Number Needed to Treat (NNT) Calculator
Calculate NNT, NNH, ARR, RRR, relative risk, and odds ratio from clinical trial event rates.
Essential tool for evidence-based medicine.
What Is NNT? The Number Needed to Treat (NNT) tells you how many patients must receive a treatment for one additional patient to benefit. It was introduced by Laupacis et al. in 1988 and is now a cornerstone of evidence-based medicine.
Key Formulas
ARR (Absolute Risk Reduction) = CER − EER NNT = 1 / ARR
Where:
- CER = Control Event Rate (proportion of events in the control/placebo group)
- EER = Experimental Event Rate (proportion of events in the treatment group)
- ARR = absolute reduction in risk due to treatment
Relative Measures
RR (Relative Risk) = EER / CER RRR (Relative Risk Reduction) = 1 − RR = ARR / CER Odds Ratio = (EER / (1 − EER)) / (CER / (1 − CER))
NNH (Number Needed to Harm) When EER > CER, the treatment increases risk. In this case: ARI (Absolute Risk Increase) = EER − CER NNH = 1 / ARI
Interpreting NNT
One scale, used by the body text, the result panel, and the chart alike. There is no agreed standard in the literature, so treat the boundaries as convention rather than fact.
| NNT | Reading | Example |
|---|---|---|
| 1 | Every patient benefits (ideal, essentially unheard of) | — |
| 2 to 4 | Excellent | Antibiotics for confirmed strep throat |
| 5 to 9 | Very good | Aspirin after a heart attack |
| 10 to 19 | Moderate, likely worthwhile | Statins in high-risk secondary prevention |
| 20 to 49 | Modest, weigh against cost and side effects | Many blood pressure targets |
| 50 to 99 | Low, needs careful benefit-risk assessment | Much primary prevention |
| 100 and above | Minimal population-level impact | Screening in low-risk groups |
NNT is conventionally rounded up, because you cannot treat 12.4 patients. An NNT of 12.4 means 13 people need the treatment for one of them to benefit.
Why ARR beats RRR
A drug that cuts risk from 2% to 1% has a 50% relative risk reduction, which sounds dramatic and is how it will be reported. The absolute reduction is 1 percentage point, and the NNT is 100. Ninety-nine of every hundred people treated get no benefit at all from it.
This is not a subtle point and it is misused constantly, in press releases and sometimes in the papers themselves. Whenever you see a relative figure with no baseline risk attached, the number is unreadable. Ask what the control event rate was.
Number needed to harm
The same arithmetic run backwards. When the treatment group has more events than control, the absolute risk increase is EER − CER and NNH is 1 divided by it. An NNH of 33 means one extra person is harmed for every 33 treated.
Worth stating plainly: NNT and NNH are not comparable to each other unless the events they count are comparable. An NNT of 20 for preventing a stroke sits very differently against an NNH of 20 for a rash than against an NNH of 20 for a fatal bleed. The numbers do not encode severity, and a calculator cannot supply it.
Odds ratio versus relative risk
Odds ratios turn up in case-control studies and logistic regression, where relative risk cannot be calculated directly. For rare outcomes, under about 10%, the odds ratio approximates the relative risk closely enough to read as one. For common outcomes it overstates the effect, sometimes badly, and reporting an odds ratio as though it were a risk ratio is one of the more common errors in the applied literature.
How we build and check this calculator
This calculator runs entirely in your browser, so the numbers you enter stay on your device. The math behind it is written by hand and tested against worked examples and standard references before the page goes live.
SuperGlobalCalculator is independently built and maintained. See how we build and verify our calculators.