Table 2 Summary of assessment of bias
Variable | Level | Trials with individual-level randomisation (n) | Trials with cluster randomisation (n) |
|---|---|---|---|
Jadad score (possible range of 0–5; higher scores indicate lower likelihood of bias) | 0 | 2 (1.4%) | 1 (1.1%) |
1 | 26 (18.8%) | 22 (23.9%) | |
2 | 44 (31.9%) | 32 (34.8%) | |
3 | 66 (47.8%) | 37 (40.2%) | |
Allocation sequence adequately generated | Low risk | 75 (53.2%) | 44 (47.3%) |
High risk | 3 (2.1%) | 1 (1.1%) | |
Unclear | 63 (44.7%) | 48 (51.6%) | |
Allocation sequence adequately concealed | Low risk | 30 (21.3%) | 18 (19.3%) |
High risk | 3 (2.1%) | 1 (1.1%) | |
Unclear | 108 (76.6%) | 74 (79.6%) | |
Randomisation after consent obtained | Low risk | 116 (82.3%) | 23 (24.7%) |
High risk | 6 (4.3%) | 39 (41.9%) | |
Unclear / NA | 19 (13.5%) | 31 (33.3%) | |
Randomisation after baseline measures were completed | Low risk | 56 (39.7%) | 15 (16.1%) |
High risk | 14 (9.9%) | 41 (44.1%) | |
Unclear | 71 (50.4) | 37 (39.8%) | |
Baseline outcome measurements similar across trial arms | Low risk | 109 (77.3%) | 70 (75.3%) |
High risk | 9 (6.4%) | 11 (11.8%) | |
Unclear / NA | 23 (16.3%) | 12 (12.9%) | |
Baseline demographic characteristics similar across trial arms | Low risk | 121 (85.8%) | 69 (74.2%) |
High risk | 3 (2.1%) | 18 (19.4%) | |
Unclear / NA | 17 (12.1%) | 6 (6.5%) | |
Knowledge of allocation adequately concealed | Low risk | 0 (0%) | 0 (0%) |
High risk | 141 (100%) | 93 (100%) | |
Blinded outcome assessment | Low risk | 63 (44.7%) | 29 (31.2%) |
High risk | 8 (5.7%) | 17 (18.3%) | |
Unclear / NA | 70 (49.6%) | 47 (50.5%) | |
Incomplete outcome data | Low risk | 35 (24.8%) | 22 (23.7%) |
High risk | 82 (58.2%) | 57 (61.3) | |
Unclear / NA | 24 (17.0%) | 14 (15.1%) | |
Similar attrition between trial arms | Low risk | 85 (60.3%) | 51 (54.8%) |
High risk | 30 (21.3%) | 25 (26.9%) | |
Unclear / NA | 26 (18.4%) | 17 (18.3%) |