Table 5.
Use of risk of bias and risk of bias domains in the Cochrane and non-Cochrane reviews that applied the Cochrane risk of bias tool for randomized clinical trials
| Use of risk of bias and risk of bias domains | 100 Cochrane reviews (100 %) | 31 non-Cochrane reviews (100 %)f | P value* |
|---|---|---|---|
| Use of risk of bias | |||
| Summarises risk of bias on an outcome levela | 12 (12 %) | 2 (6 %) | 0.73 |
| Unclear what level risk of bias was summarised onb | 88 (88 %) | 29 (94 %) | 0.88 |
| Use of risk of bias standardc domains | |||
| Review uses the 2011 tool version | 100 (100 %) | 26 (84 %) | 0.65 |
| Review uses all standardc domains | 59 (59 %) | 16 (52 %) | 0.73 |
| -Sequence generation | 100 (100 %) | 30 (97 %) | 1.00 |
| -Allocation concealment | 100 (100 %) | 30 (97 %) | 1.00 |
| -Blinding of patients and care providers | 62 (62 %) | 21 (68 %) | 0.87 |
| -Blinding of outcome assessors | 65 (65 %) | 20 (65 %) | 1.00 |
| -Incomplete outcome data | 99 (99 %) | 29 (94 %) | 0.88 |
| -Selective reporting | 87 (87 %) | 25 (81 %) | 0.88 |
| Merging and splitting of standardc domains | |||
| Review merges two standardc domains | 37 (37 %) | 8 (26 %) | 0.53 |
| -Merges risk of bias domains on an outcome leveld | 6 of 37 (16 %) | 0 of 8 (0 %) | 0.57 |
| -Does not merge risk of bias domains on an outcome level | 31 of 37 (84 %) | 8 of 8 (100 %) | 0.79 |
| Review splits a standardc domain into two or more domainse | 18 (18 %) | 7 (23 %) | 0.62 |
*P values were calculated with Fisher’s two-tailed exact test
aOne or more domains were separately assessed for more than one outcome or groups of outcomes (i.e. subjective versus objective outcomes)
bReview has a singular risk of bias assessment despite more than one outcome included in the review. No review based its risk of bias assessment on a singular or primary outcome
cThe six standard domains (not including the “other bias” domain) included in the Cochrane risk of bias tool for randomized clinical trials (i.e. the tool)
di.e. merges blinding of patients and care providers with blinding of outcome assessors into one blinding domain and evaluates blinding for subjective/objective or explicit (≥2) outcomes.
ei.e. splits blinding of patients and care providers into blinding of personnel and blinding of patients or splits incomplete outcome data into assessment of intention-to-treat and assessment of dropouts.
f31 of 100 non-Cochrane reviews used the Cochrane risk of bias tool for randomized clinical trials (i.e. the tool) and were compared to the 100 Cochrane reviews that used the tool for randomized clinical trials