>> Home >> Bookstore >> Title index >> Biostatistics and epidemiology >> Cluster Randomised Trials

Cluster Randomised Trials

Click to enlarge
See the back cover

$84.50 Print Add to cart What are VitalSource eBooks? Your access code will be emailed upon purchase. eBook not available for this title  Authors: Richard J. Hayes and Lawrence H. Moulton Publisher: Chapman & Hall/CRC Copyright: 2009 ISBN-13: 978-1-58488-816-1 Pages: 315; hardcover Price:$84.50
 Authors: Richard J. Hayes and Lawrence H. Moulton Publisher: Chapman & Hall/CRC Copyright: 2009 ISBN-13: Pages: 315; eBook Price: \$0.00
 Supplements: datasets

Comment from the Stata technical group

The cluster randomized trial (CRT) is the “gold standard” for evaluating the effectiveness of medical interventions because

• One can assemble treatment and control groups that are similar in every attribute, both measurable and unmeasurable, with only the treatment differing between groups
• The control group is followed contemporaneously with the treatment group, making CRTs robust to outcome variations over time
• Randomizing treatment allocations provides further protection from confounding
• Allocating subjects as clusters provides both convenience and protection from contamination between treatment arms

Cluster Randomised Trials describes in detail the aspects of CRT designs and the analysis of data from such trials. After presenting some basic concepts, Hayes and Moulton focus on the clustering aspect of the design. They cover between- and within-cluster variability and correlation, advantages and disadvantages of clustering, choosing clusters to minimize bias and contamination, and other related topics. Middle chapters deal with matching, stratification, randomization procedures, and sample-size calculations. The authors then focus on statistical analysis, with the discussion including calculations of rate differences and rate ratios, t tests for means, nonparametric tests of equality between treatment groups, Cox regression, logistic regression, Poisson regression, and mixed-effects models. The authors close with miscellaneous topics, such as the ethical considerations of randomization and establishing stopping rules.

All analyses are performed in Stata, and the data used are freely available, making the analyses easy to reproduce.