal Pseudo-Likelihood | p. 195 |
9.5 Comparison with Generalized Estimating Equations | p. 199 |
9.6 Analysis of NTP Data | p. 200 |
10 Fitting Marginal Models with SAS | p. 203 |
10.1 Introduction | p. 203 |
10.2 The Toenail Data | p. 203 |
10.3 GEE1 with Correlations | p. 204 |
10.4 Alternating Logistic Regressions | p. 212 |
10.5 A Method Based on Linearization | p. 215 |
10.6 Programs for the NTP Data | p. 219 |
10.7 Alternative Software Tools | p. 221 |
III Conditional Models | p. 223 |
11 Conditional Models | p. 225 |
11.1 Introduction | p. 225 |
11.2 Conditional Models | p. 226 |
11.3 Marginal versus Conditional Models | p. 233 |
11.4 Analysis of the NTP Data | p. 234 |
11.5 Transition Models | p. 236 |
12 Pseudo-Likehood | p. 243 |
12.1 Introduction | p. 243 |
12.2 Pseudo-Likelihood for a Single Repeated Binary Outcome | p. 244 |
12.3 Pseudo-Likelihood for a Multivariate Repeated Binary Outcome | p. 245 |
12.4 Analysis of the NTP Data | p. 246 |
IV Subject-specific Models | p. 255 |
13 From Subject-specific to Random-effects Models | p. 257 |
13.1 Introduction | p. 257 |
13.2 General Model Formulation | p. 257 |
13.3 Three Ways to Handle Subject-specific Parameters | p. 258 |
13.4 Random-effects Models: Special Cases | p. 260 |
14 The Generalized Linear Mixed Model (GLMM) | p. 265 |
14.1 Introduction | p. 265 |
14.2 Model Formulation and Approaches to Estimation | p. 265 |
14.3 Estimation: Approximation of the Integrand | p. 268 |
14.4 Estimation: Approximation of the Data | p. 269 |
14.5 Estimation: Approximation of the Integral | p. 273 |
14.6 Inference in Generalized Linear Mixed Models | p. 276 |
14.7 Analyzing the NTP Data | p. 277 |
14.8 Analyzing the Toenail Data | p. 278 |
15 Fitting Generalized Linear Mixed Models with SAS | p. 281 |
15.1 Introduction | p. 281 |
15.2 The GLIMMIX Procedure for Quasi-Likelihood | p. 282 |
15.3 The GLIMMIX Macro for Quasi-Likelihood | p. 287 |
15.4 The NLMIXED Procedure for Numerical Quadrature | p. 290 |
15.5 Alternative Software Tools | p. 296 |
16 Marginal versus Random-effects Models | p. 297 |
16.1 Introduction | p. 297 |
16.2 Example: The Toenail Data | p. 297 |
16.3 Parameter Interpretation | p. 298 |
16.4 Toenail Data: Marginal versus Mixed Models | p. 301 |
16.5 Analysis of the NTP Data | p. 304 |
V Case Studies and Extensions | p. 307 |
17 The Analgesic Trial | p. 309 |
17.1 Introduction | p. 309 |
17.2 Marginal Analyses of the Analgesic Trial | p. 310 |
17.3 Random-effects Analyses of the Analgesic Trial | p. 314 |
17.4 Comparing Marginal and Random-effects Analyses | p. 317 |
17.5 Programs for the Analgesic Trial | p. 318 |
18 Ordinal Data | p. 325 |
18.1 Regression Models for Ordinal Data | p. 326 |
18.2 Marginal Models for Repeated Ordinal Data | p. 329 |
18.3 Random-effects Models for Repeated Ordinal Data | p. 331 |
18.4 Ordinal Analysis of the Analgesic Trial | p. 332 |
18.5 Programs for the Analgesic Trial | p. 334 |
19 The Epilepsy Data | p. 337 |
19.1 Introduction | p. 337 |
19.2 A Marginal GEE Analysis | p. 337 |
19.3 A Generalized Linear Mixed Model | p. 340 |
19.4 Marginalizing the Mixed Model | p. 342 |
20 Non-linear Models | p. 347 |
20.1 Introduction | p. 347 |
20.2 Univariate Non-linear Models | p. 349 |
20.3 The Indomethacin Study: Non-hierarchical Analysis | p. 351 |
20.4 Non-linear Models for Longitudinal Data | p. 355 |
20.5 Non-linear Mixed Models | p. 357 |
20.6 The Orange Tree Data | p. 358 |
20.7 Pharmacokinetic and Pharmacodynamic Models | p. 360 |
20.8 The Songbird Data | p. 368 |
20.9 Discrete Outcomes | p. 376 |
20.10 Hypothesis Testing and Non-linear Models | p. 379 |
20.11 Flexible Functions | p. 379 |
20.12 Using SAS for Non-linear Mixed-effects Models | p. 384 |
21 Pseudo-Likelihood for a Hierarchical Model | p. 393 |
21.1 Introduction | p. 393 |
21.2 Pseudo-Likelihood Estimation | p. 394 |
21.3 Two Binary Endpoints | p. 397 |
21.4 A Meta-analysis of Trials in Schizophrenic Subjects | p. 401 |
21.5 Concluding Remarks | p. 403 |
22 Random-effects Models with Serial Correlation | p. 405 |
22.1 Introduction | p. 405 |
22.2 A Multilevel Probit Model with Autocorrelation | p. 406 |
22.3 Parameter Estimation for the Multilevel Probit Model | p. 408 |
22.4 A Generalized Linear Mixed Model with Autocorrelation | p. 410 |
22.5 A Meta-analysis of Trials in Schizophrenic Subjects | p. 412 |
22.6 SAS Code for Random-effects Models with Autocorrelation | p. 415 |
22.7 Concluding Remarks | p. 417 |
23 Non-Gaussian Random Effects | p. 419 |
23.1 Introduction | p. 419 |
23.2 The Heterogeneity Model | p. 421 |
23.3 Estimation and Inference | p. 423 |
23.4 Empirical Bayes Estimation and Classification | p. 427 |
23.5 The Verbal Aggression Data | p. 428 |
23.6 Concluding Remarks | p. 435 |
24 Joint Continuous and Discrete Responses | p. 437 |
24.1 Introduction | p. 437 |
24.2 A Continuous and a Binary Endpoint | p. 439 |
24.3 Hierarchical Joint Models | p. 445 |
24.4 Age Related Macular Degeneration Trial | p. 448 |
24.5 Joint Models in SAS | p. 455 |
24.6 Concluding Remarks | p. 464 |
25 High-dimensional Joint Models | p. 467 |
25.1 Introduction | p. 467 |
25.2 Joint Mixed Model | p. 469 |
25.3 Model Fitting and Inference | p. 471 |
25.4 A Study in Psycho-Cognitive Functioning | p. 473 |
VI Missing Data | p. 479 |
26 Missing Data Concepts | p. 481 |
26.1 Introduction | p. 481 |
26.2 A Formal Taxonomy | p. 482 |
27 Simple Methods, Direct Likelihood, and WGEE | p. 489 |
27.1 Introduction | p. 489 |
27.2 Longitudinal Analysis or Not? | p. 490 |
27.3 Simple Methods | p. 491 |
27.4 Bias in LOCF, CC, and Ignorable Likelihood | p. 495 |
27.5 Weighted Generalized Estimating Equations | p. 498 |
27.6 The Depression Trial | p. 499 |
27.7 Age Related Macular Degeneration Trial | p. 503 |
27.8 The Analgesic Trial | p. 507 |
28 Multiple Imputation and the EM Algorithm | p. 511 |
28.1 Introduction | p. 511 |
28.2 Multiple Imputation | p. 511 |
28.3 The Expectation-Maximization Algorithm | p. 516 |
28.4 Which Method to Use? | p. 526 |
28.5 Age Related Macular Degeneration Study | p. 527 |
28.6 Concluding Remarks | p. 529 |
29 Selection Models | p. 531 |
29.1 Introduction | p. 531 |
29.2 An MNAR Dale Model | p. 532 |
29.3 A Model for Non-monotone Missingness | p. 543 |
29.4 Concluding Remarks | p. 552 |
30 Pattern-mixture Models | p. 555 |
30.1 Introduction | p. 555 |
30.2 Pattern-mixture Modeling Approach | p. 556 |
30.3 Identifying Restriction Strategies | p. 557 |
30.4 A Unifying Framework for Selection and Pattern-mixture Models | p. 561 |
30.5 Selection Models versus Pattern-mixture Models | p. 563 |
30.6 Analysis of the Fluvoxamine Data | p. 567 |
30.7 Concluding Remarks | p. 572 |
31 Sensitivity Analysis | p. 575 |
31.1 Introduction | p. 575 |
31.2 Sensitivity Analysis for Selection Models | p. 576 |
31.3 A Local Influence Approach for Ordinal Data with Dropout | p. 578 |
31.4 A Local Influence Approach for Incomplete Binary Data | p. 585 |
31.5 Interval of Ignorance | p. 590 |
31.6 Sensitivity Analysis and Pattern-mixture Models | p. 604 |
31.7 Concluding Remarks | p. 605 |
32 Incomplete Data and SAS | p. 607 |
32.1 Introduction | p. 607 |
32.2 Complete Case Analysis | p. 607 |
32.3 Last Observation Carried Forward | p. 609 |
32.4 Direct Likelihood | p. 611 |
32.5 Weighted Estimating Equations (WGEE) | p. 613 |
32.6 Multiple Imputation | p. 618 |
32.7 The EM Algorithm | p. 633 |
32.8 MNAR Models and Sensitivity Analysis Tools | p. 635 |
References | p. 637 |
Index | p. 671 |