RESULTS
##Delta Meld as predictor
##
## Call:
## glm(formula = as.factor(responder) ~ delta.MELD, family = binomial,
## data = beax1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.5605 -1.2666 0.9284 1.0908 1.3017
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.0423 0.2792 0.151 0.880
## delta.MELD 0.1649 0.1132 1.456 0.145
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 85.369 on 61 degrees of freedom
## Residual deviance: 83.151 on 60 degrees of freedom
## AIC: 87.151
##
## Number of Fisher Scoring iterations: 4#intercepts
## # A tibble: 2 x 5
## term estimate std.error statistic p.value
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 (Intercept) 0.0423 0.279 0.151 0.880
## 2 delta.MELD 0.165 0.113 1.46 0.145
#####Delta Albumin as predictor
##
## Call:
## glm(formula = as.factor(responder) ~ delta.Albumin, family = binomial,
## data = beax1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.6753 -1.2073 0.9306 1.0914 1.3234
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.002274 0.314466 0.007 0.994
## delta.Albumin 0.678168 0.653744 1.037 0.300
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 85.369 on 61 degrees of freedom
## Residual deviance: 84.253 on 60 degrees of freedom
## AIC: 88.253
##
## Number of Fisher Scoring iterations: 4## # A tibble: 2 x 5
## term estimate std.error statistic p.value
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 (Intercept) 0.00227 0.314 0.00723 0.994
## 2 delta.Albumin 0.678 0.654 1.04 0.300
BEA SCORE VS MODEL
##
## Call:
## glm(formula = responder ~ BE3A.Score, family = "binomial", data = beax1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.006 -1.200 0.536 1.155 1.959
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.5745 1.1249 -3.178 0.001485 **
## BE3A.Score 1.8141 0.5322 3.409 0.000652 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 85.369 on 61 degrees of freedom
## Residual deviance: 67.402 on 60 degrees of freedom
## AIC: 71.402
##
## Number of Fisher Scoring iterations: 4##
## Call:
## glm(formula = responder ~ age + c_sex + Bilirubin + c_ascites +
## Albumin + INR + ALT + Encephalopathy + eGFR + Platelet, family = "binomial",
## data = beax1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.1259 -0.7637 0.1980 0.7512 1.6184
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.789527 5.968747 -0.467 0.64025
## age -0.018021 0.037882 -0.476 0.63428
## c_sexMale -0.634485 0.801651 -0.791 0.42867
## Bilirubin -1.278650 0.477822 -2.676 0.00745 **
## c_ascitesSevere -0.089119 0.797281 -0.112 0.91100
## Albumin 1.912596 1.116173 1.714 0.08662 .
## INR -0.839159 1.326783 -0.632 0.52708
## ALT 0.035072 0.019849 1.767 0.07724 .
## Encephalopathy -1.913290 1.603119 -1.193 0.23268
## eGFR 0.035958 0.024060 1.494 0.13505
## Platelet -0.001798 0.010248 -0.175 0.86071
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 85.369 on 61 degrees of freedom
## Residual deviance: 57.937 on 51 degrees of freedom
## AIC: 79.937
##
## Number of Fisher Scoring iterations: 5
Confusion matrix of our model
## Confusion Matrix and Statistics
##
## response
## predicted 0 1
## 0 19 7
## 1 9 27
##
## Accuracy : 0.7419
## 95% CI : (0.615, 0.8447)
## No Information Rate : 0.5484
## P-Value [Acc > NIR] : 0.001343
##
## Kappa : 0.4757
##
## Mcnemar's Test P-Value : 0.802587
##
## Sensitivity : 0.6786
## Specificity : 0.7941
## Pos Pred Value : 0.7308
## Neg Pred Value : 0.7500
## Prevalence : 0.4516
## Detection Rate : 0.3065
## Detection Prevalence : 0.4194
## Balanced Accuracy : 0.7363
##
## 'Positive' Class : 0
## 
confusion matrix of be3a score
## Confusion Matrix and Statistics
##
## response
## predicted 0 1
## 0 9 3
## 1 19 31
##
## Accuracy : 0.6452
## 95% CI : (0.5134, 0.7626)
## No Information Rate : 0.5484
## P-Value [Acc > NIR] : 0.079335
##
## Kappa : 0.2456
##
## Mcnemar's Test P-Value : 0.001384
##
## Sensitivity : 0.3214
## Specificity : 0.9118
## Pos Pred Value : 0.7500
## Neg Pred Value : 0.6200
## Prevalence : 0.4516
## Detection Rate : 0.1452
## Detection Prevalence : 0.1935
## Balanced Accuracy : 0.6166
##
## 'Positive' Class : 0
## THUS AUC of our Model lightly better but not by much.
##
## DeLong's test for two correlated ROC curves
##
## data: g and g1
## Z = 1.0549, p-value = 0.2915
## alternative hypothesis: true difference in AUC is not equal to 0
## sample estimates:
## AUC of roc1 AUC of roc2
## 0.8413866 0.7710084BOth curves together
##
## Call:
## roc.formula(formula = responder ~ BE3A.Score, data = beax1)
##
## Data: BE3A.Score in 28 controls (responder No) < 34 cases (responder Yes).
## Area under the curve: 0.771## [1] "1/exp^-((Intercept)*-2.7895+age*-0.018+c_sexMale*-0.6345+Bilirubin*-1.2786+c_ascitesSevere*-0.0891+Albumin*1.9126+INR*-0.8392+ALT*0.0351+Encephalopathy*-1.9133+eGFR*0.036+Platelet*-0.0018)"