Reading the Numbers

The authors of this review stated that the studies that were included had satisfactorily minimized the four major types of bias that lead to erroneous conclusions: selection bias, performance bias, attrition bias, and detection bias. Selection bias is a systematic difference between the participants in the groups that are being compared. It is prevented by proper randomization and concealment of allocation. Performance bias is a systematic difference in the care—apart from the intervention that is being studied—provided to the groups. It is prevented by blinding patients and providers (double blinding). Attrition bias is a systematic difference between groups with regard to the participants who withdraw. It is minimized by intention-to-treat analyses, high rates of follow-up, and by examining the reasons for withdrawal between comparison groups. Detection bias is a systematic difference in outcome assessment and is prevented by blinding the study personnel who perform the outcome measurements.