NNT, number needed to treat: does it have any real value?

Clinical trials usually use the relative risk (rate ratio or hazard ratio) to compare the effects of one treatment modality with others. However, the numbers needed to treat/harm (NNT/NNH) are sometimes used as another way of presenting an estimate of the effect of a medical intervention, pointing at the number of patients needed to be exposed over a certain period of time in order to achieve one beneficial or adverse event. For clinicians and patients, this is a very simple and clear tool to demonstrate the consequences of a specific intervention. Epidemiologists and statisticians are more cautious with interpretations of data of that sort. This article brings the relevant perspectives of a clinician, an epidemiologist and a statistician in regard to the value of NNT/NNH.