Table 3 Results of the multivariate meta-analysis
Interacrtions in the full model = Outcome*Baseline Value, Outcome*S/RSD Type, Outcome*Timing | |||||
Parameter | Estimated SMD | P value | 95% C.I | ||
Outcome: CAL | 0.17 | 0.76 | -0.92-1.26 | ||
Outcome: PPD | 1.17 | 0.01 | 0.27–2.07* | ||
Baseline Value | 0.01 | 0.87 | −0.10-0.12 | ||
S/RSD Type: PMD (reference value = FMD) | 0.22 | 0.50 | −0.42-0.86 | ||
Timing: Pre-therapy (reference value = Post-Therapy) | 0.13 | 0.72 | −0.57-0.83 | ||
PPD*Baseline | 0.11 | < 0.01 | 0.08–0.15*** | ||
PPD*S/RSD Type PMD | −1.25 | < 0.01 | −1.73--0.78*** | ||
PPD*Timing: Pre-therapy | −1.18 | < 0.01 | −1.48--0.87*** | ||
Model Statistics | |||||
Log Likelihood ratio | Deviance | AIC | BIC | ||
−38.06 | 76.13 | 94.20 | 99.88 | ||
Variance Components: Study, number of levels = 09, Estimated 휎2 value: 0.10 | |||||
Test for Residual Heterogeneity, Moderators and Model Comparison | |||||
Test for Residual Heterogeneity: QE (df = 14) = 88.50, p -value < 0.01 Test of Moderators: QM (df = 8) = 150.03, p-value < 0.01 Model comparsion with null model without moderators (ANOVA) | |||||
| Â | AIC | Log Likelihood ratio | Likelihood Ratio Test Statistic | P value | Model QE |
Full model (df = 9) | 94.20 | −38.06 |  |  | 88.50 |
Reduced (df = 3) | 185.23 | −89.54 | 103.10 | < 0.01 | 201.50 |