Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Multilevel meta-analysis: Forest plot

    Hello,

    I am using -meta multilevel- (Stata 18.5) to conduct a three-level random-effects meta-analysis with SMDs nested within studies to account for dependence (as studies contribute multiple effect sizes).

    Using the code below resulted in an overall SMD -0.85, 95% CI -1.15 to -0.54 and an I2 = 92.1%.
    Code:
    meta set Hedgesgrm SE, random(sjonkman) studylabel(Study) studysize(N)
    meta multilevel Hedgesgrm, relevels(Study) essevariable(SE)
    estat heterogeneity
    But when I try to make a forest plot using the below, it displays the results from a conventional meta-analysis treating each effect size as independent (Fig.1 at end of post).
    Code:
    meta forestplot, subgroup(Timepoint)
    I can change the overall effect size from -1.31 to -0.85 using the below, but I’m not sure if/how it is possible to do this with subgroups.
    Code:
    meta forestplot, customoverall(-0.85  -1.15 -0.54)
    I am hoping to display the results in the same format as Fig.1 but where the SMD for each subgroup and overall takes into account dependencies. Any suggestions greatly appreciated – thank you!

  • #2
    Fig.1

    Click image for larger version

Name:	Forest plot.png
Views:	1
Size:	69.0 KB
ID:	1773513

    Comment

    Working...
    X