Table 7: Multiple Logistic Regression Model and ROC Analysis for Mortality.
Multiple logistic regression model |
||||||||||
|
ß (SH) |
Odds ratio |
95% Confidence interval |
p |
||||||
SBP |
-0.180 (0.084) |
0.835 |
0.709-0.985 |
0.032 |
||||||
Pulse |
0.225 (0.098) |
1.252 |
1.033-1.519 |
0.022 |
||||||
CRP |
0.751 (0.399) |
2.118 |
0.969-4.632 |
0.060 |
||||||
Total bilirubin |
0.943 (0.545) |
2.567 |
0.882-7.474 |
0.084 |
||||||
Nagelkerke: 0.843. Cox and Snell R Square: 0.495. -2 Log likelihood: 14.980 Hosmer and Lemeshow test: χ2 = 2.035. df = 8. p = 0.980 |
||||||||||
ROC analysis for mortality |
||||||||||
|
AUC |
Cut-off |
Sensitivity |
Specificity |
PPV |
NPV |
||||
Ranson |
0.880 |
4 |
0.833 |
0.887 |
0.588 |
0.965 |
||||
BISAP |
0.856 |
3 |
0.833 |
0.839 |
0.5 |
0.963 |
||||
Model |
0.988 |
0.185 |
1 |
0.919 |
0.706 |
1 |
||||
ß: Model coefficient; SE: Standard Error; AUC: Area under the Curve; PPV: Positive Predictive Value; NPV: Negative Predictive Value. BISAP: Bedside index of severity in acute pancreatitis