Workshop W27

Workshop registration closed!

The facility to register for workshops has closed. If you have not registered for a workshop, workshop listings will be on the Message Board in the conference foyer of The Burlington Hotel for you to register for workshops that still have spaces available.

Missing data in meta-analysis - a practical guide.

James Carpenter.
Medical Statistics Unit, London School of Hygiene and Tropical Medicine, London, UK.

james.carpenter doesnotexist@cochrane.org lshtm.ac.uk

Objectives:

To introduce participants to the key issues raised by missing data in clinical trials;
To outline a systematic approach for thinking about these issues;
To provide guidance on reviewing trial reports where the trial has a non-trivial proportion of missing data;
To review methods for including studies with missing data in meta-analyses;
To work through examples illustrating these issues in small groups.

Summary: Missing data are ubiquitous in clinical trials. They introduce a level of ambiguity into the analysis and conclusions beyond conventional sampling imprecision. This is because additional assumptions, for example about the reason for the missing observations, must be made to justify the analysis and hence the conclusions. For a broad introduction, see www.missingdata.org.uk.

Building on the approach described by Carpenter and Kenward (2006), we argue that a principled approach to the analysis of partially observed data sets is preferable to an ad-hoc approach, and describe how such an approach might work in practice.

We show how this approach can inform meta-analysis when some contributing trials have non-trivial proportions of missing data. When faced with such a trial, we propose the initial focus should be on understanding whether the methods used for the analysis are appropriate for the intention-to-treat hypothesis. If not, we consider the likely effects on the presented results. Then we consider how to include such trials in a meta-analysis. We discuss established and recently proposed methods for doing this (Gamble and Hollis, 2005).

Small group practical work will be an integral part of the workshop and focus on critically reviewing reports of trials with non-trivial missing data issues, with a view to deciding how to allow for this in a subsequent meta-analysis.

The workshop will focus on concepts, keeping technical details in the background.

Level of knowledge required to attend: advanced.