Home›Glossary›NNT

NNT

General

Number Needed to Treat

The number of patients who need to receive a treatment for one additional patient to benefit, compared to a control — a way to express real-world treatment effectiveness.

Definition

NNT (Number Needed to Treat) is a statistic from evidence-based medicine that expresses how many patients, on average, must receive a particular treatment for one additional patient to experience a specific beneficial outcome, compared to what would have happened without the treatment (or with a control/standard treatment).

NNT translates the results of a clinical trial into a concrete, real-world number that helps clinicians and patients weigh:

  • How effective a treatment actually is in absolute terms
  • Whether the benefit justifies the cost, side effects, or inconvenience of treatment
  • How one treatment compares to another when choosing between options

A companion metric, NNH (Number Needed to Harm), uses the same logic to express how many patients need to receive a treatment for one additional patient to experience a harmful side effect.

Formula

NNT = 1 ÷ ARR

Where ARR (Absolute Risk Reduction) = Event rate in control group − Event rate in treatment group

The result is conventionally rounded up to the nearest whole number, since NNT represents a count of people.

Worked Example

In a clinical trial testing a new cholesterol drug for preventing heart attacks over 5 years:

  • Control group event rate: 15% had a heart attack
  • Treatment group event rate: 10% had a heart attack
  • ARR = 15% − 10% = 5 percentage points = 0.05
  • NNT = 1 ÷ 0.05 = 20

Interpretation: 20 patients need to take this drug for 5 years for one additional heart attack to be prevented, compared to not taking it. Use the NNT calculator to compute this from any pair of event rates.

Key Things to Know

  • NNT is outcome-specific: The same drug can have different NNT values for different outcomes (e.g., preventing a heart attack vs. preventing death) — always check which outcome an NNT figure refers to.
  • Time frame matters: NNT is only meaningful alongside the study duration it was measured over (e.g., "NNT = 20 over 5 years") — a treatment with NNT = 20 over 1 year is far more potent than one with NNT = 20 over 10 years.
  • NNT complements statistical significance: A result can be statistically significant (unlikely due to chance) while still having a large NNT, meaning the effect is real but small in absolute, practical terms.
  • Compare against NNH: A favorable treatment decision generally requires NNT to be meaningfully lower than NNH for a comparably serious outcome — if a drug is nearly as likely to harm as to help, its NNT alone doesn't tell the full story.
  • Baseline risk changes NNT: The same relative treatment effect produces a much lower (better) NNT in high-risk patients than in low-risk patients, which is why NNT should be interpreted in the context of the population it was measured in, not assumed to generalize universally.

Related Calculators

NNT Calculator (Number Needed to Treat) →Calculate Number Needed to Treat (NNT) and absolute risk reduction from control and treatment group event rates. A standard evidence-based medicine statistic.

Related Terms

Statistical SignificanceStatistical Significance (p-value / Confidence Level)Pearl IndexPearl Index (Contraceptive Failure Rate)

Frequently Asked Questions

A low NNT (e.g., NNT = 2) means a treatment is highly effective — only 2 patients need to be treated for one additional patient to benefit compared to no treatment or a control. A high NNT (e.g., NNT = 100) means the treatment benefit is much smaller in absolute terms, even if it's still statistically significant. NNT is most useful for comparing the real-world impact of different treatments side by side.
NNT is calculated as 1 divided by the Absolute Risk Reduction (ARR), where ARR is the difference in event rates between the control group and the treatment group. For example, if 30% of a control group has an event and 20% of a treated group has an event, the ARR is 10 percentage points (0.10), and NNT = 1 ÷ 0.10 = 10. The [NNT calculator](/nnt-calculator/) automates this from your trial's event rates.
The raw calculation (1 ÷ ARR) can produce a decimal, which is conventionally rounded up to the next whole number since NNT represents a count of patients. A negative ARR (meaning the treatment group had worse outcomes than control) produces a negative NNT, which is instead reported as NNH (Number Needed to Harm) — the number of patients who need to receive the treatment for one additional patient to be harmed.
Generally yes for effectiveness, but NNT should always be interpreted alongside the severity of the outcome being prevented and the treatment's side-effect profile (often summarized as NNH). A treatment with NNT = 20 for preventing a minor, self-resolving symptom may be far less compelling than a treatment with NNT = 50 for preventing a heart attack or death, because the value of the prevented outcome differs enormously.
Relative risk reduction can make small absolute benefits look dramatic — a 50% relative risk reduction sounds impressive whether the underlying risk drops from 20% to 10% or from 0.02% to 0.01%. NNT anchors the effect in absolute, patient-relevant terms, making it much easier for clinicians and patients to judge whether a treatment is worth pursuing given its costs, side effects, and alternatives.