Number Needed to Treat and Number Needed to Harm
The number needed to treat (NNT) and the number needed to harm (NNH) are two useful metrics for understanding an intervention’s clinical potential (Cook & Sackett, 1995; Zapletal, LeMaitre, Menard, & Degoulet, 1996). NNT refers to the number of patients or participants who would have to receive a treatment for a particular benefit to occur or to prevent a particular negative outcome such as a death; conversely, NNH refers to the number of patients who would have to receive a treatment for a particular harmful outcome to occur. NNT is a simple measure of the impact of an intervention on one person. For example, if an intervention has an NNT of 10, it means that 10 persons would have to receive the treatment for one person to benefit or to prevent one additional negative outcome. As not all persons will benefit from a treatment and some may be harmed or not affected, the NNT offers a measurement of how many persons are needed for each scenario. To calculate NNT, the absolute risk reduction (ARR) needs to be determined; the NNT is the inverse of ARR (NNT = 1/ARR). NNT and NNH are associated with effect sizes. A small effect size may reflect statistical but not clinical significance. In order for an intervention to be worth implementing, a trial must yield more than a trivial effect to also reflect a clinically meaningful difference. NNH is an important metric particularly for high-risk interventions. As most behavioral interventions report adverse events and are of low risk, NNH may not be as relevant as NNT.