The studies were assessed for:
selection bias (allocation concealment and baseline comparability of the treatment groups),
performance bias (treatment adjustments applied equally to both groups, groups treated the same apart from the intervention of interest),
exclusion bias (intention-to-treat analysis, similar drop-out rates between treatment groups), and
detection bias (blinding of the outcome assessors and similar methods used to assess the outcomes for each treatment group).
The studies were classified as A if all quality criteria were met, B if one or more criteria were only partially met, or C if one or more criteria were not met.
Two reviewers independently assessed validity and resolved any disagreements through reaching consensus.