Classical competing risks.

*(English)*Zbl 0979.62078
Boca Raton, FL: Chapman & Hall/ CRC. 186 p. (2001).

This book is about likelihood-based inference for competing risks data in discrete and continuous time. The situation is first described in terms of bivariate random variables, with the first component being a survival time and the second a discrete random variable giving the cause of death. Then the traditional approach with latent survival times of which only the smallest, together with information of which is the smallest, is observed is introduced.

Models for competing risks are described both for continuous time and discrete time data, and both in a parametric as well as in a nonparametric fashion. Frailty is briefly introduced, as are alternatives to the proportional hazards model, namely, proportional odds and accelerated failure time models.

More mathematical approaches to survival analysis, like counting processes and martingales, are avoided througout the book. The exception is the last chapter, which contains a brief introduction to this challenging field.

Models for competing risks are described both for continuous time and discrete time data, and both in a parametric as well as in a nonparametric fashion. Frailty is briefly introduced, as are alternatives to the proportional hazards model, namely, proportional odds and accelerated failure time models.

More mathematical approaches to survival analysis, like counting processes and martingales, are avoided througout the book. The exception is the last chapter, which contains a brief introduction to this challenging field.

Reviewer: Göran Broström (Umea)