Request PDF on ResearchGate | Analysis of Multivariate Survival Data | Introduction.- Univariate survival data. Philip Hougaard at Lundbeck. Philip Hougaard. This book is, at it states in the preface, a tool box rather than a cookbook, for those wishing to analyse multivariate survival data. It would thus be. Analysis of Multivariate Survival Data. Philip Hougaard, Springer, New York, No. of pages: xvii+ Price: $ ISBN 0‐‐‐4.
|Published (Last):||24 September 2005|
|PDF File Size:||16.18 Mb|
|ePub File Size:||13.13 Mb|
|Price:||Free* [*Free Regsitration Required]|
This book is, at it states in the preface, a tool box rather than a cookbook, for those wishing to analyse multivariate survival data.
It would thus be of most relevance to applied statisticians or epidemiologists requiring a theoretical and practical grounding in the analysis of such data.
Analysis of Multivariate Survival Data. | International Journal of Epidemiology | Oxford Academic
The book divides into three main sections: The first chapter briefly describes the main features of survival data, and the two main types of multivariate survival data parallel and longitudinal.
The various datasets used as examples throughout the text are then detailed, and the five main aims of multivariate survival analysis hoguaard in a table. The description of each dataset is helpfully cross-referenced to the later sections in which the dataset is analysed. A chapter summarizing approaches to univariate survival data follows, with indications as to which sections are most important as forming the basis for development of the different multivariate models.
The three dependence mechanisms—common events, common risks and event-related dependence—are outlined in a non-mathematical chapter, with a useful table showing common data types relating to these three mechanisms.
A chapter describing various measures of bivariate dependence follows. The main part of the book consists of ten chapters outlining each of the four main approaches to multivariate survival analysis: Various aspects of the theory and statistical inference for each of these approaches are discussed, including helpful sections on assessing goodness-of-fit and choosing between the different models available within each approach.
These chapters contain much theoretical development, including statistical derivation and issues around estimation of the various models, and are more mathematically-orientated than the rest of the book.
Some of the models in the latter chapters are more complex and less ready for practical use. The last chapter provides a very useful summary of the text, with cross-references to the appropriate sections throughout. The summary of the theory includes a table outlining questions to consider when identifying the best model to use in a given situation.
A practical section on the course of analysis includes tables and discussion ddata which models are appropriate for which type of data and the relevance of each approach eurvival various purposes. Questions to consider before choosing between specific multi-state models, frailty models, marginal models and non-parametric approaches are considered in more detail in four separate tables.
A table outlines the limitations of each of the four main approaches. The chapter concludes with a summary of the datasets discussed throughout the text, discussing the main questions and which models are used to answer them. One of the most useful aspects of this book, in my opinion, is the extensive use made of practical examples.
The datasets are described fully in the introduction, and include several examples of each of the more common types of multivariate data. The example discussed the most often, the Danish twins study, is one which will be of particular relevance to those involved in genetics studies.
The datasets on length of leukaemia remissions, number of epileptic seizures, exercise test times and competing risks all show types of data which occur in different types of epidemiological study. These datasets are analysed throughout the text, and results from the various different models presented, interpreted and compared.
Analysis of Multivariate Survival Data
For some of the datasets, the data are given in the introduction in tabular form, so the reader could attempt to analyse the data and compare the results with those presented. Code for statistical programs mostly in SAS, with some examples in Hougaaard is given for some of the examples.
There are exercises at the end of each chapter. In the case of the main chapters describing the different approaches, these are theoretically-based, and include examples of deriving transition probabilities for the multi-state model and survivor functions frailty models. These would be of most use for those seeking to understand fully the underlying mathematical statistics of these models.
The exercises at the end of the more applied chapters relate more to the identification of sources of bias, dependence mechanisms and time-frames, study design and choice of analysis. Several daata the exercises suggest analyses of specific datasets described in the introduction. The organization of the book, and the good use of cross-referencing, mean that it can be read in varying degrees of depth.
This book should prove an informative extension to the literature on survival analysis. Oxford University Press is a department of the University of Oxford. It furthers the University’s survifal of excellence in research, scholarship, and education by publishing worldwide. Sign In or Create an Account. Close mobile search navigation Article navigation.
Analysis of Multivariate Survival Data. Email alerts New issue alert. Receive exclusive offers and updates from Oxford Academic.
Analysis of Multivariate Survival Data : Philip Hougaard :
Related articles in Google Scholar. Citing articles via Google Scholar.
Circulating vitamin D concentrations and risk of breast and prostate cancer: Poor diet quality in pregnancy is associated with increased risk of excess fetal growth: Socioeconomic position during pregnancy and DNA methylation signatures at three stages across early life: