Table 1 —

Assessment of internal validity based on Cochrane Collaboration Criteria (20)

Type of biasDefinition
Selection biasSystematic differences in control and intervention groups at baseline To avoid requires randomization and no significant differences between baseline variables in control and intervention groups, or adequate statistical consideration of potential confounders if baseline differences exist
Performance biasSystematic differences in care provided to control and intervention groups, apart from the intervention being evaluated To avoid requires no evidence of contamination or cointervention, including no additional contacts with researcher or providers for the intervention group compared with the control group
Attrition biasSystematic differences between study groups in withdrawals from the study To avoid requires attrition <20% of total n, or dropouts must resemble completers in baseline characteristics
Detection biasSystematic differences in outcomes assessment between study groups To avoid requires blinding for any outcome subject to assessor interpretation