Number Needed to Treat (NNT)

The number needed to treat is the reciprocal of the absolute risk reduction, and it was proposed by Laupacis, Sackett and Roberts in 1988 as a deliberate corrective to relative risk reporting. If 30 % of the control arm has the outcome and 20 % of the treated arm does, the absolute reduction is 10 percentage points and the NNT is 10: ten people treated for one to benefit. The unit is people, and the number is a summary of a completed trial rather than any kind of prescription — this page computes no dose and no quantity that goes to anybody.

Its purpose is to make the same result feel its actual size. That trial can be reported as a 33 % relative risk reduction, which sounds transformative, or as an NNT of 10, which tells you nine of every ten people treated got nothing from it. Both are true. The second is the one that supports a decision, and the gap between how the two feel is exactly why the measure was invented.

Round it up, always. An NNT of 12.3 is reported as 13, not 12. The reason is that rounding down claims the benefit arrives sooner than the trial demonstrated: asserting an NNT of 12 states a result the data does not support, while 13 merely understates the benefit slightly and errs toward caution. The fractional value is kept in the calculation above because it is the honest arithmetic and because it is what carries forward into any further work — a cost-per-outcome-avoided, or a comparison between two treatments.

Three things an NNT does not travel with, and all three are routinely lost. It is specific to a baseline risk: the same treatment applied to a lower-risk population has a larger NNT, often much larger, so an NNT from a trial of high-risk patients does not describe a low-risk one. It is specific to a time horizon, and "NNT of 20" means nothing without "over five years". And it says nothing about harms — the companion number needed to harm has to sit beside it, because a treatment with an NNT of 20 and a number needed to harm of 15 is a bad bargain no matter how impressive the first figure looks alone.