Sensitivity analysis
Sensitivity analysis was conducted using leave one-out method, that is, iteratively removing one study each time and repeating the analysis. The overall estimate of effect size was not significantly driven by a single study (Fig 4).