[R-meta] Quick question on metaprop output
Joao Afonso
jot@|on@o @end|ng |rom gm@||@com
Wed Feb 19 11:58:19 CET 2020
Dear all,
I think I have managed to set the model for running the meta-analysis
using metaprop function and run the outlier and influential case
diagnostic using the InfluenceAnalysis function.
I have a question regarding the output of my subgroup analysis. After
running the code below:
metaprop(nlameanimal, ssizeanimal, author, data=prevalence_2020_noout,
method="inverse", sm="PLOGIT", method.tau="DL", method.ci="NAsm",
byvar=lcmbi, tau.common=TRUE,prediction = TRUE)
I get
Number of studies combined: k = 40
proportion 95%-CI z p-value
Fixed effect model 0.1934 [0.1920; 0.1949] -- --
Random effects model 0.2526 [0.1991; 0.3150] -- --
Prediction interval [0.0445; 0.7104]
Quantifying heterogeneity:
tau^2 = 0.9339; H = 30.19 [29.42; 30.98]; I^2 = 99.9% [99.9%; 99.9%]
Quantifying residual heterogeneity:
tau^2 = 0.7064; H = 24.86 [24.12; 25.62]; I^2 = 99.8% [99.8%; 99.8%]
Test of heterogeneity:
Q d.f. p-value
35545.91 39 0
Results for subgroups (fixed effect model):
k proportion 95%-CI
Q tau^2
lcmbi = Records 12 0.1351 [0.1337; 0.1366]
20819.72 0.7064
lcmbi = Locomotion Scoring Method 28 0.3157 [0.3124; 0.3189]
2661.60 0.7064
I^2
lcmbi = Records 99.9%
lcmbi = Locomotion Scoring Method 99.0%
Test for subgroup differences (fixed effect model):
Q d.f. p-value
Between groups 12064.59 1 0
Within groups 23481.32 38 0
Results for subgroups (random effects model):
k proportion 95%-CI
Q tau^2
lcmbi = Records 12 0.1834 [0.1207; 0.2687]
20819.72 0.7064
lcmbi = Locomotion Scoring Method 28 0.2877 [0.2265; 0.3578]
2661.60 0.7064
I^2
lcmbi = Records 99.9%
lcmbi = Locomotion Scoring Method 99.0%
Test for subgroup differences (random effects model):
Q d.f. p-value
Between groups 3.83 1 0.0505
Within groups 23481.32 38 0
Details on meta-analytical method:
- Inverse variance method
- DerSimonian-Laird estimator for tau^2 (assuming common tau^2 in subgroups)
- Logit transformation
- Normal approximation confidence interval for individual studies
Previously I would get the amount of heterogeneity explain by the
moderator with r^2. However I can't find this parameter in the output
above. Am I missing something in my code to generate this parameter?
Many thank and sorry for flooding your email box with so many messages.
Wishing all a great day,
Joao
--
João Afonso
DVM, MSc Veterinary Epidemiology
PhD Student
Department of Infection and Global Health
University of Liverpool
+351914812305
More information about the R-sig-meta-analysis
mailing list