Number Needed to Treat Calculator

The Number Needed to Treat (NNT) is the number of patients who must receive a treatment over a specified period to prevent one additional adverse outcome compared to a control group. An NNT of 1 means every patient treated benefits, while a higher NNT means fewer patients benefit per person treated. Lower NNT values indicate a more effective treatment.

NNT is calculated as 1 divided by the Absolute Risk Reduction (ARR). ARR is the difference between the control group event rate and the treatment group event rate: ARR = Control Event Rate − Treatment Event Rate. Then NNT = 1 / ARR. For example, if the control rate is 26% and treatment rate is 16%, ARR = 10%, and NNT = 1 / 0.10 = 10 patients.

When your data is expressed as simple proportions or percentages (e.g., 26 out of 100 patients had an event), use the Percentage/Proportion type. If your study tracks events over observation time (e.g., 12 events per 1,000 patient-years), use the Patient-Years type. The patient-years method uses an exponential formula (1 − e^(−events/time)) to calculate incidence rates, which accounts for varying follow-up durations.

An NNT of 10 means you need to treat 10 patients with the experimental therapy for one additional patient to benefit (i.e., avoid the adverse outcome) compared to the control treatment. In clinical terms, the lower the NNT, the more effective the treatment is at a population level.

Absolute Risk Reduction (ARR), also called the risk difference, is the arithmetic difference between the event rate in the control group and the event rate in the treatment group. It directly quantifies how much the treatment reduces the probability of the adverse outcome. ARR = Control Event Rate − Treatment Event Rate.

The Number Needed to Harm (NNH) is the flip side of NNT — it represents how many patients must receive a treatment before one additional patient is harmed by it. When the treatment event rate exceeds the control event rate, the ARR becomes negative and the NNH is calculated as 1 / |ARR|. A high NNH is desirable, indicating harm occurs rarely.

Relative Risk Reduction (RRR) expresses the ARR as a proportion of the control group event rate: RRR = ARR / Control Event Rate. Unlike ARR, RRR does not reflect the baseline risk, which makes it appear more impressive. For instance, a treatment that reduces risk from 2% to 1% has an ARR of 1% but an RRR of 50%. ARR and NNT are considered more clinically meaningful.

The confidence interval for NNT is derived from the confidence interval of the ARR. Using the standard error of the risk difference (based on event counts and group sizes), we compute the Z-score for the chosen confidence level and apply it to ARR ± Z × SE(ARR). The NNT CI bounds are then 1 / ARR_upper and 1 / ARR_lower. A wide confidence interval indicates more uncertainty around the NNT estimate.