Number Needed to Treat Calculator

Find the number needed to treat.

Result

16.67
6.00%

How it is calculated

Enter the event rate in the control group (CER) and in the treated group (EER), both as percentages between 0 and 100. The tool computes the absolute risk reduction ARR = (CER − EER) ÷ 100 and then NNT = 1 ÷ ARR, showing NNT to two decimals and ARR in percent. The event is whatever outcome the trial counted, such as a stroke or relapse. Only a treatment that lowers the rate gives a result: if EER is equal to or higher than CER there is no benefit to express, so the tool returns nothing, and values outside 0–100 are rejected. NNT is not rounded up to a whole patient here, and it is tied to the follow-up time of the trial it came from; it carries no confidence interval, so small trials give a very uncertain figure.

Worked example

In a trial, 15% of the control group had the event and 9% of the treated group did. The absolute risk reduction is 15 − 9 = 6.00%, or 0.06 as a fraction. The number needed to treat is 1 ÷ 0.06 = 16.67. In practice this reads as: about 17 patients must be treated over the trial's follow-up period for one of them to avoid the event who would otherwise have had it.

Try this example →

Common mistakes

  • Entering the rates as fractions (0.15 and 0.09). They are percentages, and 0.15 would be read as 0.15%, giving an NNT of 1,666.67.
  • Entering the relative risk reduction, or the percentage who responded, instead of the event rates in each group.
  • Swapping the groups. The control rate goes first; if the treated group is higher, the tool returns no result because the treatment did not reduce risk.
  • Comparing NNTs from trials with different follow-up periods or different outcomes as if they were interchangeable.
  • Quoting the figure as exact. With a small trial the uncertainty is large, and 16.67 is a point estimate that is normally rounded up to 17 in clinical talk.

Sources

About this calculator

The number needed to treat (NNT) is a clinical measure of treatment effect, calculated as NNT = 1/ARR, where ARR (absolute risk reduction) is the difference between the event rate in the control group and the event rate in the treatment group from a clinical trial. NNT tells you, on average, how many patients need to receive a treatment for one additional patient to benefit — an NNT of 5 means treating 5 people prevents one bad outcome that would otherwise have occurred.

Physicians and clinicians use NNT alongside relative risk to weigh a treatment's real-world benefit against its cost, side effects, or number needed to harm (NNH) when counseling patients, while epidemiologists and health-policy analysts use it to compare the cost-effectiveness of different drugs or interventions across a population. Medical students and pharmacists also use NNT as a standard evidence-based-medicine metric when interpreting clinical trial results.

Was this helpful?

You can pick up to 3 reactions.

Comments

  • Be the first to comment.

Popular calculators

All Calculators