Copyright 2007 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2007
We appreciate Cheng's interest in our study,1 and we agree that it can be very useful to use additional measures of association to quantify the impact of risk factors on breast cancer incidence. The focus of our analysis was to evaluate whether physical activity was differentially associated with breast cancer subtypes defined by estrogen and progesterone receptor status, and therefore we used the RR as our main measure of association. The advantage of the RR is that it is generally independent of the underlying baseline risk and therefore can be compared more easily across studies. This facilitates replication of findings, which is critical before assessing the impact of a risk or protective factor. However, readers are also interested in the potential public health significance of putative associations, and we also provided an estimate of the population-attributable risk of our findings, which suggested that 11% of breast cancer incidence in our population could have been prevented if all women had engaged in high levels of physical activity. As Cheng and others have discussed, the NNE is a newer measure of public health impact that can be applied to our study.2 We suggest using an RR of 0.86 rather than 0.91 from our study (see Table 2 in our article1) because the former is adjusted for breast cancer risk factors, while the latter is less appropriate because it is further adjusted for obesity measures, which are thought to be part of the same mechanistic pathway as physical activity. Based on an RR of 0.86, the NNE is 1526. There is little guidance on an acceptable NNE threshold for a preventive intervention in the general population. For comparison, the number needed to screen for 5 years using the fecal occult blood test to prevent 1 colorectal cancer death was 1374, and the number needed to screen for 5 years using mammography in women aged 60 to 69 years to prevent 1 breast cancer death was 1251.3 Both the population-attributable risk and NNE help evaluate the impact of potential interventions but must be interpreted with caution.4,5 Cheng appears to imply that physical activity is not an important protective factor for breast cancer at the population level. We believe that it is premature to rule out the importance of this factor, given that this conclusion is based on a single study and the fact that there are currently few readily modifiable risk factors for breast cancer. We must stress, however, that evaluating a specific preventive intervention at the population level was not the goal of our study, which was focused on addressing a specific question in breast cancer etiology.
Cerhan JR, Bardia A, Vierkant RA, Vachon CM. 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—Reply. Arch Intern Med. 2007;167(15):1690-1691. doi:10.1001/archinte.167.15.1690-b