Table 3. Area under curves (AUC) as a measure of predictive strength for risk-prediction models based on different indicators.
AUC | 95% CI | PAUC | Pa | |
---|---|---|---|---|
NJMU GWAS | ||||
Epidemiologic model | 0.61 | 0.597–0.623 | 1 | |
Genetic model | 0.653 | 0.639–0.668 | < 0.001 | |
The extended model | 0.697 | 0.683–0.711 | < 0.001 | 0.483 |
The EAGLE study | ||||
Epidemiologic model | 0.625 | 0.613–0.637 | 1 | |
Genetic model | 0.558 | 0.540–0.576 | < 0.001 | |
The extended model | 0.647 | 0.630–0.664 | 0.004 | 0.662 |
The two data sets | ||||
Epidemiologic model | 0.625 | 0.615–0.634 | 1 | |
Genetic model | 0.604 | 0.593–0.616 | < 0.001 | |
The extended model | 0.658 | 0.647–0.669 | < 0.001 | 0.792 |
aCalculated by Hosmer-Lemeshow test.