Table 5 Review flow reporting, included studies, outcomes and analyses in the included systematic reviews (the HTA data were collected for this study; all other data are from Page et al. [1], i.e. the data collected for 300 systematic reviews for a single month in 2004 and 2014)
From: Nature and reporting characteristics of UK health technology assessment systematic reviews
Characteristic | 2004 n = 300 [1] | HTA 2004 n = 23 | HTA 2014 n = 30 | Cochrane 2014 n = 45 [1] | 2014 n = 300 [1] |
|---|---|---|---|---|---|
Review flow reporting | |||||
PRISMA-like flow diagram (and text) | - (7%) | 13 (57%) | 29 (97%) | 23 (51%) | 206 (69%) |
Text/tables or flowchart only | - (35%) | 8 (35%) | 1 (3%) | 5 (11%) | 20 (7%) |
Partially reported | - (33%) | 0 | 0 (0%) | 7 (16%) | 38 (13%) |
Not reported | - (31%) | 2 (9%) | 0 (0%) | 10 (22%) | 36 (12%) |
Reasons for exclusion of full text articles reported | |||||
PRISMA-like flow diagram and text/tables | - (48%) | 15 (65%) | 27 (91%) | 41 (91%) | 211 (70%) |
Partial (only reasons for some exclusions provided) | - (40%) | 5 (22%) | 1 (3%) | 4 (9%) | 28 (9%) |
Not reported | - (17%) | 3 (13%) | 2 (7%) | 0 (0%) | 61 (20%) |
Grey literature included in review | |||||
Yes | NR | 18 (78%) | 9 (30%)* | 8 (18%) | 26 (9%) |
Total number of included participants reported | |||||
In main text | NR | 2/23 (9%) | 4/28 (14%) | 39/42 (93%) | 194/296 (66%) |
In abstract | NR | 1/23 (4%) | 4/28 (14%) | 37/42 (88%) | 147/296 (50%) |
At least one outcome reported in Methods | NR | 21/23 (96%) | 30 (100%) | 45 (100%) | 234 (78%) |
A specified primary outcome | - (51%) | 6 (26%) | 16 (53%) | 43(96%) | 136/288 (47%) |
Meta-analysis | |||||
No (includes empty reviews or reviews with a statement that studies could not be combined) | NR | 12 (52%) | 15 (50%) | 13 (29%) | 111 (37%) |
Yes | NR | 11 (47%) | 15 (50%) | 32 (71%) | 189 (63%) |
Statistical heterogeneity investigated | |||||
Quantitative | - (91%) | 9/11 (82%) | 13 (87%) | 32/32 (100%) | 175/189 (93%) |
Not reported | NR | 2/11 (18%) | 1/15 (7%) | 0 (0%) | 12 (7%) |
Risk of bias incorporated in meta-analysis | NR | 0/23 (0%) | 1/15 (7%) | 4/32 (13%) | 31/189 (16%) |
Publication bias | |||||
Formally assessed (Funnel plot etc.) | NR | 0/23 (0%) | 6/30 (20%) | 7 (16%) | 93 (31%) |
Not assessed, but planned if sufficient studies | NR | 1/23 (4%) | 2/30 (7%) | 28 (62%) | 37 (12%) |
Possibility discussed/considered in results, discussion etc. | - (31%) | 0/23 (0%) | 6/30 (20%) | 29 (64%) | 141 (47%) |
GRADE assessment reported | NR | 0 (0%) | 0 (0%) | 27 (60%) | 32 (11%) |
Limitations reported in Discussion/Conclusion | |||||
At study level and review level | NR | 11 (49%) | 19 (63%) | 32 (71%) | 173 (58%) |
Study-level only | NR | 10 (44%) | 6 (20%) | 10 (22%) | 67 (22%) |
Review-level only | NR | 0 | 3 (10%) | 0 (0%) | 27 (9%) |
Not reported | NR | 2 (7%) | 2 (7%) | ||
Risk of bias / quality assessment reported | |||||
In Abstract (Therapeutic reviews only) | NR | 7/16 (44%) [63%‡] | 8/14a (57%) [79%b] | 42 (93%) | 99/164 (60%) |