Citations 0
Editor's Correspondence
August 13/27, 2007

The Number Needed to Be Exposed: A Potential Use for Quantifying the Strength of an Individual Risk Factor Including a Protective Factor in a Cohort Study

Author Affiliations

Copyright 2007 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2007

Arch Intern Med. 2007;167(15):1690-1691. doi:10.1001/archinte.167.15.1690-a

A recent cohort study in the Archives showed that higher recreational physical activity in postmenopausal women was independently associated with low incidence of breast cancer after adjusting for multiple confounders.1 Based on the data reported by Bardia et al,1 there were 4.7 (dividing 1216 by 259 247) and 4.1 (dividing 582 by 141 180) breast cancer cases per 1000 person-years among women who had low- and high-recreational physical activity, respectively, at baseline. Compared with low-recreational physical activity, high-recreational physical activity was associated with a reduced breast cancer incidence by 0.6 per 1000 person-years (4.7 per 1000 person-years minus 4.1 per 1000 person-years). The potential limitation of such a small absolute risk reduction of breast cancer incidence related to high-recreational physical activity was not quantified and fully discussed.1 Based on the concept and calculation of the number needed to treat, “the number needed to be exposed” (NNE) was recently developed.2 The NNE is defined as the number of study participants who have a harm exposure (ie, have a risk factor) that is needed to be associated with 1 outcome, or NNE is defined as the number of study participants who have a beneficial exposure (ie, have a protective factor) that is needed to be associated with 1 less outcome. In assuming that the difference of breast cancer incidence between the low- and high-recreational physical activity groups was solely owing to the level of physical activity, the NNE in women who had high-recreational physical activity for 1 less case of breast cancer is 1667 (dividing 1000 by 0.6). However, a cohort study always has multiple confounders. The following adjusted NNE was proposed for a risk factor (harmful exposure)2:

First Page Preview View Large
First page PDF preview
First page PDF preview