TTE

##   threshold specificity sensitivity 
##  49.5000000   0.8846154   0.6153846

MIF

##   threshold specificity sensitivity 
##  59.2130300   1.0000000   0.6153846
## 
##  DeLong's test for two ROC curves
## 
## data:  mod2 and mod1
## D = 0.12005, df = 80.589, p-value = 0.9047
## alternative hypothesis: true difference in AUC is not equal to 0
## sample estimates:
## AUC of roc1 AUC of roc2 
##   0.7717122   0.7544379

Compare with TTE, AUC increased from 0.75 to 0.77, p-value=0.9047.

MIF+TTE

##  threshold.mif threshold.tte specificity sensitivity
##       60.24626            50   0.8846154   0.9230769

Logistic Model

##   threshold specificity sensitivity 
##   0.1767037   0.7692308   0.9230769
## 
##  DeLong's test for two ROC curves
## 
## data:  mod3 and mod1
## D = 1.3588, df = 54.377, p-value = 0.1798
## alternative hypothesis: true difference in AUC is not equal to 0
## sample estimates:
## AUC of roc1 AUC of roc2 
##   0.9082840   0.7544379

Compare with TTE, AUC increased from 0.75 to 0.91, p-value=0.1798.

ROC curve (TTE vs. MIF vs. Logistic Model)

Quick Rule 1: Event if MIF > 60 or TTE > 50

##           POPH
## pred.rule1  0  1
##      FALSE 24  1
##      TRUE   2 12
## specificity sensitivity 
##   0.9230769   0.9230769

Quick Rule 2: Event if .4*MIF + .6*TTE > 50

##           POPH
## pred.rule2  0  1
##      FALSE 24  5
##      TRUE   2  8
## specificity sensitivity 
##   0.9230769   0.6153846