5 Bias and confounding
Research into the eects of therapeutic interventions is bedevilled by the dual problems of bias and confounding. Bias refers to any systematic error in the design, conduct, analysis, publication or review of a study leading to conclusions that are systematically dierent from the truth. Confounding literally means confusion of eects. It occurs when there are other dierences between two (or more) groups that inuence the apparent outcomes of an intervention and are independent of the eects of the intervention itself. ese other factors are known as ‘confounders’ or ‘confounding factors’. For a factor to be a confounder it must dier between the comparison groups and be a predictor of the outcome of interest.