Big Picture: relative vs absolute risk

Trial results are often reported as a relative risk reduction (RRR): the proportional difference in event rates between two groups. They can also be reported as an absolute risk reduction (ARR): the plain difference in event rates. The two are calculated from the same figures and can look very different when the baseline event rate is low.

Example arithmetic, using made-up numbers: if 2 in 100 people in a comparison group have an event and 1 in 100 in a treated group do, then RRR = (2 − 1) ÷ 2 = 50% and ARR = 2% − 1% = 1 percentage point. The "number needed to treat" is 1 ÷ ARR = 100.