Number Needed to Treat (NNT) and Harm (NNH) Calculator
Result
Calculate NNT or NNH from comparable control and treatment event rates (0–1); handles zero risk difference and undefined RRR safely.
Formula
- Enter the event proportion in the control group and the event proportion in the treatment group, both between 0 and 1 for the SAME outcome and follow-up interval. If treatment lowers the event rate, NNT is the reciprocal of the absolute risk reduction; if treatment raises it, the tool labels the reciprocal of the absolute risk increase as NNH. Whole-person NNT/NNH is conventionally rounded up. With equal event rates, no finite NNT/NNH exists. If the control event rate is zero, relative risk reduction is undefined. The numbers alone cannot establish whether a treatment works, causes harm, or is appropriate for an individual; study design, confidence intervals and baseline risk matter.
- The main inputs are Control event rate, Experimental event rate. Enter them in the units and format shown in the calculator.
- Use the measurement units shown beside each field. The result is an estimate and should be interpreted in context.
- Changing any input above recalculates the result with the same calculation engine used by this page.
Example: Number Needed to Treat (NNT) and Harm (NNH)
- Control event rate: 0.2
- Experimental event rate: 0.1
Using Control event rate = 0.2, Experimental event rate = 0.1, the calculated result is ARR: 0.1 | NNT: 10 | RRR: 50.
Sample result: Calculating…
Frequently asked questions
What does the Number Needed to Treat (NNT) and Harm (NNH) calculate?
Calculate NNT or NNH from comparable control and treatment event rates (0–1); handles zero risk difference and undefined RRR safely.
Which values should I enter?
Enter Control event rate, Experimental event rate using the units and format shown beside each field.
What formula does this calculator use?
The calculator uses the following relationship: Absolute risk reduction (ARR) = control event rate (CER) − treatment event rate (TER). For ARR > 0: NNT = 1/ARR, conventionally rounded UP to the next whole person. For ARR < 0: the absolute risk INCREASE is TER − CER and NNH = 1/(TER − CER), rounded up. When ARR = 0, NNT/NNH is not finite (there is no observed absolute risk difference). Relative risk reduction (%) = 100 × (CER − TER)/CER only when CER > 0; when CER = 0 it is undefined. Both rates must lie in [0, 1], represent comparable groups and outcome periods, and describe the same event definition. This is arithmetic based on supplied event rates, not proof of efficacy, causal effects, treatment safety, significance, or patient-level clinical guidance..
Why might my result differ from another source?
Results can differ because measurement methods, rounding and assumptions vary. This calculator is an estimate and is not a medical diagnosis.
Does the result update automatically?
Yes. Changing an input updates the result using the same calculation engine shown on this page.