Calculate NNT, ARR, and RRR from your clinical trial data
Try It FreeThe Number Needed to Treat (NNT) is the average number of patients who need to receive a treatment for one additional patient to benefit compared to the control group. NNT = 1 / ARR (absolute risk reduction). A lower NNT indicates a more effective treatment. NNT is considered one of the most clinically meaningful measures because it translates statistical significance into practical impact.
NNT is derived from the absolute risk reduction:
Where CER = control event rate and EER = experimental event rate
CrossTabs.com calculates all clinical risk measures from a 2×2 table:
NNT must always be interpreted in clinical context. An NNT of 10 for preventing death is very different from NNT of 10 for preventing a mild headache. Confidence intervals for NNT can be discontinuous when they cross infinity (when the ARR CI crosses zero), which CrossTabs.com handles correctly.
There is no universal threshold. It depends on the severity of the outcome and the side effects of treatment. For serious outcomes like mortality, NNT < 50 may be considered worthwhile. For mild conditions, NNT < 10 might be required. Always consider NNT alongside the NNH (number needed to harm) from side effects.
A negative NNT means the treatment increases the event rate (harm). This is reported as Number Needed to Harm (NNH). CrossTabs.com automatically distinguishes between NNT and NNH.
NNT CI = 1/upper ARR CI to 1/lower ARR CI. When the ARR CI crosses zero, the NNT CI is discontinuous (goes through infinity). CrossTabs.com computes these CIs using the Wald method.