A clinical trial reports that a new medication reduces cardiovascular events by 25%. Impressive—until you realize this relative reduction might mean preventing one event per thousand patients treated, or one per twenty. The difference matters enormously for treatment decisions, yet relative risk reductions dominate medical headlines precisely because they sound more dramatic.

Number needed to treat (NNT) cuts through this statistical fog. This single figure answers the question patients and clinicians actually need answered: how many people must receive this treatment for one person to benefit? An NNT of 5 means treating five patients prevents one adverse outcome. An NNT of 500 means treating five hundred.

Understanding NNT transforms how you evaluate treatment evidence. It moves discussion from abstract percentages to concrete clinical reality—the actual patients in your waiting room, the actual resources required, the actual likelihood that this particular intervention will make a difference for this particular person.

NNT derives from absolute risk reduction, not relative risk reduction. Consider a trial where 4% of control patients experience heart attacks versus 3% of treated patients. The relative risk reduction is 25%—a quarter fewer events. The absolute risk reduction is 1 percentage point. Invert that absolute reduction (1 divided by 0.01), and you get an NNT of 100.

That same 25% relative reduction looks entirely different when baseline risk changes. If control group risk is 40% and treatment reduces it to 30%, the absolute reduction is 10 percentage points. NNT becomes 10. Identical relative efficacy produces vastly different real-world impact depending on how common the outcome is without treatment.

This calculation exposes why relative risk reductions can mislead. Drug marketing frequently emphasizes relative figures because they maximize apparent benefit. A 50% reduction in disease recurrence sounds transformative. But if recurrence drops from 2% to 1%, that 50% reduction requires treating 100 patients to prevent one recurrence. The treatment may still be worthwhile—but the scale of benefit differs dramatically from what headlines suggest.

NNT also enables direct comparison across treatments and conditions. Aspirin for secondary prevention of cardiovascular events carries an NNT around 50-100 depending on population. Statins for primary prevention in low-risk patients might exceed 200. These numbers allow rational prioritization when resources, time, or patient tolerance limit what treatments can be implemented.

Takeaway

Always ask for absolute risk reduction alongside relative figures—then calculate or look up the NNT to understand how many patients must be treated for one to benefit.

An NNT of 20 means different things for different interventions. For a daily medication with potential side effects, ongoing costs, and indefinite duration, treating 20 patients to benefit one raises legitimate questions. For a single vaccine preventing a potentially fatal infection, an NNT of 20 represents remarkable efficiency.

Disease severity fundamentally shapes NNT interpretation. Preventing one death among 50 treated patients differs morally and practically from preventing one episode of mild nausea among 50 treated. The same numerical NNT can represent urgent clinical priority or marginal benefit depending on what outcome is being prevented.

Treatment burden must enter the calculation. An NNT of 15 for a well-tolerated once-daily pill differs from an NNT of 15 for chemotherapy with significant toxicity. Number needed to harm (NNH)—the inverse calculation for adverse effects—provides the counterbalance. When NNT and NNH approach similar values, treatment becomes a closer decision requiring careful individual assessment.

Alternative options provide essential context. An NNT of 30 might be excellent if no other treatment exists, or poor if a safer option achieves an NNT of 15. Clinical decisions rarely involve single interventions in isolation. Comparing NNTs across available treatments, while accounting for their respective harms and burdens, allows genuinely informed choice.

Takeaway

Evaluate NNT against three factors: the severity of the outcome being prevented, the burden and risks of the treatment itself, and the NNT of alternative options.

Consider explaining statin therapy to a patient with elevated cardiovascular risk. Rather than stating 'this medication reduces your risk by 30%,' NNT enables concrete conversation: 'If I prescribe this to 40 patients like you for five years, we'd expect to prevent one heart attack or stroke. That might be you, or it might be one of the other 39—we can't know in advance.'

This framing respects patient autonomy. Some patients hear 'one in forty' and decide the daily medication, potential muscle aches, and monitoring are worthwhile protection. Others conclude the odds don't justify the burden. Both responses are rational given identical evidence—they reflect different values, not different understanding.

NNT also guides resource allocation decisions. When formularies restrict access or clinicians face time constraints, NNT helps identify where interventions deliver most value. Treatments with NNTs under 20 for serious outcomes generally warrant aggressive implementation. Those exceeding 100 require careful justification, particularly when harms, costs, or alternatives exist.

Published NNT databases and calculators assist clinical application. TheNNT.com provides evidence-based NNT values for common interventions, color-coded by whether benefits clearly outweigh harms. Many electronic medical records now incorporate NNT into clinical decision support, presenting this figure alongside prescribing recommendations.

Takeaway

Use NNT to structure treatment conversations around concrete probabilities rather than abstract percentages—it transforms statistical evidence into terms patients can meaningfully weigh against their own values.

NNT strips away the statistical decoration that often obscures treatment evidence. By answering 'how many must be treated for one to benefit,' it forces honest reckoning with what interventions actually accomplish.

This metric doesn't make decisions—it clarifies them. The same NNT can support either treating or withholding depending on context, alternatives, and values. What changes is the quality of reasoning behind the choice.

Fluency with NNT marks sophisticated engagement with medical evidence. It protects against being swayed by dramatic-sounding relative reductions while missing modest absolute benefits. Every treatment decision improves when this single number enters the conversation.