Table 3.

Reported ROBINS-I risk of bias judgments in included systematic reviews that used the standard ROBINS-I scale

		AMSTAR 2 confidence rating
Characteristic	Overall, N = 93	High, N = 5	Moderate, N = 16	Low, N = 31	Critically low, N = 41
Mean (SD) proportion of studies rated
Low	0.10 (0.23)	0.02 (0.03)	0.10 (0.26)	0.08 (0.24)	0.12 (0.23)
Moderate	0.36 (0.33)	0.22 (0.40)	0.27 (0.25)	0.34 (0.32)	0.41 (0.35)
Serious	0.39 (0.33)	0.44 (0.44)	0.28 (0.25)	0.45 (0.31)	0.38 (0.35)
Critical	0.15 (0.29)	0.32 (0.41)	0.34 (0.42)	0.11 (0.23)	0.08 (0.23)
No information	0.01 (0.03)	0.00 (0.00)	0.00 (0.02)	0.01 (0.02)	0.01 (0.04)
Range of risk of bias categories used (reviews with >5 studies only)
One category (all studies rated the same)	9 (13%)	0 (0%)	0 (0%)	3 (12%)	6 (21%)
Two categories	30 (45%)	2 (50%)	4 (36%)	11 (46%)	13 (46%)
Three categories	27 (40%)	2 (50%)	7 (64%)	10 (42%)	8 (29%)
Four categories	1 (1%)	0 (0%)	0 (0%)	0 (0%)	1 (4%)
(5 or fewer studies)	26	1	5	7	13
Type of intervention studied
Clinical intervention (drug)	23 (25%)	0 (0%)	5 (31%)	11 (35%)	7 (17%)
Clinical intervention (other)	25 (27%)	0 (0%)	7 (44%)	7 (23%)	11 (27%)
Clinical intervention (surgical)	17 (18%)	0 (0%)	0 (0%)	6 (19%)	11 (27%)
Environmental exposure	3 (3%)	2 (40%)	0 (0%)	0 (0%)	1 (2%)
Non-clinical intervention	3 (3%)	0 (0%)	2 (12%)	1 (3%)	0 (0%)
Non-interventional	11 (12%)	1 (20%)	2 (12%)	1 (3%)	7 (17%)
Public health intervention	11 (12%)	2 (40%)	0 (0%)	5 (16%)	4 (10%)
Number of included studies
Median (IQR)	14 (9, 29)	14 (9, 27)	16 (11, 40)	15 (10, 27)	12 (8, 28)
Range	2, 124	3, 29	3, 91	4, 40	2, 124
Review included randomized trials	53 (57%)	1 (20%)	12 (75%)	17 (55%)	23 (56%)
Review included meta-analysis	58 (62%)	4 (80%)	6 (38%)	16 (52%)	32 (78%)

Excludes reviews not reporting risk of bias using the standard ROBINS-I categories (low, moderate, serious, critical, no information)