General characteristics and method of risk of bias assessment in individual systematic reviews
| Overall (n=200) | Cochrane (n=100) | Non-Cochrane (n=100) | Difference between proportions (95% CI) | |
|---|---|---|---|---|
| Common medical specialty | ||||
| Cardiology 20 (10%) | Neurology 16 (16%) | Cardiology 15 (15%) | ||
| Neurology 19 (9.5%) | Obs/Gynae 15 (15%) | Oncology 15 (15%) | ||
| Obs/Gynae 19 (9.5%) | Infectious diseases 7 (7%) | Endocrinology 9 (9%) | ||
| Oncology 18 (9%) | Musculoskeletal 7 (7%) | Psychiat/psychol 6 (6%) | ||
| Endocrinology 13 (6.5%) | Cardiology 5 (5%) | Surgery 6 (6%) | ||
| Type of intervention | ||||
| Drug | 109 (54.5%) | 55 (55%) | 54 (54%) | 1.0 (−12.8 to 14.8) |
| Surgery/procedure | 38 (19%) | 20 (20%) | 18 (18%) | 2.0 (−8.9 to 12.9) |
| Counselling/lifestyle | 41 (20.5%) | 17 (17%) | 24 (24%) | −7.0 (−18.1 to 4.1) |
| Equipment | 12 (6%) | 8 (8%) | 4 (4%) | 4.0 (−2.6 to 10.6) |
| Number of included studies | ||||
| Median (IQR) | 10 (5 to 23) | 6.5 (4 to 14) | 16.5 (7 to 33.5) | |
| Minimum/maximum | 1 to 385 | 1 to 102 | 3 to 385 | |
| Number of randomised trials | ||||
| Median (IQR) | 7 (4 to 17) | 6 (4 to 14) | 8 (5 to 22) | |
| Minimum/maximum | 1 to 104 | 1 to 86 | 1 to 104 | |
| Number of meta-analyses | ||||
| Median (IQR) | 4 (2 to 12) | 11 (3 to 22) | 3 (1 to 5) | |
| Minimum/maximum | 0 to 367 | 0 to 277 | 0 to 367 | |
| Method of assessing risk of bias* | ||||
| Single components | 128 (62%) | 92 (90%) | 36 (34%) | 55.9 (45.1 to 66.7) |
| Scale | 49 (24%) | 9 (9%) | 40 (38%) | −29.2 (−40.0 to −18.4) |
| Checklist | 9 (4%) | 1 (1%) | 8 (6%) | −6.6 (−12.0 to −1.2) |
| Not specified | 21 (10%) | 0 (0%) | 21 (20%) | −20.0 (−27.6 to −12.3) |
| Specific tool used† | ||||
| Cochrane RoB tool | 107 (51%) | 86 (82%) | 21 (20%) | 55.9 (45.1 to 66.7) |
| Modified Cochrane RoB tool | 15 (7%) | 9 (9%) | 6 (6%) | 2.8 (−4.2 to 9.8) |
| Jadad scale | 26 (12%) | 7 (7%) | 19 (18%) | −11.6 (−20.4 to −2.8) |
| Other | 34 (16%) | 3 (2%) | 31 (30%) | −26.9 (−36.3 to −17.6) |
| Not specified | 27 (13%) | 0 | 27 (26%) | −25.9 (−34.3 to −17.5) |
| Who did the assessment? | ||||
| One person | 5 (2.5%) | 1 (1%) | 4 (4%) | −3.0 (−7.3 to 1.3) |
| Two people | 167 (83.5%) | 98 (98%) | 69 (69%) | 29.0 (19.5 to 38.4) |
| Not reported | 28 (14%) | 1 (1%) | 27 (27%) | −26.0 (−34.9 to −17.1) |
| Assessment used as eligibility criteria | ||||
| Yes | 10 (5%) | 2 (2%) | 8 (8%) | −6.0 (−11.9 to 0.0) |
| No | 182 (91%) | 98 (98%) | 84 (84%) | 14.0 (6.3 to 21.7) |
| Not reported | 8 (4%) | 0 | 8 (8%) | −8.0 (−13.3 to −2.7) |
| Number of items assessed‡ | ||||
| Median (IQR) | 6 (5 to 7) | 6 (6 to 7) | 5 (3 to 7) | |
| Minimum/maximum | 1 to 27 | 3 to 15 | 1 to 27 | |
*Method of assessing risk of bias: seven reviews used more than one method; Cochrane n=2; non-Cochrane n=5.
†Type of tool used: nine reviews used more than one tool; Cochrane n=5; non-Cochrane n=4. Other tool n=34: Pedro n=8; Downs and Black n=4; own scale n=4; Schulz n=2; Maastricht criteria n=2; Critical Skills Appraisal Programme (CASP) n=1; Consolidated Standards for Reporting Trials (CONSORT) n=1; Centre for Reviews and Dessimination n=1; Detsky scale n=1; Grading of Recommendations Assessments, Development and Evaluation (GRADE) n=1; Heyland score n=1; Juni n=1; Modified Jadad scale n=1; PRISMA n=1.
‡Number of items assessed: nine reviews unclear number of items assessed; Cochrane n=0; non-Cochrane n=9.