Analysis of Longitudinal Data, Second Edition

"The new edition of this important text has been completely revised and expanded to become the most up-to-date and thorough professional reference text in this fast-moving and important area of biostatistics. Two new chapters have been added on fully parametric models for discrete repeated measures data and on statistical models for time-dependent predictors where there may be feedback between the predictor and response variables. It also contains the many useful features of the previous edition such as, design issues, exploratory methods of analysis, linear models for continuous data, and models and methods for handling data and missing values."--OUP. ANALYSIS OF LONGITUDINAL DATA, 2ND ED.......Page 1 Oxford Statistical Science Series......Page 2 Title Page......Page 3 Copyright Page......Page 4 Dedication......Page 5 Preface......Page 7 Contents......Page 11 1.1 Longitudinal studies......Page 17 1.2 Examples......Page 19 1.3 Notation......Page 31 1.4 Merits of longitudinal studies......Page 32 1.5 Approaches to longitudinal studies......Page 33 1.6 Organization of subsequent chapters......Page 36 2.2 Bias......Page 38 2.3 Efficiency......Page 40 2.4 Sample size calculations......Page 42 2.5 Further reading......Page 47 3.1 Introduction......Page 49 3.2 Graphical presentation of longitudinal data......Page 50 3.3 Fitting smooth curves to longitudinal data......Page 57 3.4 Exploring correlation structure......Page 62 3.5 Exploring association amongst categorical responses......Page 68 3.6 Further reading......Page 69 4.1 Motivation......Page 70 4.2.1 The uniform correlation model......Page 71 4.2.2 The exponential correlation model......Page 72 4.2.3 Two-stage least-squares estimation and random effects models......Page 73 4.3 Weighted least-squares estimation......Page 75 4.4 Maximum likelihood estimation under Gaussian assumptions......Page 80 4.5 Restricted maximum likelihood estimation......Page 82 4.6 Robust estimation of standard errors......Page 86 5.1 Introduction......Page 97 5.2 Models......Page 98 5.2.1 Pure serial correlation......Page 100 5.2.2 Serial correlation plus measurement error......Page 105 5.2.3 Random intercept plus serial correlation plus measurement error......Page 106 5.2.4 Random effects plus measurement error......Page 107 5.3 Model-fitting......Page 109 5.3.1 Formulation......Page 110 5.3.2 Estimation......Page 111 5.3.3 Inference......Page 113 5.3.4 Diagnostics......Page 114 5.4 Examples......Page 115 5.5 Estimation of individual trajectories......Page 126 5.6 Further reading......Page 129 6.1 Preliminaries......Page 130 6.2 Time-by-time ANOVA......Page 131 6.3 Derived variables......Page 132 6.4 Repeated measures......Page 139 6.5 Conclusions......Page 141 7.1 Marginal models......Page 142 7.2 Random effects models......Page 144 7.3 Transition (Markov) models......Page 146 7.4 Contrasting approaches......Page 147 7.5 Inferences......Page 153 8.1 Introduction......Page 157 8.2.1 The log-linear model......Page 158 8.2.2 Log-linear models for marginal means......Page 159 8.2.3 Generalized estimating equations......Page 162 8.3 Examples......Page 164 8.4.1 Parametric modelling for count data......Page 176 8.4.2 Generalized estimating equation approach......Page 178 8.5 Sample size calculations revisited......Page 181 8.6 Further reading......Page 183 9.1 Introduction......Page 185 9.2.1 Conditional likelihood......Page 187 9.2.2 Maximum likelihood estimation......Page 188 9.3.1 Conditional likelihood approach......Page 191 9.3.2 Random effects models for binary data......Page 194 9.3.3 Examples of logistic models with Gaussian random effects......Page 196 9.4.1 Conditional likelihood method......Page 200 9.4.2 Random effects models for counts......Page 202 9.4.3 Poisson-Gaussian random effects models......Page 204 9.5 Further reading......Page 205 10.1 General......Page 206 10.2 Fitting transition models......Page 208 10.3 Transition models for categorical data......Page 210 10.3.1 Indonesian children's study example......Page 213 10.3.2 Ordered categorical data......Page 217 10.4 Log-linear transition models for count data......Page 220 10.5 Further reading......Page 222 11.1 Introduction......Page 224 11.2 Generalized linear mixed models......Page 225 11.2.1 Maximum likelihood algorithms......Page 228 11.2.2 Bayesian methods......Page 230 11.3 Marginalized models......Page 232 11.3.1 An example using the Gaussian linear model......Page 234 11.3.2 Marginalized log-linear models......Page 236 11.3.3 Marginalized latent variable models......Page 238 11.3.4 Marginalized transition models......Page 241 11.4.1 Crossover data......Page 247 11.4.2 Madras schizophrenia data......Page 250 11.5 Summary and further reading......Page 259 12.1 Introduction......Page 261 12.2 An example: the MSCM study......Page 263 12.3 Stochastic covariates: full and partly conditional means......Page 269 12.3.1 Estimation issues with cross-sectional models......Page 270 12.3.2 A simulation illustration......Page 272 12.3.3 MSCM data and cross-sectional analysis......Page 273 12.3.4 Summary......Page 274 12.4.1 A single lagged covariate......Page 275 12.4.2 Multiple lagged covariates......Page 276 12.4.3 MSCM data and lagged covanates......Page 277 12.5 Time-dependent confounders......Page 281 12.5.1 Feedback: response is an intermediate and a confounder......Page 282 12.5.2 MSCM data and endogeneity......Page 284 12.5.3 Targets of inference......Page 285 12.5.4 Estimation using g-computation......Page 289 12.5.5 MSCM data and g-computation......Page 291 12.5.6 Estimation using inverse probability of treatment weights (IPTW)......Page 292 12.5.7 MSCM data and marginal structural models using IPTW......Page 295 12.6 Summary and further reading......Page 296 13.1 Introduction......Page 298 13.2 Classification of missing value mechanisms......Page 299 13.3 Intermittent missing values and dropouts......Page 300 13.4.1 Last observation carried forward......Page 303 13.5 Testing for completely random dropouts......Page 304 13.6 Generalized estimating equations under a random missingness mechanism......Page 309 13.7.1 Selection models......Page 311 13.7.2 Pattern mixture models......Page 315 13.7.3 Random effect models......Page 317 13.7.4 Contrasting assumptions: a graphical representation......Page 319 13.8 A longitudinal trial of drug therapies for schizophrenia......Page 321 13.9 Discussion......Page 332 14.1 Non-parametric modelling of the mean response......Page 335 14.2 Non-linear regression modelling......Page 342 14.2.1 Correlated errors......Page 344 14.3 Joint modelling of longitudinal measurements and recurrent events......Page 345 14.4 Multivariate longitudinal data......Page 348 A.2 The linear model and the method of least squares......Page 353 A.3 Multivariate Gaussian theory......Page 355 A.4 Likelihood inference......Page 356 A.5.1 Logistic regression......Page 359 A.5.2 Poisson regression......Page 360 A.5.3 The general class......Page 361 A.6 Quasi-likelihood......Page 362 Bibliography......Page 365 Index......Page 385 Back Cover......Page 396 The first edition of Analysis for Longitudinal Data has become a classic. Describing the statistical models and methods for the analysis of longitudinal data, it covers both the underlying statistical theory of each method, and its application to a range of examples from the agricultural and biomedical sciences. The main topics discussed are design issues, exploratory methods of analysis, linear models for continuous data, general linear models for discrete data, and models and methods for handling data and missing values. Under each heading, worked examples are presented in parallel with the methodological development, and sufficient detail is given to enable the reader to reproduce the author's results using the data-sets as an appendix. This second edition, published for the first time in paperback, provides a thorough and expanded revision of this important text. It includes two new chapters; the first discusses fully parametric models for discrete repeated measures data, and the second explores statistical models for time-dependent predictors.

Aiming their work at the first year postgraduate statistic students, the authors (from Lancaster U., the U. of Washington, and Johns Hopkins U.) describe statistical models and methods for the analysis of longitudinal data, strongly emphasizing applications in biological and health sciences. After covering basic issues of design and exploratory analysis, they develop linear models and associated statistical methods for data sets in which the response variable is a continuous measurement. Later chapters examine generalized linear models for discrete response variables. Also explored are stochastic processes that may interact with response processes, solutions to the problems of missing values in longitudinal studies, and additional topics. Annotation c. Book News, Inc., Portland, OR