Handbook of survival analysis (Q2865202)

This handbook presents methodology of modern survival analysis developed within the past thirty years including both frequentist and Bayesian techniques. The aims of the book are to provide introductory as well as more advanced material for graduate students and new researchers, to give a reference of modern survival analysis as well as to help practitioners with their survival data experiments. The book consists of 30 chapters grouped into the following six parts:NEWLINENEWLINE (I) Regression models for right censoring: Statistical methods for right-censored and left-truncated survival data are presented with a strong focus on the Cox model.NEWLINENEWLINE (II) Competing risks: In many applications, there may be more than one cause for failures and the observed failure time is given as the minimum of failures due to each possible cause. In such a situation it is of interest to calculate the probability of failure due to a specific cause within a given time frame, the so called crude probabilities. In this part these probabilities are examined reaching from more classical techniques such as regression modeling based on a Cox model to more recent alternatives. Finally, a series of illustrations on medical data sets is given.NEWLINENEWLINE (III) Model selection and validation: While the focus lies on classical model selection and validation, one chapter is also devoted to robustness of the Cox regression model.NEWLINENEWLINE (IV) Other censoring schemes: In this part estimation for models with more difficult censoring schemes such as interval censoring are discussed.NEWLINENEWLINE Multivariate/multistate models: Multistate models deal with a complete disease/recovery process of a patient involving several health states. A second focus of this part lies on frailty models, which are the most common example of multivariate survival data.NEWLINENEWLINE (V) Clinical trials: In this part methods useful for both the design and analysis of clinical trials where the time to some event is the main interest are discussed.