NNT vs NNH: Risk-Benefit Comparator

Treatment benefit

NNT — Number Needed to Treat 50

Severity of outcome prevented 8 / 10

Treatment harm

NNH — Number Needed to Harm 200

Severity of harm risked 2 / 10

Scenario presets

Out of 100 patients treated

Benefit (favorable outcome)

–

Harmed (side effect)

–

Both (overlap)

–

Neither

–

Severity-weighted net clinical value

– Adjust the sliders above.

Benefit weight/100

–

Harm weight/100

–

How it's computed

NNT and NNH are both the reciprocal of an absolute rate: NNT = 1 / (absolute risk reduction), NNH = 1 / (absolute risk increase). A lower NNT means benefit happens more often; a lower NNH means harm happens more often — so on the raw scale a "worse" (higher) NNT than NNH looks bad. But raw counts alone can't decide anything: a rare benefit that prevents a stroke can still outweigh a more common but trivial side effect.

patients benefiting  ≈ 100 / NNT
patients harmed       ≈ 100 / NNH
net value / patient   = severityB/NNT − severityH/NNH

NNT — how many patients must be treated for one extra favorable outcome. Lower = more common benefit.
NNH — how many patients must be treated for one extra harmful outcome. Lower = more common harm.
Severity weights — a 1–10 scale for how bad the prevented outcome (benefit) or the side effect (harm) actually is. These make the NNT-vs-NNH comparison decision-relevant instead of purely statistical.
Net clinical value — the severity-weighted balance. Positive means the treatment nets out beneficial across the population even if its NNT is numerically worse than its NNH; negative means the opposite.

This is a simplified teaching model — real risk-benefit analysis also weighs baseline risk, reversibility of harm, and patient preference. Benefit and harm are drawn from opposite ends of the 100-patient grid; if the two groups would exceed 100 people, the overlap is shown as patients who could plausibly experience both.

NNT 50

NNH 200

Benefits Harmed Both Neither

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.