Reporting RRR rather than ARR may exaggerate the perceptions of benefit or harm, and can be misleading.

Consider drug A, which reduces mortality from 40 to 30 percent if taken for one year. In this case:

RRR is 25 percent
[(40–30)/40] x 100
ARR is 10 percent
40–30
NNT is 10 for one year
1/ARR, which is 1/10 percent (or 1/0.10)

Now consider use of drug A by patients with a lower baseline risk of dying (e.g., the drug reduces mortality from 1.00 to 0.75 percent if taken for one year). In this case:

RRR is still 25 percent
[(1.00–0.75)/1.00] x 100
ARR is only 0.25 percent
1.00–0.75
NNT is 400
1/ARR, which is 1/0.25 percent (or 1/0.0025)

Patients with the highest baseline risk benefit the most from treatment, but all patients need to take the drug for the few to benefit. However, all patients taking the therapy risk the possibility of adverse effects, and the treatment must be funded for all.