Table 2 Characteristics of the synthesis methods
Characteristic | Studies |
|---|---|
n (%) | |
Primary studies included (median [IQR]) | 10 [7–17] |
Number of RCTs included (median [IQR]) | 3 [1–6] |
Number of OSs included (median [IQR]) | 6.5 [3–12] |
Type of outcome analysed | |
Primary analysis | 113 (86) |
Secondary analysis | 12 (9) |
Efficacy outcome | 84 (64) |
Safety outcome | 28 (21) |
Type of data pooled | |
Adjusted | 7 (5) |
Not adjusted | 5 (4) |
Both adjusted and non-adjusted | 7 (5) |
Other | 5 (4) |
Not reported | 108 (82) |
Synthesis model | |
Random effects model | 73 (55) |
Common effect model | 3 (2) |
Both models | 53 (40) |
Other | 1 (1) |
Not reported | 2 (2) |
Types of effects reported | |
Effect estimates for each primary study | 121 (92) |
Meta-analysis summary effect | 77 (58) |
Effects stratified by study design | 44 (34) |
Statistical heterogeneity evaluation | N = 132 |
I2 | 123 (93) |
Q test | 25 (19) |
Tau | 32 (24) |
Statistical heterogeneity exploration | N = 105 |
Sensitivity analyses | 71 (68) |
Subgroup analyses | 75 (70) |
Meta-regression | 23 (22) |
Source of heterogeneity studied by the meta-analyses’ authors | N = 105a |
Study design | 283 (79) |
Risk of bias of the primary studies | 47 (45) |
Study population characteristics | 37 (35) |
Other confounding factors | 11 (10) |
Outcome assessment | 24 (23) |
Publication bias and small study effect evaluation | |
Graphical evaluation | 85 (64) |
Statistical evaluation | 66 (50) |
No evaluation | 40 (30) |