Translating effect size into people treated
Number needed to treat is the reciprocal of the absolute risk reduction. If an intervention drops an event rate from 10 percent to 5 percent, the absolute reduction is 5 percent and the NNT is 20, meaning about 20 people must be treated for one to benefit who would not have otherwise. A small NNT signals a strong, efficient intervention; a large NNT means many treated for each one helped.
Because it is built from absolute risk, NNT depends on baseline risk. The same drug has a smaller NNT in high-risk patients and a larger one in low-risk patients, so a single quoted NNT only applies to the population it was calculated in.
Weighing benefit against harm
Number needed to harm mirrors NNT but uses the absolute increase in a side effect, so it counts how many people must be treated for one extra harm to occur. Comparing NNT against NNH frames the core tradeoff of any therapy: how many gain a benefit versus how many incur a harm.
This is general educational information, not medical advice. These numbers summarize trial populations and do not predict an individual outcome, so treatment choices should be made with a qualified clinician.