Number needed to treat
The number needed to treat (NNT) is the expected number of people treated, compared with a control treatment, for one additional person to benefit over a stated time period.
Remember
Divide 1 by the absolute risk difference (the risk in the control group minus the risk in the treated group, written as a decimal) to get the NNT: the bigger the absolute benefit, the smaller the NNT.
The number needed to treat (NNT) converts the benefit of a treatment into a number of patients. A drug is compared with a comparator, which is the control treatment, such as a placebo. The NNT is how many people need the drug instead of the comparator for one additional person to benefit, over a stated outcome and time period.
The calculation uses the absolute risk difference, also written as the absolute risk reduction (ARR). This is the risk of the outcome in the control group minus the risk in the treated group, written as a decimal. Then the NNT is 1 divided by that difference.
Take a trial in which 12% of the placebo group have a stroke over 5 years, compared with 8% of the drug group. The difference is 4 percentage points (4 in every 100), or 0.04 as a proportion, so the NNT is 1 divided by 0.04, which is 25. A result such as 1 divided by 0.06, which is 16.7, is conventionally rounded up to the next whole number, here 17.
A smaller NNT means a more effective treatment, because fewer patients need treating when the absolute benefit is larger. The NNT is an average expectation, so an NNT of 10 does not guarantee one benefit in every group of 10.
Do not confuse the ARR with the relative risk reduction, which is the fall in risk as a fraction of the starting risk. A drug that halves a stroke risk of 2 in 1000 sounds dramatic, but it prevents only 1 stroke per 1000 people treated, an NNT of 1000. The matching measure for harm is the number needed to harm (NNH).
Key facts
- Formula
- the NNT is 1 divided by the absolute risk difference, written as a decimal.
- Worked example
- a fall in risk from 12% to 8% is a difference of 0.04, so the NNT is 25.
- Rounding
- an answer such as 16.7 is conventionally rounded up to the next whole number, 17.
- Direction
- a smaller NNT means a more effective treatment.
- Average
- an NNT of 10 does not guarantee one benefit in every group of 10.
- Harm
- the matching measure for harm is the number needed to harm (NNH).
Exam traps
- Dividing 1 by the relative risk reduction does not give the NNT, because the NNT uses the absolute difference in risk, and when the starting risk is low that division gives a misleadingly small result.
- Subtracting the risk in the treated group from the risk in the control group gives the absolute risk reduction, not the NNT, because that difference still has to be turned upside down by dividing 1 by it.
- A figure worked out as the fall in risk divided by the starting risk is the relative risk reduction, not the NNT, because the NNT uses the absolute difference in risk.
Read the full notes
Tested in 7 SimpleMed+ questions
Practice it in Pharmacology, with the answer and a full explanation after every question.
Practice these questions