Table 3.

Sensitivity, specificity, positive predictive value, and negative predictive value for prediction of mortality determined by logistic regression

Variable	Cutoff point*	SE	SP	PPV	NPV
MELD score > 17	0.3056	57.89	78.99	30.56	92.16
∆MELD ≥ 1	0.5455	85.71	79.17	54.55	95.00
CTP score > 8	0.7700	93.33	60.61	26.42	98.36
∆CTP ≥ 1	0.5333	80.00	82.93	53.33	94.44
2 variables (CTP score > 8 + ∆MELD ≥ 1)	0.9018	75.00	97.73	90.00	93.48
3 variables (CTP score > 8 + ∆MELD ≥ 1 + ∆CTP ≥ 1)	0.1550	90.00	84.62	60.00	97.06
4 variables (CTP score > 8 + ∆MELD ≥ 1 + MELD score > 17 + ∆CTP ≥ 1)	0.1565	90.00	84.62	60.00	97.06

SE – sensitivity; SP – specificity; PPV – positive predictive value; NPV – negative predictive value

^*Cutoff point means the classification threshold probability. The probability is used to classify patients as surviving (the probability less than this critical value) or non-surviving (the probability equal to or more than the cutoff point) in a logistic regression.