Is there Publication Bias?

Publication bias occurs when the results of a study influence whether it gets published or not. Trials with positive results are more likely to be published and then included in a systemic review. The systematic review will then be biased towards finding a positive result. There are some very clever ways to assess for publication bias in a systematic review.

A funnel plot, like the examples below, uses some mathematical wizardry to do this. You certainly dont need to know how to do this, but you do need to know that they exist and how to interpret them.

The bottom of the funnel represents the smaller studies and the middle line represents the estimated effect. The idea is that smaller studies have more random error around the true estimate. If all the studies that investigated the question are included, then the funnel plot should have symmetrical bases and the point of the funnel would be in the middle, indicating there isnt much publication bias.

However, if small trials with negative results were excluded, then the shape of the funnel would appear to have a hole on one side and the point of the funnel would not be in the middle. This indicates publication bias.

Fig 1: Funnel plot 1 Fig 2: Funnel plot 2

 

Which funnel plot has come from a meta-analysis with little publication bias?

Funnel plot 1 suggests publication bias and funnel plot 2 suggests a low chance of publication bias.