Table 3 Discrepancies of treatment effects of Chinese- versus non-Chinese-language acupuncture trials stratified by ROB domains with high/unclear risk of bias
ROB domains | No. reviews | No. RCTs (Chinese- vs. non-Chinese-language) | Sample size (Chinese- vs. non-Chinese-language) | ROR* [95% CI] |
|---|---|---|---|---|
Random sequence generation (high/unclear risk) | 9 | 20 vs. 26 | 1524 vs. 3425 | 0.87 [0.36, 2.11] |
Allocation concealment (high/unclear risk) | 12 | 41 vs. 39 | 3419 vs. 4556 | 0.60 [0.32, 1.15] |
Blinding of participants and personnel (high/unclear risk) | 13 | 51 vs. 23 | 4509 vs. 2804 | 0.67 [0.37, 1.21] |
Blinding of outcome assessment (high/unclear risk) | 13 | 49 vs. 20 | 4422 vs. 1452 | 0.59 [0.29, 1.19] |
Incomplete outcome data (high/unclear risk) | 4 | 16 vs. 8 | 1660 vs. 711 | 1.26 [0.56, 2.84] |
Selective reporting (high/unclear risk) | 12 | 49 vs. 40 | 4339 vs. 4605 | 0.52 [0.27, 1.00] |