Reference
|
Study design*
|
Randomization
|
Blinding
|
Selection bias
|
Measurement bias
|
Confounding
|
Overall quality based on score average**
|
---|
Real world settings
|
---|
Elbel et al. (2011) [17]
|
1
|
NA
|
0
|
1
|
2
|
0
|
Fair
|
Elbel et al. (2009) [18]
|
1
|
NA
|
0
|
1
|
2
|
0
|
Fair
|
Finkelstein et al. (2011)[19]
|
1
|
NA
|
2
|
2
|
2
|
1
|
Good
|
Chu et al. (2009) [6]
|
0
|
NA
|
2
|
2
|
2
|
1
|
Fair
|
Dumanovsky et al. (2011) [16]
|
0
|
NA
|
0
|
1
|
2
|
0
|
Fair
|
Laboratory settings
|
Harnack et al. (2008) [20]
|
2
|
2
|
1
|
2
|
2
|
1
|
Good
|
Roberto et al. (2010) [21]
|
2
|
2
|
0
|
2
|
2
|
1
|
Fair
|
- NA, not applicable
- * Study design: Randomized controlled trials received a score of 2, cohort studies with comparison groups received a score of 1, and single group studies (with no comparison group) received a score of 0
- **Quality scoring: 2 for good (or low potential for bias), 1 for fair (or moderate potential for bias), 0 for poor (or high potential for bias). To achieve a final quality rating of good, studies had to have an average score greater than or equal to 1.5 and could not receive scores of 0 for any individual variable. An average score less than or equal to 0.5 would have been considered to be a study with poor quality. Scores in between were considered fair quality studies.