Table 4. Ability to predict 10-year renal outcome with and without J-score.
Model 1 | C-statistics | NRI | IDI |
Clinical | 0.661 (0.614, 0.709) | ||
Clinical + J-score | 0.685 (0.634, 0.736) | 0.154 (-0.040, 0.349), p = 0.12 | 0.015 (0.003, 0.028), p = 0.02 |
Clinical + D-score | 0.751 (0.672, 0.830) | 0.691 (0.355, 1.027), p<0.001 | 0.031 (-0.008, 0.062), p = 0.06 |
Model 2 | C-statistics | NRI | IDI |
Clinical | 0.722 (0.674, 0.769) | ||
Clinical + J-score | 0.724 (0.672, 0.775) | 0.194 (-0.010, 0.398), p = 0.06 | 0.013 (0.001, 0.024), p = 0.03 |
Clinical + D-score | 0.673 (0.571, 0.775) | 0.539 (0.185, 0.893), p = 0.003 | 0.043 (-0.010, 0.398), p = 0.06 |
Clinical model 1: estimated glomerular filtration rate (eGFR) and albuminuria category
Clinical model 2: age, sex, eGFR, blood pressure, body mass index, hemoglobin A1c, and albuminuria category
NRI, net reclassification improvement; IDI, integrated discrimination improvement