Methods characteristics in Bayesian MTCs
| Characteristic | n/N (%) |
|---|---|
| Conducted traditional meta-analysis | 26/34 (76.5) |
| Model | |
| Fixed effects | 1/34 (2.9) |
| Random effects | 20/34 (58.8) |
| Fixed and random effects | 7/34 (20.6) |
| Not reported | 6/34 (17.6) |
| Adjustment for covariates | 9/34 (25.6) |
| Adjustment for multiple arms in MTCs including trials with three or more arms | 10/28 (35.7) |
| Model fit tested | 15/34 (44.1) |
| Residual deviance | 6/15 (40.0) |
| Deviance information criterion | 2/15 (13.3) |
| Residual deviance and deviance information criterion | 3/15 (20.0) |
| Q–Q plots | 1/15 (6.7) |
| Mean sum deviation | 1/15 (6.7) |
| Method not reported | 2/15 (13.3) |
| Code published | 7/34 (20.6) |
| Online supplement | 5/7 (71.4) |
| External website | 2/7 (28.6) |
| Aggregate study-level data published | 21/34 (61.8) |
| Manuscript | 18/21 (85.7) |
| Online supplement | 2/21 (9.5) |
| External website | 1/21 (4.8) |
| Evaluation of convergence* | 12/34 (35.3) |
| Gelman Rubin statistic | 7/12 (58.3) |
| Kernel density plot | 1/12 (8.3) |
| Visual plot inspection | 1/12 (8.3) |
| Observation of chain mix | 2/12 (16.7) |
| Method not reported | 2/12 (16.7) |
| Priors | |
| Use of non-informative | 15/34 (44.1) |
| Use of informative priors | 3/34 (8.8) |
| Not specified | 16/34 (47.1) |
| Prior distribution of d reported | 11/34 (32.4) |
| Prior distribution for σ reported | 10/34 (29.4) |
| Sensitivity analysis based on priors | 4/34 (11.8) |
| Evaluation of heterogeneity in traditional meta-analysis* | 16/26 (61.5) |
| I2 | 13/16 (81.3) |
| Cochrane-Q statistic | 7/16 (43.8) |
| PICO statement | 1/16 (6.3) |
| Plot visualisation | 2/16 (12.5) |
| L'Abbe plot | 1/16 (6.3) |
| Evaluation of heterogeneity in network meta-analysis* | 11/34 (32.4) |
| Precision (τ2) | 6/11 (54.5) |
| Between study SD | 5/11 (45.5) |
| Heterogeneity p values | 1/11 (9.1) |
| Evaluation of inconsistency* | 24/34 (70.6) |
| Comparison to traditional or prior meta-analysis† | 12/24 (50.0) |
| Inconsistency/incoherence factors | 4/12 (33.3) |
| Posterior mean residual deviance | 3/12 (25.0) |
| Method not reported | 4/12 (33.3) |
| Trial sequential analysis | 1/12 (8.3) |
| Overall inconsistency (σ2w) | 1/12 (8.3) |
*Studies that used multiple methods to test heterogeneity were counted multiple times, in the respective categories.
†Authors either compared results of the MTC with a traditional meta-analysis that they conducted concurrently or with a traditional meta-analysis that was previously published.
MTC, mixed treatment comparison; PICO, patient, intervention, comparator, outcome.