Table 2.
Systematic review methods of therapeutic meta-analyses published in 2013 and including non-randomized studies (n = 188)
| Characteristic | Included NRSI only | Included NRSI&RCTs | All Meta-analyses |
|---|---|---|---|
| (n = 69) | (n = 119) | (n = 188) | |
| No (%) | No (%) | No (%) | |
| Search strategy | |||
| No. of electronic databases searched, median [IQR] | 3 [2–4] | 3 [3, 4] | 3 [3, 4] |
| > 2 databases searched | 51 (74) | 96 (81) | 147 (78) |
| Electronic databases searcheda | |||
| Medline/PubMed/Ovid | 69 (100) | 118 (99) | 187 (99) |
| EMBASE | 52 (75) | 97 (82) | 149 (79) |
| Cochrane | 45 (65) | 81 (68) | 126 (67) |
| Web of science/web of knowledge | 14 (20) | 22 (18) | 36 (19) |
| Time-frame of search strategy reported | 62 (90) | 109 (92) | 171 (91) |
| Time end of search-publication date (months), median [min-max] | 10 [0–24] | 9 [1–51] | 10 [0–51] |
| Search strategy provided for at least one database | 18 (26) | 44 (37) | 62 (33) |
| Searched reference lists for relevant studies | 57 (83) | 105 (88) | 162 (86) |
| Hand searched journals | 5 (7) | 7 (6) | 12 (6) |
| Searched for grey literaturea | 23 (33) | 49 (41) | 72 (38) |
| Conference abstracts | 14 (20) | 27 (23) | 41 (22) |
| Registries (eg, ClinicalTrials.gov) | 8 (12) | 25 (21) | 33 (18) |
| Experts | 6 (9) | 9 (8) | 15 (8) |
| No restriction by language reported | 20 (29) | 62 (52) | 82 (44) |
| Study selection | |||
| Duplicate study selection | 47 (68) | 84 (71) | 131 (70) |
| Methodological quality/risk of bias | |||
| Assessment of MQ/RoB | 47 (68) | 88 (74) | 135 (72) |
| Tools used for RCTsa | |||
| Cochrane RoB Tool [41] | NA | 42 (35) | NA |
| Jadad scale [42] | NA | 22 (18) | NA |
| GRADE [43] | NA | 9 (8) | NA |
| Other | NA | 21 (18) | NA |
| Authors made their own tool | NA | 3 (3) | NA |
| Unclear | NA | 7 (6) | NA |
| Tools used for NRSIa | |||
| Newcastle-Ottawa scale [44] | 25 (36) | 43 (36) | 68 (36) |
| GRADE [43] | 5 (7) | 8 (7) | 13 (7) |
| Cochrane RoB Tool [41] | 3 (4) | 7 (6) | 10 (5) |
| Other tools | 13 (19) | 24 (20) | 37 (20) |
| Authors made their own tool | 6 (9) | 6 (5) | 12 (6) |
| Unclear | 3 (4) | 9 (8) | 12 (6) |
| Confounding taken into account in risk of bias | 13 (19) | 20 (17) | 33 (18) |
| Same tool used for RCTs and NRSI | NA | 27 (23) | NA |
| Duplicate MQ/RoB assessment | 30/47 (64) | 57/88 (65) | 87/135 (64) |
| Data extraction | |||
| Duplicate data extraction | 52 (75) | 86 (72) | 138 (73) |
| Contact authors for clarification or additional results | 18 (26) | 43 (36) | 61 (32) |
Data are no. (%), unless indicated. IQR interquartile range; MQ methodological quality; NRSI non-randomized studies of interventions; RCTs randomized controlled trials; Cochrane RoB Cochrane risk of bias tool; NA Not applicable
aMore than one item could have been reported for each meta-analysis