Analysis of Multivariate Survival Data

Rating:
Summary: first book on multivariate survival analysis
Review: The author is one of the pioneers in the newly developing field of multivariate survival analysis. His work goes back to his Ph.D. dissertation in the mid 1980s. These methods come into play when one is studying more than one survival curve and the event times are correlated rather than independent. Practical applications include situations when multiple events are studied on the same patients, such as time until contracting the disease, followed by time to complications and then possibly by time to death from the disease. Studies involving events related to twins can also be analyzed by these methods.

He gives an excellent exposition and a number of good examples. He provides the reader with a very current list of references from the literature.

The author presents the four common approaches to the problem and concedes that the field is in its infancy. He believes that while some of the methods described will prove not to be as fruitful as others, at this point it is still difficult to determine which are the most promising. His aim is to expand the toolbox for researchers in medical and biological fields who have experience with univariate survival analysis and may be faced with multivariate problems. He covers such important current topics as fraility models and competing risks.

In my opinion the author has succeeded in his goal and provided biostatisticians with a reference source that will be useful to them for many years. It should not be your first book in survival analysis though. See the book by Lawless or Kalbfleish and Prentice before attaching this book.