Table 3. The number and proportion of RCTs with low risk of bias by income, publication language, sample size and year of publication.
| Total (n = 1177) | Sequence generation | Allocation concealment | Blinding | Incomplete outcome | Reporting |
|---|---|---|---|---|---|
| 530 (45.0%) | 391 (33.2%) | 438 (37.2%) | 678 (57.6%) | 520 (44.2%) | |
| Country income levels | |||||
| High-income (n = 921) | 446 (48.4%) | 323 (35.1%) | 344 (37.4%) | 536 (58.2%) | 431 (46.8%) |
| Middle-income (n = 215) | 58 (27.0%) | 40 (18.6%) | 69 (32.1%) | 117 (54.4%) | 58 (27.0%) |
| Multiple income (n = 41) | 26 (63.4%) | 28 (68.3%) | 25 (61.0%) | 25 (61.0%) | 31 (75.6%) |
| Publication language of RCTs | |||||
| English (n = 1056) | 506 (47.9%) | 382 (36.1%) | 415 (39.3%) | 619 (58.6%) | 501 (47.4%) |
| Chinese (n = 92) | 14 (15.4%) | 3 (3.3%) | 15 (16.5%) | 43 (47.3%) | 8 (8.7%) |
| Other languages (n = 29) | 10 (34.5%) | 6 (20.7%) | 8 (27.6%) | 16 (55.2%) | 11 (37.9%) |
| Total sample size of RCTs | |||||
| <50 (n = 366) | 128 (35.0%) | 95 (26.0%) | 126 (34.4%) | 213 (58.2%) | 141 (38.5%) |
| 50–99 (n = 295) | 127 (43.1%) | 75 (25.4%) | 102 (34.6%) | 163 (53.9%) | 115 (39.0%) |
| 100–199 (n = 224) | 103 (46.0%) | 75 (33.5%) | 84 (37.5%) | 136 (60.7%) | 101 (45.1%) |
| 200–499 (n = 177) | 87 (49.2%) | 66 (37.3%) | 68 (38.4%) | 104 (58.8%) | 90 (50.9%) |
| ≥500 (n = 115) | 85 (73.9%) | 80 (69.6%) | 58 (50.4%) | 62 (53.9%) | 73 (63.5%) |
| Year of publication of RCTs | |||||
| Before 1980 (n = 35) | 11 (31.4%) | 12 (34.3%) | 13 (37.1%) | 19 (54.3%) | 17 (48.6%) |
| 1980–1989 (n = 133) | 39 (29.3%) | 24 (18.1%) | 40 (30.1%) | 50 (37.6%) | 40 (30.1%) |
| 1990–1999 (n = 214) | 81 (37.9%) | 51 (23.8%) | 75 (35.1%) | 103 (48.1%) | 94 (43.9%) |
| 2000–2009 (n = 626) | 301 (48.1%) | 231 (36.9%) | 243 (38.8%) | 402 (64.2%) | 281 (44.9%) |
| Since 2010 (n = 169) | 98 (58.0%) | 73 (43.2%) | 67 (39.6%) | 104 (61.5%) | 88 (52.1%) |