Skip to main content

Advertisement

Table 3 AUCs for risk models developed and validated in various populations: Approach I

From: Impact of correlation of predictors on discrimination of risk models in development and external populations

Validated in population Developed in population
A B C D E F G H I J K L M N 0 P
A 0.663 * * * * 0.656 0.663 0.652 0.663 0.556 ** ** * * * *
B 0.645 0.645 * * * 0.632 0.645 0.631 0.644 0.534 ** ** * * * *
C 0.639 * 0.639 * * 0.626 0.639 0.622 0.639 0.534 ** ** * * * *
D 0.660 * * 0.660 * 0.649 0.660 0.649 0.659 0.545 ** ** * * * *
E 0.676 * * * 0.676 0.670 0.676 0.670 0.676 0.570 ** ** * * * *
F 0.624 * * * * 0.629 0.624 0.602 0.625 0.578 ** ** * * * *
G 0.575 * * * * 0.571 0.575 0.570 0.575 0.526 ** ** * * * *
H 0.770 * * * * 0.728 0.770 0.789 0.767 0.502 ** ** * * * *
I 0.593 * * * * 0.590 0.593 0.587 0.593 0.534 ** ** * * * *
J 0.531 * * * * 0.529 0.530 0.531 0.530 0.632 ** ** * * * *
K 0.540 * * * * 0.571 0.540 0.502 0.543 0.615 0.616 ** * * * *
L 0.542 * * * * 0.541 0.540 0.541 0.542 0.602 ** 0.603 * * * *
M 0.804 * * * * 0.792 0.804 0.800 0.804 0.693 ** ** 0.804 * * *
N 0.781 * * * * 0.759 0.781 0.775 0.781 0.692 ** ** * 0.781 * *
O 0.795 * * * * 0.782 0.795 0.790 0.795 0.686 ** ** * * 0.795 *
P 0.810 * * * * 0.802 0.810 0.806 0.810 0.695 ** ** * * * 0.810
  1. *Risk models with the same adjusted ORs will have equal impact on an external validation population. Therefore, prediction models developed in population A-E and M-P will perform similarly in an external validation population, and thus the values indicated as ‘*’ in these columns are identical to those in column A
  2. **The adjusted ORs in population J-K are the same and therefore perform similarly in an external validation population. Thus, the values indicated as ‘**’ in columns K and L are identical to those in column J
  3. The numbers in bold indicate the AUC estimated in the development population