The Statistical Analysis of Recurrent Events

Abstract

Recurrent event data arise in diverse fields such as medicine, publichealth, insurance, social science, economics, manufacturing and reliability.The purpose of this book is to present models and statistical methodsfor the analysis of recurrent event data. No single comprehensivetreatment of these areas currently exists. The authors provide broadbut detailed coverage of the major approaches to analysis, whilealso emphasizing the modeling assumptions that they are based on.Thus, they consider important models such as Poisson and renewalprocesses, with extensions to incorporate covariates or random effects.More general intensity-based models are also considered, as wellas simpler models that focus on rate or mean functions. Parametric,nonparametric and semiparametric methodologies are all covered, withclear descriptions of procedures for estimation, testing and modelchecking. Important practical topics such as observation schemesand selection of individuals for study, the planning of randomizedexperiments, events of several types, and the prediction of futureevents are considered. Methods of modeling and analysis are illustratedthrough many examples taken from health research and industry. Theobjectives and interpretations of different analyses are discussedin detail, and issues of robustness are addressed. Statistical analysisof the examples is carried out with S-PLUS software and code is givenfor some examples. This book is directed at graduate students, researchers,and applied statisticians working in industry, government or academia.Some familiarity with survival analysis is beneficial since survivalsoftware is used to carry out many of the analyses considered. Thisbook can be used as a textbook for a graduate course on the analysisof recurrent events or as a reference for a more general course onevent history analysis. Problems are given at the end of chaptersto reinforce the material presented and to provide additional backgroundor extensions to certain topics. Richard J. Cook is Professor inthe Department of Statistics and Actuarial Science at the Universityof Waterloo and Canada Research Chair in Statistical Methods forHealth Research. He is an Associate Editor for Lifetime Data Analysis.Jerald F. Lawless is Professor in the Department of Statistics andActuarial Science at the University of Waterloo. He is a former Editorof Technometrics and from 1994-2004 held the General Motors Canada-NSERCIndustrial Research Chair in Quality and Productivity. He is theauthor of Statistical Models and Methods for Lifetime Data, SecondEdition (2003).