β Scribed by Michael J. Campbell; Allan Donner; Diana R. Elbourne
No coin nor oath required. For personal study only.
A total of 69 people were in attendence. Invited speakers were Patrick Heagerty of the University of Seattle, who spoke on the analysis of cluster trials using generalized estimating equations, David Spiegelhalter of the MRC Biostatistics Unit in Cambridge, U.K., who spoke on Bayesian analysis of cluster trials and Jane Hutton of the University of Newcastle, U.K., who spoke on ethics and cluster trials. Min Yang from the Multilevel Project at the Institute for Education, London, gave a demonstration of MlWin and Gilda Piaggio from the World Health Organization in Geneva demonstrated software for the design and analysis of cluster randomized trials. Twenty-one people gave contributed papers on a variety of aspects of cluster trials. Twelve speakers subsequently submitted manuscripts, which, after subsequent refereeing, appear in this special issue of the journal.
As would be expected, the majority of attendees were from the U.K., but (including invited speakers) there were three from the U.S.A. and Holland, two from Finland, Denmark and Australia, and one each from Norway, France, Spain, Thailand, Canada and Switzerland. In general the feedback was very positive, and both the course and workshop were felt to be very successful.
We have grouped the papers to cover issues of design, sample size, analysis, Bayesian methods, ethics and reporting, although there is obviously overlap between these. The issue begins with three papers looking at di erent aspects of design, two relating to conducting cluster randomized trials in general (family) practice (Moore et al. and Yudkin and Moher) and the third looking at the role of balancing covariates (Raab and Butcher). A group of three papers then examines issues related to sample size and estimation of intracluster correlation coe cients (Eldridge et al., Kerry and Bland and Campbell et al.). Piaggio et al. describe design and analysis issues arising from a stratiΓΏed equivalence cluster randomized trial. The analysis issues related to repeated cross-sectional binary measurements are considered by Ukoumunne and Thompson. Two further papers explore the role of Bayesian methods for continuous and binary outcomes (Spiegelhalter and Turner et al., respectively). Ethical issues are the focus of the paper by Hutton. The issue concludes with a proposed extension of the CONSORT guidelines [2] for the reporting of cluster randomized trials (Elbourne and Campbell).
We would like to thank the following who helped in the refereeing process: Marion Campbell;
π SIMILAR VOLUMES
This information is provided as a professional service to the medical profession by Merck & Co., Inc. The views expressed in this publication reflect the experience and opinions of the authors. The prescribing information from the Company (manufacturer) for any product described in this publication