Free Tools › Number Needed to Treat Calculator
Work out the number needed to treat or harm from trial data, or convert a risk ratio or odds ratio into an NNT for a given baseline risk.
FreeNo account neededYour data stays in your browser
| Had the event | No event | |
|---|---|---|
| Treatment | ||
| Control |
| Event rate, treatment (EER) | 15.0% |
|---|---|
| Event rate, control (CER) | 30.0% |
| Absolute risk reduction | 15.0% (95% CI 3.6% to 26.4%) |
| Number needed to treat (NNTB) | 6.7 |
| 95% CI for NNT | 3.8 to 27.7Altman (1998) method |
For harmful events (e.g. death), an NNT for benefit means fewer events with treatment.
A 2×2 table, or a risk ratio or odds ratio with an assumed baseline risk.
Events in each group, or the ratio and baseline risk.
With the absolute risk reduction and, for 2×2 data, a confidence interval.
The number needed to treat (NNT) is the expected number of people who need to receive the intervention rather than the comparator for one extra person to benefit. When treatment increases an event, the same calculation gives the number needed to harm (NNH).
Because the NNT depends on baseline risk, the Cochrane Handbook recommends expressing relative effects as absolute effects across a range of assumed comparator risks.
It is conventionally rounded up to the next whole number when reported, because you cannot treat part of a person.
When the risk difference could be zero, the NNT could be infinitely large, so the interval runs from a benefit through infinity to a harm.