How this calculator works
Absolute risk reduction subtracts the intervention event probability from the baseline probability. Relative reduction divides this difference by baseline risk. For an undesirable event, a positive reduction represents benefit; a negative reduction represents increased risk or harm. The reciprocal absolute difference is rounded upward to a whole-person number needed to treat or harm. Both risks must concern the same population definition, outcome and follow-up period. Estimates have uncertainty and can be confounded; no confidence interval or causal conclusion is supplied here.
Formula
ARR = control risk - intervention risk. RRR = ARR/control risk. NNT or NNH = ceil[1/|ARR|], with risks expressed as fractions. Equal risks give an infinite number needed.
Worked example
A baseline event risk of 20% and intervention risk of 10% give a 10 percentage-point absolute reduction, a 50% relative reduction and a number needed to treat of 10 over the stated period.
Advertisement
Frequently asked questions
Is a 50% reduction always a large absolute benefit?
No. Reducing risk from 2% to 1% is a 50% relative reduction but only a one percentage-point absolute reduction.
What does a negative absolute reduction mean?
For an undesirable event, the intervention risk is higher. The reciprocal absolute difference is then a number needed to harm.
Can I compare a one-year risk with a five-year risk?
No. The risks need a common follow-up period and outcome definition. The period field labels that shared basis; it does not convert probabilities between periods.