- Shopping Bag ( 0 items )
Doody's Review ServiceReviewer: Mary E Charlton, PhD (University of Iowa College of Public Health)
Description: This book describes a modeling framework for analysis and interpretation of data comparing measurement methods, and includes practical examples of the application of methods in some specific situations using Stata and SAS.
Purpose: The purpose is to provide practical tools to help readers interpret and conduct method comparison studies. Given the challenges of conducting these types of studies that involve the need for both clinical and statistical perspectives, this is a worthy objective, which the book meets.
Audience: The author assumes that his readers are familiar with standard statistical theory, linear models, and mixed models, but he also states that the general concepts can be understood by a more general audience. I think those without a statistical background would likely struggle to understand even the general ideas.
Features: Most of the early chapters in the book describe various common situations, such as two methods with a single measurement on each, replicate measurements, and several methods of measurements with single and replicate measures. Useful summaries of how to proceed with analyses appear at the end of most chapters. Chapter 7 describes a modeling framework to compare measurements taken from different methods. The later chapters describe transformation of data, and include discussions on repeatability, reproducibility, measures of association and agreement, and design of method comparison studies. The final chapters contain examples using SAS and Stata and information on the MethComp package for R. The book has a strange flow, but could serve as a good reference once readers know what they are looking for.
Assessment: This book presents useful information about the complexities of method comparison studies specific to clinical/biomedical research. It is not necessarily written at an easily accessible level for graduate students, clinicians, or researchers who do not have a statistical background, but it would be helpful for those who have a background in biostatistics/epidemiology and conduct method comparison studies. There is some overlap with Statistical Evaluation of Measurement Errors: Design and Analysis of Reliability Studies, Dunn (Wiley-Blackwell, 2009), but this book focuses more on common situations that arise in clinical healthcare research. I would consider using it in a course intended for students seeking advanced degrees in biostatistics and epidemiology.