|
Bayesian methods are being used more often than ever before in biology and
medicine. For example, at the University of Texas MD Anderson Cancer
Center, Bayesian sequential stopping rules routinely are used for the design
of clinical trials. This book is based on the author’s experience working with
a variety of researchers, including radiologists, pathologists, and medical
oncologists. The majority of that experience has been with the Division of
Diagnostic Imaging, where radiologists determine the extent of disease
among patients undergoing treatment. Diagnosis, via medical imaging, is
essential in order to assess the effect of the various therapies provided to the
patient. Another source of information for the author has been the ability to
work with medical oncologists in their design of Phase I, II, and III clinical
trials. The author has found Bayesian methods for the design and analysis
of clinical trials to be quite useful because prior information, in the form of
previous related studies, is always available and easily incorporated into the
design of future studies.
Based on this experience and the wealth of information available to the
author, this book should give the biostatistics student a good idea of what
to expect and how to work with healthcare researchers. It is an introductory
book with a Bayesian flavor and is directed toward diagnostic medicine.
Students with a good background in the basic methods courses of regression
and the analysis of variance and in the introductory courses in probability
and mathematical statistics should benefit greatly from the book. With this
type of background, the student will be able to learn Bayesian statistics and
how to apply it to important problems in medicine and biology. In addition,
it should serve as a useful reference for those providing statistical assistance
to medical scientists.
In the book, the reader is introduced to various diagnostic medical procedures,
then presented with the fundamentals of Bayesian statistics and associated
computing methods. Next, the foundation for the analysis of
diagnostic test accuracy is outlined and the Bayesian way to analyze such
data is explained, using many author-assisted studies. Of special interest is
the estimation of the area under the receiver operating characteristic (ROC)
curve for determining diagnostic accuracy. Also described in the book is a
novel way to estimate the area when the image data are clustered. |
