Meta-analysis of continuous data

This module will discuss analysing data from continuous outcomes. We address issues of data extraction and meta-analysis of continuous data, and consider some of the problems reviewers frequently encounter.

Learning objectives

  • Identify continuous data
  • Identify scales and ordinal outcomes that can reasonably be treated as continuous
  • Understand the concepts of mean and standard deviation
  • Be aware that you can convert between standard deviations, standard errors and confidence intervals
  • Understand the use of differences between means and standardized differences between means as measures of treatment effect
  • Be aware of problems when dealing with skewed data and non-parametric summaries
  • Understand the methods of combining continuous data in meta-analysis
  • Be aware of issues in choosing between change scores and final values

Relevant sections of the Reviewers’ Handbook

  • Section 8

Where does this go in a Cochrane review?

  • Details of how continuous data are analyzed should be given in the Methods section. When presenting results of analyses involving continuous data, it should always be clear what method has been used, what the units of measurement are and, for scale based outcomes, a description of what the measurement scale means, noting whether the scale goes up or down with improving outcome. If assumptions are made when performing analyses, these should be assessed and if potentially important, addressed in the Discussion