In Reply: Dr Cook suggests that our findings differ from those of prior studies. However, the proportion of individuals correctly reclassified in some of the studies she cites was relatively modest as well. For instance, in the Physicians' Health Study, the NRI for cardiovascular events was only 3.2% with the addition of both CRP and parental history of cardiovascular disease.1 Furthermore, although Cook describes risk reclassification measures as helpful in determining whether new markers can alter clinical decisions, not all reclassification is clinically meaningful. Indeed, our findings suggest that most reclassification does not involve movement across a treatment threshold and thus may not affect decision making.
Melander O, Newton-Cheh C, Wang TJ. Biomarkers for Prediction of Cardiovascular Events—Reply. JAMA. 2009;302(19):2089-2090. doi:10.1001/jama.2009.1638