TABLE 1

Summary of Reviews Documenting Risk of Bias by Domain Using Cochrane Collaboration Risk of Bias Tool

ReferenceHartling et al7Hamm et al8Crocetti et al9
Publication periodLate 1990s2007July 2007–June 2008a
N trials163300146
Risk of bias
Low, %48N/A
High, %3759N/A
Unclear, %5933N/A
Sequence generation,b%685141
Allocation concealment,b%687557
Blinding,b%405019
 Incomplete data,b%473811
 Selective outcome reporting,b%32182
 Other sources of bias,b%61662
Effect size
 Low bias0.230.16N/A
 High/unclear bias0.5c0.28dN/A
  • N/A, not available.

  • a Inclusion criteria: pediatric randomized controlled trials published in 8 high-impact journals.

  • b Percentage at high or unclear risk of bias.

  • c Significantly increased effect size in studies with high risk of bias.

  • d Trend toward increased effect size in studies with high risk of bias.