BiostatKit / Risk & odds / NNT

NNT Calculator

Enter the number of events in a treated and a control group and get the number needed to treat (or harm), the absolute risk reduction and a 95% confidence interval, plus a sentence for your manuscript.

Your data never leaves your browser. No sign-up, no ads.

Result

Number needed to treat, treated vs control.

Sentence for your manuscript

Written in AMA style from your numbers. Check it against your own study design before you submit.

Reporting checklist

CONSORT 2010 (trials): for a binary outcome give both relative and absolute effect sizes (item 17b) and the precision of each (item 17a). Report the follow-up period next to the NNT.
STROBE 2007 (cohort studies): state the statistical methods (item 12) and the number of outcome events per group (item 15).

Methods wording. The absolute risk reduction and its 95% CI (Newcombe hybrid score method) were computed with BiostatKit [1]; the number needed to treat is its reciprocal, rounded up, with limits from inverting the CI of the absolute risk reduction [2].
[1] BiostatKit. NNT calculator. Accessed [date]. [URL]
[2] Altman DG. Confidence intervals for the number needed to treat. BMJ. 1998;317(7168):1309-1312.

Reproduce it in R

The script is filled in with your numbers. Paste it into R or RStudio and compare.

R 4.x · base R only

This calculator was checked against SciPy and statsmodels on 2 reference cases and matched to 4 significant digits. See the verification log.

How to read a number needed to treat

What it is

The NNT is the number of people you must treat for one extra person to avoid the outcome during the same follow-up period. A smaller NNT means a stronger absolute effect. If treatment raises risk, the same number is the number needed to harm (NNH).

How this page computes it

ARR = risk(control) - risk(treated) and NNT = 1 / ARR (ARR as a proportion, for example 0.075, not 7.5).

The 95% CI of the ARR uses the Newcombe hybrid score method. The NNT limits are the reciprocals of the ARR limits (Altman). When the ARR interval includes 0, the NNT interval runs through infinity and is written as NNTB (benefit) x to infinity to NNTH (harm) y.

Assumptions

A binary outcome and a fixed follow-up period that is the same in both groups.
Independent participants; for a trial, randomised allocation.
The NNT applies to the baseline risk of the control group in your data.

Common mistakes

Quoting an NNT with no time frame. An NNT of 14 over 1 year is not an NNT of 14 over 5 years.
Rounding down. Round up to the next whole person.
Reporting an NNT with no CI, or a CI that excludes infinity when the ARR interval includes 0.
Applying the NNT to patients whose baseline risk differs from the control group.

Frequently asked questions

How do I interpret an NNT of 14?

On average, 14 people need to be treated for the follow-up period for one more person to avoid the outcome compared with control. Always state the period and the outcome.

What does an NNT confidence interval running through infinity mean?

The CI of the absolute risk reduction includes 0, so the data are compatible with no effect, with benefit and with harm. Write it as NNTB to infinity to NNTH, as in the result panel.

Can I get an NNT from an odds ratio or hazard ratio?

Not from this page. It needs the event counts in both groups. Converting an OR or HR needs an assumed baseline risk and gives a different, model-based number.

Is my data stored or sent anywhere?

No. All calculations run in your browser tab. Nothing is uploaded, and the page works offline once loaded.

Related calculators: Relative risk calculator · Odds ratio calculator