From: Association between trial registration and treatment effect estimates: a meta-epidemiological study
Characteristics | Registered N = 225 | Not registered N = 97 | |||
---|---|---|---|---|---|
Prospectively N = 74 | Retrospectively before PCDb N = 88 | Retrospectively after PCDb N = 54 | Not reported N = 9 | ||
Type of interventiona | |||||
Pharmacological (N = 265) | 66 (25) | 72 (27) | 46 (17) | 7 (3) | 74 (28) |
Non-pharmacological (N = 57) | 8 (14) | 16 (28) | 8 (14) | 2 (4) | 23 (40) |
Publication yeara | |||||
2006–2009 (N = 169) | 17 (10) | 53 (31) | 31 (18) | 2 (1) | 66 (39) |
2010–2014 (N = 153) | 57 (37) | 35 (23) | 23 (15) | 7 (5) | 31 (20) |
Start datea | |||||
Before July 2005 (N = 162) | 5 (3) | 59 (36) | 41 (25) | 6 (4) | 51 (31) |
In July 2005 or after (N = 133) | 69 (52) | 29 (22) | 11 (8) | 2 (1) | 22 (17) |
Not reported (N = 27) | 0 (0) | 0 (0) | 2 (7) | 1 (4) | 24 (89) |
Sample size | |||||
Median (Q1–Q3) | 242 (100–632) | 152 (66–439) | 153 (64–274) | 97 (63–103) | 87 (50–154) |
Fundinga | |||||
Non-profit (N = 154) | 36 (23) | 50 (32) | 29 (19) | 4 (3) | 35 (23) |
Industry (N = 104) | 35 (34) | 31 (30) | 19 (18) | 4 (4) | 15 (14) |
Both (N = 22) | 2 (9) | 6 (27) | 6 (27) | 0 (0) | 8 (37) |
Not reported (N = 42) | 1 (2) | 1 (2) | 0 (0) | 1 (2) | 39 (93) |
Risk of bias | |||||
Sequence generation (high/unclear) | 19 (26) | 17 (19) | 16 (30) | 5 (56) | 27 (28) |
Allocation concealment (high/unclear) | 25 (34) | 26 (29) | 21 (39) | 6 (67) | 47 (48) |
Incomplete outcome data (high/unclear) | 24 (34) | 35 (40) | 12 (22) | 1 (11) | 33 (34) |