Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.

Background and Objectives: In systematic reviews of interventions of studies where randomization was done by individual but data are paired (such as eyes, ears), it is necessary to account for the natural clustering present. The Cochrane Handbook suggests treating these as examples of cluster randomized trials. An incorrect analysis (without adjustment) would usually overestimate the precision of the estimate. We discuss a simple method of adjustment that deals with this problem. Methods: From a cross-tabulation of the event being present on the "left" and "right" body part, we estimate the design effect which is a measure of the inflation on the variance due to clustering. This estimate is then used to obtain an adjusted effect size per trial by reducing the number of events and the sample size in each intervention group. Results: In a systematic review on Auto-inflation for Glue Ear, data on improvement were obtained for pairs of ears. The design effect obtained from these data was 1.25. In a meta-analysis, the weights given to the trials changed after adjustment from 33% to 11% in one case. Conclusion: In a systematic review, when dealing with paired data, it is possible to give adequate weighting to each trial using a simple adjusting method. © 2007 Elsevier Inc. All rights reserved.

Original publication

DOI

10.1016/j.jclinepi.2006.12.004

Type

Journal article

Journal

Journal of Clinical Epidemiology

Publication Date

01/09/2007

Volume

60

Pages

975 - 978