Table 3 True and predicted BoS values for two Cochrane reviews, alongside their multivariate and univariate meta-analysis results
Review (total number of studies) | Outcome | Number of studies with the outcome | Predicted BoS from Eq. (1) | Observed BoS after bivariate meta-analysis | Bivariate common-effect meta-analysis | Univariate common-effect meta-analysis for each outcome separately | ||
|---|---|---|---|---|---|---|---|---|
Summary odds ratio (95% CI) | Standard error of log odds ratio | Summary odds ratio (95% CI) | Standard error of log odds ratio | |||||
Buzzetti et al. [30] (12) | 1: Mortality at 30 days | 9 | 20.1% | 47.7% | 0.89 (0.69 to 1.15) | 0.13 | 0.69 (0.48 to 0.98) | 0.18 |
2: Mortality at maximal follow-up | 12 | 10.8% | 3.2% | 0.89 (0.69 to 1.14) | 0.13 | 0.91 (0.70 to 1.17) | 0.13 | |
Feinberg et al. [31] (15) | 1: All-cause mortality at end of intervention | 11 | 21.0% | 32.8% | 0.97 (0.81 to 1.16) | 0.09 | 0.97 (0.78 to 1.20) | 0.11 |
2: All-cause mortality at maximum follow-up | 15 | 10.4% | 0.2% | 0.96 (0.82 to 1.12) | 0.08 | 0.96 (0.82 to 1.13) | 0.08 | |