In Reply: Dr Williamson correctly points out an alternative formula for the computation of attributable fractions using adjusted effect estimates:
PAR = Pe[(HR-1) ÷ HR]
where Pe represents the proportion of the population exposed to the risk factor.1 This alternative formula was used to recalculate all of the PARs that were reported in the article. The interpretation of the data are virtually identical using this formula compared with using the formula in the article. An example is the comparative data of the principal results, which were presented in Table 2 of the article. The PARs in the article vs using the alternative formula were 53% vs 52% (3 healthy lifestyle factors), 58% vs 57% (4 factors), 72% vs 71% (5 factors), and 78% vs 78% (6 factors).
Forman JP. Population Attributable Risk of Incident Hypertension in Women—Reply. JAMA. 2009;302(23):2550-2551. doi:10.1001/jama.2009.1827