Introduction to meta-analysis 5: including non-standard studies and non-standard data

Article type
Authors
Beyene J1, McKenzie J2
1McMaster University
2Statistical Methods Group, Australasian Cochrane Centre
Abstract
Objective:
The Cochrane Statistics Methods Group has developed a series of workshops addressing statistical guidelines as formulated in the Cochrane Handbook for Systematic Reviews of Interventions. This workshop will provide review authors with the knowledge to recognise and address situations in which a simple meta-analysis may not be appropriate.

Description:
Cochrane authors frequently face data that do not apparently fit into the usual dichotomous or continuous data meta-analyses that can be performed by RevMan. Typical examples are: (i) study designs, such as cross-over trials or cluster-randomised trials; (ii) types of outcome data, such as time-to-event data, ordered categorical data or rates. In this workshop we will review such ‘non-standard’ studies and data, and consider the possibilities for analysing individual studies and combining results in meta-analysis. RevMan offers a ‘generic inverse variance’ outcome type that can, in principle, be used to perform meta-analyses in all of these ‘non-standard’ situations.