In this study, we first compared the ability of two phenotype based risk algorithms, the Cambridge risk score and the Framingham offspring risk score,7 8 with genotype based risk models for the discrimination of incident cases of type 2 diabetes in the prospective Whitehall II study of civil servants in the United Kingdom. We then evaluated the effect of addition of genotypes to the two non-genetic risk algorithms on the ability to discriminate between people who developed type 2 diabetes and those who remained free of the disease. We further assessed the effect of adding genotype to the calibration of the risk models (which involves assessing the closeness of the observed event rate to the predicted risk in different risk categories) as well as reclassification (which assesses the extent to which addition of genetic information shifts people to categories of predicted risk that better reflect their eventual outcome).19
