Number needed to treat

Depiction of the results of 2 hypothetical placebo controlled trials involving 100 patients. The number needed to treat (NNT) is sensitive to the control rate (ie, the frequency that events such as death occur at baseline). The NNT decreases as the number of events (such as deaths) in the control arm (placebo arm) increases. The relative risk in these studies (the death rate in those receiving drug divided by the death rate in those receiving placebo) is equal; both demonstrate a 50% reduction in the risk of death in patients who received the drug. However, the NNT (the reciprocal of the death rate in the placebo minus the death rate in the drug group) is different (5 versus 2.5). This difference reflects the higher death rate (ie, the control rate) in the second study.