Table 2 Results for the topical antibiotics data using importance sampling, MCMC in OpenBUGS and REML estimation. Estimates are log odds ratios
From: Two new methods to fit models for network meta-analysis with random inconsistency effects
Parameter | Estimate | SD | MCSE | 2.5 % | 50 % | 97.5 % | ||
|---|---|---|---|---|---|---|---|---|
Importance | B: Quinolone | δ AB | -1.85 | 0.62 | 0.0012 | -3.18 | -1.82 | -0.69 |
sampling | C: Non-quinolone | δ AC | -1.31 | 0.71 | 0.0014 | -2.78 | -1.29 | 0.03 |
D: Antiseptic | δ AD | -0.64 | 0.65 | 0.0013 | -2.01 | -0.62 | 0.60 | |
Heterogeneity | \(\tau ^{2}_{\beta }\) | 0.28 | 0.31 | 0.0006 | 0.01 | 0.18 | 1.09 | |
Inconsistency | \(\tau ^{2}_{\omega }\) | 0.22 | 0.38 | 0.0007 | 0.00 | 0.08 | 1.20 | |
MCMC | B: Quinolone | δ AB | -1.85 | 0.62 | 0.0052 | -3.15 | -1.82 | -0.68 |
C: Non-quinolone | δ AC | -1.30 | 0.71 | 0.0061 | -2.77 | -1.28 | 0.04 | |
D: Antiseptic | δ AD | -0.64 | 0.65 | 0.0053 | -1.99 | -0.62 | 0.61 | |
Heterogeneity | \(\tau ^{2}_{\beta }\) | 0.28 | 0.31 | 0.0017 | 0.01 | 0.18 | 1.09 | |
Inconsistency | \(\tau ^{2}_{\omega }\) | 0.22 | 0.38 | 0.0030 | 0.00 | 0.08 | 1.19 | |
Parameter | Estimate | SE | CI lower | CI upper | p-value | |||
REML | B: Quinolone | δ AB | -1.97 | 0.68 | -3.30 | -0.65 | <0.01 | |
estimation | C: Non-quinolone | δ AC | -1.40 | 0.81 | -2.98 | 0.18 | 0.08 | |
D: Antiseptic | δ AD | -0.66 | 0.72 | -2.07 | 0.75 | 0.36 | ||
Heterogeneity | \(\tau ^{2}_{\beta }\) | 0.10 | – | 0.00 | 1.67 | – | ||
Inconsistency | \(\tau ^{2}_{\omega }\) | 0.54 | – | 0.00 | 3.96 | – |