Comparing Methods to Detect Publication Bias in Meta-Analysis

Egger's regression test detected publication bias in 15.7% of meta-analyses of binary outcomes.

• Seven publication bias tests were compared using 28,655 meta-analyses from the Cochrane Library.
• Larger meta-analyses showed a greater proportion of statistically significant publication bias tests, especially with Begg's rank test and trim-and-fill method.
• Agreement among Tang's, Macaskill's, Deeks', and Peters' regression tests for binary outcomes was moderately strong, with most κ's around 0.6.
• Tang's and Deeks' tests demonstrated very high agreement, with κ > 0.9.
• Weak to moderate agreement was observed among Begg's rank test, trim-and-fill method, and Egger's regression test, with κ < 0.5.
• Multiple tests may be necessary for meta-analysts due to the low agreement among the tests.

Simplified