TABLE 4

Estimates of Model Performance for the Prediction Models

	Goodness of fit			Discrimination		Calibration			Validation
Model	−2LLH	AIC	R²	AUC	DS	Intercept	Slope	HLp	R² boot	AUC boot
1	94.46	98.46	0.02	0.58	0.01	2.09	2.46	0.75	0.00	0.54
2–3	81.80	87.80	0.22	0.71	0.14	−0.04	0.94	1.00	0.17	0.70
4	80.59	88.59	0.23	0.71	0.15	−0.15	0.87	0.45	0.17	0.69
5	79.82	87.82	0.24	0.79	0.16	−0.14	0.86	0.42	0.18	0.76
6	78.03	88.03	0.27	0.78	0.17	−0.22	0.81	0.46	0.18	0.74

−2LLH = −2 loglikelihood; R² = Nagelkerke R²; DS = discrimination slope; HLp = Hosmer–Lemeshow P value; boot = internal validated with bootstrapping.