NNT

Number Needed to Treat (NNT) is a statistic used in evidence-based medicine to express the number of patients who need to be treated with a given therapy for one additional person to benefit compared with a control over a specified time period. It is the reciprocal of the absolute risk reduction (ARR).

ARR = control event rate minus experimental event rate. If a trial reports 10% events in control and 6% in treatment, ARR = 0.04, and NNT = 1 / 0.04 = 25. This means treating 25 patients prevents one additional event, on average, over the time horizon of the study. For harms, the analogous measure is Number Needed to Harm (NNH).

Important considerations: NNT depends on the chosen time frame and the baseline risk of the population studied; different populations or follow-up durations can yield different NNT values. It also depends on how the outcome is defined (e.g., all-cause mortality vs. a composite endpoint). Confidence intervals for NNT can be wide and are awkward to interpret because NNT is the reciprocal of a risk difference; standard methods use the CI for ARR to derive CIs for NNT.

Limitations: NNT conveys average benefit and can obscure variation among individuals; it should not be used to compare therapies across different outcomes or different time horizons without care. It should be considered alongside other measures such as relative risk, ARR, baseline risk, patient preferences, and safety data.