Number Needed to Treat

The number needed to treat (NNT) is defined as the number of patients that would need to be treated to prevent an adverse outcome in one additional patient compared with control treatment over a specified time period. The term was first introduced in 1988 by Andreas Laupacis, David Sackett, and Robin Roberts as the number needed to be treated but has been shortened to number needed to treat or its abbreviation NNT.

In medicine and other clinical practices, the well-performed randomized clinical trial is often considered the gold standard for judging the effectiveness of therapeutic interventions. The outcome measures from such studies are reported in many ways that involve both group and individual patient outcomes. An example of the former might be the mean reduction in blood pressure in the group of patients receiving one antihypertensive medication compared with the mean reduction in the group receiving another drug or a placebo. Such grouped data outcomes have often been the evidence for making clinical decisions. It makes sense, at first glance, that if all other factors (e.g., cost, access, side effects) are equal between the two antihypertensive medications, patients in the group with the average lower blood pressure would benefit from the medication compared with patients given the other drug. Missing from such analysis, however, is whether, in fact, individual patients benefited in clinically meaningful ways and how many individuals benefited. The researchers could have alternatively reported the results as the number or percentage of individuals in each medication group with outcome blood pressures in the normal range.

Largely through the development of the field of evidence-based medicine, there has been and continues to be an understanding of the unfavorable impact on the practice of medicine that can result not only from methodologically flawed clinical trials but also from the failure to report treatment effects in clinically relevant outcomes and in terms of individual patient responses.

In the current literature, when clinically relevant outcomes are reported for individual patients, randomized controlled trials and systematic reviews frequently report the treatment effect as relative risk (RR), relative risk reduction (RRR), absolute risk reduction (ARR), or the number needed to treat (NNT). This entry focuses on NNT and its clinical utility relative to the other measures, as well as the strengths and weakness of the NNT as a measure of clinical effect. Additionally, useful resources for calculating the NNT and its precision are provided.