Table 13.
Comparison of model fit for monotherapy network meta-analysis models
| Analysis | DIC | Average residual deviance†† | SD in treatment effects |
|---|---|---|---|
| ACR 20 monotherapy DMARD-experienced (fixed effects) | 280.685 | 1.531 | NA |
| ACR 20 monotherapy DMARD-experienced (random effects) | 265.259 | 1.012 | 0.817 |
| ACR 50 monotherapy DMARD-experienced (fixed effects) | 255.059 | 1.224 | NA |
| ACR 50 monotherapy DMARD-experienced (random effects) | 252.167 | 1.042 | 0.554 |
| ACR 70 monotherapy DMARD-experienced (fixed effects with cc) | 208.357 | 1.244 | NA |
| ACR 70 monotherapy DMARD-experienced (random effects with cc) | 205.674 | 1.045 | 0.761 |
| ACR 70 monotherapy DMARD-experienced (random effects, no cc) | DNC | ||
| ACR 70 monotherapy DMARD-experienced (fixed effects, no cc) | DNC |
Note:
††
Sum of the residual deviance divided by total number of arms.
Abbreviations: ACR 20/50/70, American College of Rheumatology (ACR) criteria scores of 20, 50, and 70; DMARD, disease-modifying antirheumatic drug; NA, not applicable to fixed-effect models; DNC, did not converge; cc, continuity correction; DIC, deviance information criterion; SD, standard deviation.