Absolute risk reduction

Absolute risk reduction measures benefit on the scale that actually matters to people: the real drop in the number of events. You calculate it by subtracting the treated group's event rate from the control group's event rate. Unlike relative measures, it shrinks when the outcome is rare, which is exactly why it resists the exaggeration that relative figures invite. When reading a study, look for both groups' event rates and do the subtraction yourself if the paper only reports a relative reduction. For example, if 4 percent of a placebo group and 3 percent of a treated group have a stroke, the absolute risk reduction is 1 percentage point, even though the relative reduction is a more impressive-sounding 25 percent. A related term, absolute risk increase, applies the same subtraction to harms. Absolute risk reduction also lets you compute the number needed to treat, which many readers find more intuitive.