Table 3.
Quality assessment of binary logistic regression models.
| Chi-square p value | HL test | R2 | Variables entered (regression weights/p value) | |
|---|---|---|---|---|
| Stepwise model | ||||
| Step 1 | 5.19 × 10–7 | 0.833 | 0.360 | K+ (0.223/7.4 × 10–5) |
| Step 2 | 1.85 × 10–8 | 0.215 | 0.479 |
K+ (0.223/1.4 × 10–4) GST (−0.255/0.007) |
| Step 3 | 1.82 × 10–9 | 0.840 | 0.560 |
MRC1 (0.509/0.014) GST (−0.257/0.007) K+ (0.225/3.3 × 10–4) |
| All-inclusive model (all predictor variables) | ||||
| One step | 4.7 × 10–9 | 0.446 | 0.596 |
MRC1 (0.419/0.048) GST (−0.222/0.015) K+ (0.194/0.003) Na+ (−0.010/0.227) LDH (−0.004/0.209) |
R2 Nagelkerke’s pseudo-R2 (0–1; 1 corresponds to perfect identification of group membership), HL Hosmer–Lemeshow test p value (cut-off > 0.05).