An Accurate Risk Score Based on Anthropometric, Dietary, and Lifestyle Factors to Predict the Development of Type 2 Diabetes: Response to Chien

We appreciate Chien's comment (1) and his interest in our article (2). He noted an alleged error in our formula for estimating the probability to develop diabetes in the next 5 years. He proposed to subtract the sample mean values of the risk factors in advance, as should be done in applying the Framingham coronary heart disease risk score. However, the Framingham score is based on a slightly different statistical approach that works with a baseline survival rate at the mean values of the risk factors (3). In contrast, our approach is based on a baseline survival rate at the 0 values of the risk factors; thus, standardizing the risk factors by subtracting the sample means would create biased estimates of the probability of interest. Doing so would strongly underestimate the diabetes incidence in the European Prospective Investigation into Cancer and Nutrition Potsdam cohort compared with the observed incidence rate.