- This line was added.
- This line was removed.
- Formatting was changed.
Performing Meta Analysis
After the forest-plot has been generated, click the blue [Export] button and select the format and resolution required for your export
Computational problems can occur when there are cells with zero events for the meta-analysis of dichotomous values.
When using the Inverse Variance or the Mantel-Haenszel statistical methods, systematic reviewers will need to add a continuity correction of 0.5 to each cell, for which a zero event occurs. E.g. If I have zero events out of 100 in the control group and 1 event out of 100 in the experimental group, this will need to change to 0.5 events out of 100.5 in the control group and 1.5 events out of 100.5 in the experimental group. The appropriateness of doing such a correction is debatable but yields consistent results in most scenarios.
When using the Peto Odds Ratio as the statistical method, this continuity correction does not need to be manually input the reviewer. The Peto Odds Ratio is capable of dealing with zero event cells through the underlying statistical formula. However, using such a method is usually reserved for outcome data that is extremely rare, or uncommon.
No matter the approach you take to dealing with zero event data withing SUMARI, it needs to be clearly and transparently documented in the review protocol and the systematic review itself.
Performing Qualitative Synthesis
To export the meta-aggregative flowchart hover your cursor over the flowchart. A pop-up should appear providing you with different image formats in which the flowchart can be exported.