Table 2.
Analyzed databases | Detailed databasesa | Included studies | OR and CI | Overall Z test and p value | I 2 and CI (%) |
---|---|---|---|---|---|
Overall | All eligible papers | 15 | 1.86(1.65–2.10) | Z = 10.07, p < 0.001 | 89.8(84.8–93.1) |
All eligible papers (fixed-effects model) | 15 | 1.90(1.87–1.94) | Z = 73.67, p < 0.001 | 89.8(84.8–93.1) | |
Studies with the most variation in effect were dropped | Excluded max-effect paper | 14 | 1.81(1.61–2.03) | Z = 9.87, p < 0.001 | 89.6(84.3–93.1) |
Excluded min-effect paper | 14 | 1.92(1.71–2.17) | Z = 10.70, p < 0.001 | 89.5(84.1–93.0) | |
Excluded max- and min-effect papers | 13 | 1.87(1.67–2.10) | Z = 10.55, p < 0.001 | 89.2(83.4–93.0) | |
Studies with the most variation in sample size were dropped | Excluded max-sample size paper | 14 | 1.91(1.60–2.28) | Z = 7.22, p < 0.001 | 90.5(85.8–93.6) |
Excluded min-sample size paper | 14 | 1.82(1.61–2.05) | Z = 9.70, p < 0.001 | 90.1(85.2–93.4) | |
Excluded max- and min-sample size papers | 13 | 1.85(1.55–2.20) | Z = 6.76, p < 0.001 | 90.9(86.2–93.9) |
OR odds ratio, CI confidence interval, max maximum, min minimum
aRandom-effects models were used to combine the effects unless otherwise specified