Table 3.

The use of methods to reduce the risk of selection bias in trials that stratified by site of recruitment

	Trials (n = 67)
Recruiters blinded? – number (%)
No	1 (1)
Yes	1 (1)
Not stated	65 (97)
Randomisation method – number (%)
Stratified permuted block	39 (58)
Minimisation	13 (19)
Other	3 (4)
Not stated	12 (18)
Block sizes stated? – number (%)
No	16/39 (41)
Yes	23/39 (59)
Were block sizes random? - number (%)
No	16/39 (41)
Yes	6/39 (15)
Not stated	17/39 (44)
Largest block size used – median (IQR)	8 (4 to 11)
Was minimisation deterministic? – number (%)
No	1/13 (8)
Yes	0/13 (0)
Not stated	12/13 (92)
Balanced for prognostic factors? – number (%)
No	27 (40)
Yes	40 (60)
Number of prognostic factors balanced – median (IQR)	2 (1 to 3)

IQR, interquartile range