Table 3

Risk of bias in randomized controlled trials
Overall, n (%) N=67
Sequence generation
Adequate 57 (85)
Not clear 10 (15)
Inadequate 0 (0)
Allocation concealment
Adequate 41 (61)
Inadequate 3 (5)
Not clear 23 (34)
Participants blinded
Definitely yes 5 (8)
Probably yes 21 (31)
Probably no 13 (19)
Definitely no 28 (42)
Caregiver blinded
Definitely yes 5 (8)
Probably yes 22 (33)
Probably no 13 (19)
Definitely no 27 (40)
Data collector blinded
Definitely yes 3 (5)
Probably yes 26 (39)
Probably no 13 (19)
Definitely no 25 (37)
Outcome assessor blinded
Definitely yes 34 (51)
Probably yes 16 (24)
Probably no 7 (10)
Definitely no 10 (15)
Data analyst blinded
Definitely yes 7 (10)
Probably yes 3 (5)
Probably no 37 (55)
Definitely no 20 (30)
Intention to treat
Yes 38 (57)
Unclear 15 (22)
No 14 (21)
Free of selective reporting of
outcomes
Adequate 59 (88)
Not clear 5 (7)
Inadequate 3 (5)
Stopped early
Total 7 (10)
For benefit 2 (3)
For harm 1 (2)
For insufficient accrual 4 (6)
Not reported 0 (0)
Follow up
Mean percentage (range) 96% (57-100)

Rada et al.

Rada et al. BMC Cancer 2013 13:76   doi:10.1186/1471-2407-13-76

Open Data