This module will discuss differences between studies. We learn how to recognize and deal with differences between studies in a review. Some statistical methods such as random effects meta-analysis, subgroup analysis and meta-regression can help address these differences, and this module will explain these techniques.
Learning objectives
- Understand that studies usually differ both clinically and methodologically
- Appreciate that differences between studies can result in (statistical) heterogeneity – differences in their results
- Be able to identify heterogeneity
- Know some strategies for dealing with heterogeneity
- Understand the difference between fixed effect and random effects meta-analysis
- Understand when to use subgroup analyses
- Be aware of meta-regression as a tool for exploring differences between studies
Relevant sections of the Reviewers’ Handbook
- Section 8
Where does this go in a Cochrane review?
- In the protocol where you will need to specify factors that you are planning to investigate as potential causes of heterogeneity
- The Methods section of the review should detail what you have done to identify or examine heterogeneity. It is especially important to tell the reader which analyses were pre-specified and which were ‘post hoc’, i.e. designed after collecting all the studies for your review
- The Results section should present results of subgroup analyses and meta-regressions (if used). Remember to interpret the results with caution, and keep in mind the possibility that findings may be spurious if you do more than very few analyses