What NNT means
Number needed to treat (NNT) is one of the most honest ways to express how useful something is. It answers: *how many people need to take this for one extra person to benefit?* [1]
If a supplement has an NNT of 10 for a given outcome, then on average 10 people take it for 1 to get the benefit — the other 9 do not, at least not for that outcome.
How it is calculated
NNT is simply 1 divided by the absolute risk reduction (ARR). If an outcome happens in 10% of a placebo group and 5% of a supplement group, the ARR is 5 percentage points (0.05), and the NNT is 1 / 0.05 = 20.
Notice this needs the *absolute* difference, not the relative one. Dropping 'from 10% to 5%' is a 50% relative reduction — which sounds dramatic — but the absolute reduction is only 5 points. See relative vs. absolute risk.
Why a lower NNT is better
| NNT | Interpretation |
|---|---|
| 2–5 | Strong, easily felt benefit |
| 10–20 | Modest benefit |
| 50+ | Small benefit for most individuals |
| 100+ | Most people notice no change for that outcome |
Why supplement marketing rarely mentions NNT
Relative-risk headlines ('cuts risk in half!') sound far more exciting than 'NNT of 50.' Reporting only the relative number is a classic way to make a small effect look large. NCCIH's guidance on evaluating research encourages translating findings into plain, absolute terms before deciding whether a result matters to you [1].
Putting it to use
When you read that a supplement 'significantly reduced' something, look for the absolute numbers so you can estimate the NNT. A small absolute change spread across a large study can be statistically real yet barely meaningful for any one person.