feature selection cohort without transplant
Model1 : age sex_cat
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.592
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.320
## Area under the curve: 0.6635
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 3185 828
## died 37568 30736
##
## Accuracy : 0.4691
## 95% CI : (0.4654, 0.4727)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.0459
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.97377
## Specificity : 0.07815
## Pos Pred Value : 0.44999
## Neg Pred Value : 0.79367
## Prevalence : 0.43647
## Detection Rate : 0.42502
## Detection Prevalence : 0.94451
## Balanced Accuracy : 0.52596
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.335130 0.006014 -55.725 <2e-16 ***
## age 0.619440 0.004335 142.895 <2e-16 ***
## sex_catMale 0.067073 0.007902 8.488 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 372662 on 289265 degrees of freedom
## AIC: 372668
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.86 |
1.84, 1.87 |
<0.001 |
sex_catMale |
1.07 |
1.05, 1.09 |
<0.001 |
Model2 : age sex_cat ALBUMIN
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.682
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.381
## Area under the curve: 0.7309
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 7448 1416
## died 33305 30148
##
## Accuracy : 0.5199
## 95% CI : (0.5162, 0.5235)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1238
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9551
## Specificity : 0.1828
## Pos Pred Value : 0.4751
## Neg Pred Value : 0.8403
## Prevalence : 0.4365
## Detection Rate : 0.4169
## Detection Prevalence : 0.8774
## Balanced Accuracy : 0.5689
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.413269 0.006294 -65.66 <2e-16 ***
## age 0.607550 0.004516 134.52 <2e-16 ***
## sex_catMale 0.181555 0.008308 21.85 <2e-16 ***
## ALBUMIN -0.690252 0.004610 -149.72 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 346451 on 289264 degrees of freedom
## AIC: 346459
##
## Number of Fisher Scoring iterations: 3
Characteristic |
OR |
95% CI |
p-value |
age |
1.84 |
1.82, 1.85 |
<0.001 |
sex_catMale |
1.20 |
1.18, 1.22 |
<0.001 |
ALBUMIN |
0.50 |
0.50, 0.51 |
<0.001 |
Model_3 : age sex_cat ALBUMIN COMO_CHF
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.690
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.387
## Area under the curve: 0.7374
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8071 1518
## died 32682 30046
##
## Accuracy : 0.5271
## 95% CI : (0.5234, 0.5307)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1349
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9519
## Specificity : 0.1980
## Pos Pred Value : 0.4790
## Neg Pred Value : 0.8417
## Prevalence : 0.4365
## Detection Rate : 0.4155
## Detection Prevalence : 0.8674
## Balanced Accuracy : 0.5750
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.569073 0.007008 -81.20 <2e-16 ***
## age 0.587618 0.004550 129.15 <2e-16 ***
## sex_catMale 0.185538 0.008352 22.21 <2e-16 ***
## ALBUMIN -0.681172 0.004621 -147.42 <2e-16 ***
## COMO_CHFY 0.457387 0.008682 52.68 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 343675 on 289263 degrees of freedom
## AIC: 343685
##
## Number of Fisher Scoring iterations: 3
Characteristic |
OR |
95% CI |
p-value |
age |
1.80 |
1.78, 1.82 |
<0.001 |
sex_catMale |
1.20 |
1.18, 1.22 |
<0.001 |
ALBUMIN |
0.51 |
0.50, 0.51 |
<0.001 |
COMO_CHFY |
1.58 |
1.55, 1.61 |
<0.001 |
model_4 : age sex_cat ALBUMIN COMO_CHF COMO_DM_INS
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.690
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.387
## Area under the curve: 0.7376
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8069 1507
## died 32684 30057
##
## Accuracy : 0.5272
## 95% CI : (0.5236, 0.5309)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1352
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9523
## Specificity : 0.1980
## Pos Pred Value : 0.4791
## Neg Pred Value : 0.8426
## Prevalence : 0.4365
## Detection Rate : 0.4156
## Detection Prevalence : 0.8676
## Balanced Accuracy : 0.5751
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.593365 0.007968 -74.469 < 2e-16 ***
## age 0.592089 0.004608 128.491 < 2e-16 ***
## sex_catMale 0.187209 0.008356 22.403 < 2e-16 ***
## ALBUMIN -0.678758 0.004634 -146.488 < 2e-16 ***
## COMO_CHFY 0.451465 0.008729 51.718 < 2e-16 ***
## COMO_DM_INSY 0.054199 0.008424 6.434 1.24e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 343633 on 289262 degrees of freedom
## AIC: 343645
##
## Number of Fisher Scoring iterations: 3
Characteristic |
OR |
95% CI |
p-value |
age |
1.81 |
1.79, 1.82 |
<0.001 |
sex_catMale |
1.21 |
1.19, 1.23 |
<0.001 |
ALBUMIN |
0.51 |
0.50, 0.51 |
<0.001 |
COMO_CHFY |
1.57 |
1.54, 1.60 |
<0.001 |
COMO_DM_INSY |
1.06 |
1.04, 1.07 |
<0.001 |
model_5 : age sex_cat ALBUMIN COMO_CHF COMO_DM_INS COMO_HTN
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.691
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.388
## Area under the curve: 0.7383
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8157 1526
## died 32596 30038
##
## Accuracy : 0.5282
## 95% CI : (0.5245, 0.5318)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1366
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9517
## Specificity : 0.2002
## Pos Pred Value : 0.4796
## Neg Pred Value : 0.8424
## Prevalence : 0.4365
## Detection Rate : 0.4154
## Detection Prevalence : 0.8661
## Balanced Accuracy : 0.5759
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.382622 0.013970 -27.388 < 2e-16 ***
## age 0.593702 0.004612 128.727 < 2e-16 ***
## sex_catMale 0.185036 0.008362 22.128 < 2e-16 ***
## ALBUMIN -0.675364 0.004638 -145.616 < 2e-16 ***
## COMO_CHFY 0.453370 0.008736 51.896 < 2e-16 ***
## COMO_DM_INSY 0.059086 0.008434 7.006 2.45e-12 ***
## COMO_HTNY -0.239224 0.013062 -18.314 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 343298 on 289261 degrees of freedom
## AIC: 343312
##
## Number of Fisher Scoring iterations: 3
Characteristic |
OR |
95% CI |
p-value |
age |
1.81 |
1.79, 1.83 |
<0.001 |
sex_catMale |
1.20 |
1.18, 1.22 |
<0.001 |
ALBUMIN |
0.51 |
0.50, 0.51 |
<0.001 |
COMO_CHFY |
1.57 |
1.55, 1.60 |
<0.001 |
COMO_DM_INSY |
1.06 |
1.04, 1.08 |
<0.001 |
COMO_HTNY |
0.79 |
0.77, 0.81 |
<0.001 |
model_6 : age sex_cat ALBUMIN COMO_CHF COMO_DM_INS COMO_HTN
CLM_FROM_3year_cat
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.698
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.399
## Area under the curve: 0.7489
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9515 1533
## died 31238 30031
##
## Accuracy : 0.5468
## 95% CI : (0.5432, 0.5505)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1672
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9514
## Specificity : 0.2335
## Pos Pred Value : 0.4901
## Neg Pred Value : 0.8612
## Prevalence : 0.4365
## Detection Rate : 0.4153
## Detection Prevalence : 0.8472
## Balanced Accuracy : 0.5925
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.394771 0.014096 -28.01 < 2e-16 ***
## age 0.600553 0.004649 129.18 < 2e-16 ***
## sex_catMale 0.203954 0.008439 24.17 < 2e-16 ***
## ALBUMIN -0.649967 0.004662 -139.42 < 2e-16 ***
## COMO_CHFY 0.432255 0.008812 49.05 < 2e-16 ***
## COMO_DM_INSY 0.051478 0.008508 6.05 1.45e-09 ***
## COMO_HTNY -0.237231 0.013176 -18.00 < 2e-16 ***
## CLM_FROM_3year_cat 0.291374 0.004271 68.23 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 338541 on 289260 degrees of freedom
## AIC: 338557
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.82 |
1.81, 1.84 |
<0.001 |
sex_catMale |
1.23 |
1.21, 1.25 |
<0.001 |
ALBUMIN |
0.52 |
0.52, 0.53 |
<0.001 |
COMO_CHFY |
1.54 |
1.51, 1.57 |
<0.001 |
COMO_DM_INSY |
1.05 |
1.04, 1.07 |
<0.001 |
COMO_HTNY |
0.79 |
0.77, 0.81 |
<0.001 |
CLM_FROM_3year_cat |
1.34 |
1.33, 1.35 |
<0.001 |
model_7 : age ALBUMIN COMO_CHF
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.689
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.386
## Area under the curve: 0.7367
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8021 1491
## died 32732 30073
##
## Accuracy : 0.5268
## 95% CI : (0.5231, 0.5304)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1346
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9528
## Specificity : 0.1968
## Pos Pred Value : 0.4788
## Neg Pred Value : 0.8433
## Prevalence : 0.4365
## Detection Rate : 0.4158
## Detection Prevalence : 0.8685
## Balanced Accuracy : 0.5748
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.462695 0.005077 -91.13 <2e-16 ***
## age 0.584412 0.004540 128.73 <2e-16 ***
## ALBUMIN -0.672429 0.004592 -146.43 <2e-16 ***
## COMO_CHFY 0.455627 0.008674 52.53 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 344170 on 289264 degrees of freedom
## AIC: 344178
##
## Number of Fisher Scoring iterations: 3
Characteristic |
OR |
95% CI |
p-value |
age |
1.79 |
1.78, 1.81 |
<0.001 |
ALBUMIN |
0.51 |
0.51, 0.52 |
<0.001 |
COMO_CHFY |
1.58 |
1.55, 1.60 |
<0.001 |
model_8 : age ALBUMIN COMO_CHF IRON_SAT_PERCENT
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.697
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.392
## Area under the curve: 0.7424
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8596 1593
## died 32157 29971
##
## Accuracy : 0.5333
## 95% CI : (0.5297, 0.5369)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1447
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9495
## Specificity : 0.2109
## Pos Pred Value : 0.4824
## Neg Pred Value : 0.8437
## Prevalence : 0.4365
## Detection Rate : 0.4144
## Detection Prevalence : 0.8591
## Balanced Accuracy : 0.5802
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.456212 0.005103 -89.40 <2e-16 ***
## age 0.600756 0.004581 131.13 <2e-16 ***
## ALBUMIN -0.650195 0.004599 -141.38 <2e-16 ***
## COMO_CHFY 0.424024 0.008735 48.54 <2e-16 ***
## IRON_SAT_PERCENT -0.222513 0.004234 -52.55 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 341365 on 289263 degrees of freedom
## AIC: 341375
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.82 |
1.81, 1.84 |
<0.001 |
ALBUMIN |
0.52 |
0.52, 0.53 |
<0.001 |
COMO_CHFY |
1.53 |
1.50, 1.55 |
<0.001 |
IRON_SAT_PERCENT |
0.80 |
0.79, 0.81 |
<0.001 |
model_9 : age ALBUMIN COMO_CHF IRON_SAT_PERCENT HGB
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.703
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.397
## Area under the curve: 0.746
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8886 1631
## died 31867 29933
##
## Accuracy : 0.5368
## 95% CI : (0.5331, 0.5404)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1502
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9483
## Specificity : 0.2180
## Pos Pred Value : 0.4844
## Neg Pred Value : 0.8449
## Prevalence : 0.4365
## Detection Rate : 0.4139
## Detection Prevalence : 0.8546
## Balanced Accuracy : 0.5832
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.458925 0.005123 -89.58 <2e-16 ***
## age 0.609162 0.004599 132.46 <2e-16 ***
## ALBUMIN -0.591095 0.004773 -123.83 <2e-16 ***
## COMO_CHFY 0.435216 0.008773 49.61 <2e-16 ***
## IRON_SAT_PERCENT -0.194505 0.004288 -45.36 <2e-16 ***
## HGB -0.202541 0.004549 -44.52 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 339349 on 289262 degrees of freedom
## AIC: 339361
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.84 |
1.82, 1.86 |
<0.001 |
ALBUMIN |
0.55 |
0.55, 0.56 |
<0.001 |
COMO_CHFY |
1.55 |
1.52, 1.57 |
<0.001 |
IRON_SAT_PERCENT |
0.82 |
0.82, 0.83 |
<0.001 |
HGB |
0.82 |
0.81, 0.82 |
<0.001 |
model_10 : age ALBUMIN COMO_CHF IRON_SAT_PERCENT
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.704
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.402
## Area under the curve: 0.7522
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9865 1568
## died 30888 29996
##
## Accuracy : 0.5512
## 95% CI : (0.5476, 0.5548)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1742
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9503
## Specificity : 0.2421
## Pos Pred Value : 0.4927
## Neg Pred Value : 0.8629
## Prevalence : 0.4365
## Detection Rate : 0.4148
## Detection Prevalence : 0.8419
## Balanced Accuracy : 0.5962
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.459411 0.005151 -89.19 <2e-16 ***
## age 0.607074 0.004614 131.57 <2e-16 ***
## ALBUMIN -0.625283 0.004623 -135.27 <2e-16 ***
## COMO_CHFY 0.403660 0.008806 45.84 <2e-16 ***
## IRON_SAT_PERCENT -0.214749 0.004265 -50.35 <2e-16 ***
## CLM_FROM_3year_cat 0.282746 0.004282 66.04 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 336916 on 289262 degrees of freedom
## AIC: 336928
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.84 |
1.82, 1.85 |
<0.001 |
ALBUMIN |
0.54 |
0.53, 0.54 |
<0.001 |
COMO_CHFY |
1.50 |
1.47, 1.52 |
<0.001 |
IRON_SAT_PERCENT |
0.81 |
0.80, 0.81 |
<0.001 |
CLM_FROM_3year_cat |
1.33 |
1.32, 1.34 |
<0.001 |
model_11 : age,ALBUMIN
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.681
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.380
## Area under the curve: 0.7301
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 7404 1421
## died 33349 30143
##
## Accuracy : 0.5192
## 95% CI : (0.5156, 0.5228)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1227
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9550
## Specificity : 0.1817
## Pos Pred Value : 0.4748
## Neg Pred Value : 0.8390
## Prevalence : 0.4365
## Detection Rate : 0.4168
## Detection Prevalence : 0.8780
## Balanced Accuracy : 0.5683
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.309793 0.004116 -75.26 <2e-16 ***
## age 0.604396 0.004506 134.12 <2e-16 ***
## ALBUMIN -0.681663 0.004582 -148.77 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 346930 on 289265 degrees of freedom
## AIC: 346936
##
## Number of Fisher Scoring iterations: 3
Characteristic |
OR |
95% CI |
p-value |
age |
1.83 |
1.81, 1.85 |
<0.001 |
ALBUMIN |
0.51 |
0.50, 0.51 |
<0.001 |
model_12 : age ALBUMIN IRON_SAT_PERCENT
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.690
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.387
## Area under the curve: 0.7372
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8086 1529
## died 32667 30035
##
## Accuracy : 0.5271
## 95% CI : (0.5235, 0.5308)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.135
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9516
## Specificity : 0.1984
## Pos Pred Value : 0.4790
## Neg Pred Value : 0.8410
## Prevalence : 0.4365
## Detection Rate : 0.4153
## Detection Prevalence : 0.8670
## Balanced Accuracy : 0.5750
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.314236 0.004142 -75.87 <2e-16 ***
## age 0.620486 0.004549 136.39 <2e-16 ***
## ALBUMIN -0.657290 0.004589 -143.24 <2e-16 ***
## IRON_SAT_PERCENT -0.236508 0.004211 -56.16 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 343720 on 289264 degrees of freedom
## AIC: 343728
##
## Number of Fisher Scoring iterations: 3
Characteristic |
OR |
95% CI |
p-value |
age |
1.86 |
1.84, 1.88 |
<0.001 |
ALBUMIN |
0.52 |
0.51, 0.52 |
<0.001 |
IRON_SAT_PERCENT |
0.79 |
0.78, 0.80 |
<0.001 |
model_13 : age ALBUMIN IRON_SAT_PERCENT HGB
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.697
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.391
## Area under the curve: 0.7407
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8388 1555
## died 32365 30009
##
## Accuracy : 0.531
## 95% CI : (0.5273, 0.5346)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.141
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9507
## Specificity : 0.2058
## Pos Pred Value : 0.4811
## Neg Pred Value : 0.8436
## Prevalence : 0.4365
## Detection Rate : 0.4150
## Detection Prevalence : 0.8625
## Balanced Accuracy : 0.5783
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.313152 0.004156 -75.36 <2e-16 ***
## age 0.629107 0.004567 137.76 <2e-16 ***
## ALBUMIN -0.600225 0.004760 -126.08 <2e-16 ***
## IRON_SAT_PERCENT -0.209681 0.004263 -49.19 <2e-16 ***
## HGB -0.196437 0.004531 -43.36 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 341809 on 289263 degrees of freedom
## AIC: 341819
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.88 |
1.86, 1.89 |
<0.001 |
ALBUMIN |
0.55 |
0.54, 0.55 |
<0.001 |
IRON_SAT_PERCENT |
0.81 |
0.80, 0.82 |
<0.001 |
HGB |
0.82 |
0.81, 0.83 |
<0.001 |
model_14 : age ALBUMIN IRON_SAT_PERCENT HGB CLM_FROM_3year_cat
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.703
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.401
## Area under the curve: 0.7509
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9718 1592
## died 31035 29972
##
## Accuracy : 0.5488
## 95% CI : (0.5452, 0.5525)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1701
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9496
## Specificity : 0.2385
## Pos Pred Value : 0.4913
## Neg Pred Value : 0.8592
## Prevalence : 0.4365
## Detection Rate : 0.4145
## Detection Prevalence : 0.8436
## Balanced Accuracy : 0.5940
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.323447 0.004199 -77.04 <2e-16 ***
## age 0.634053 0.004599 137.88 <2e-16 ***
## ALBUMIN -0.577922 0.004785 -120.77 <2e-16 ***
## IRON_SAT_PERCENT -0.202689 0.004294 -47.20 <2e-16 ***
## HGB -0.184980 0.004550 -40.65 <2e-16 ***
## CLM_FROM_3year_cat 0.283274 0.004280 66.19 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 337338 on 289262 degrees of freedom
## AIC: 337350
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.89 |
1.87, 1.90 |
<0.001 |
ALBUMIN |
0.56 |
0.56, 0.57 |
<0.001 |
IRON_SAT_PERCENT |
0.82 |
0.81, 0.82 |
<0.001 |
HGB |
0.83 |
0.82, 0.84 |
<0.001 |
CLM_FROM_3year_cat |
1.33 |
1.32, 1.34 |
<0.001 |
model_15 : age ALBUMIN IRON_SAT_PERCENT HGB CLM_FROM_3year_cat
race_cat_4
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.709
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.407
## Area under the curve: 0.7564
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 10361 1634
## died 30392 29930
##
## Accuracy : 0.5571
## 95% CI : (0.5535, 0.5608)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 0.9997
##
## Kappa : 0.1836
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9482
## Specificity : 0.2542
## Pos Pred Value : 0.4962
## Neg Pred Value : 0.8638
## Prevalence : 0.4365
## Detection Rate : 0.4139
## Detection Prevalence : 0.8341
## Balanced Accuracy : 0.6012
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.781039 0.023605 -33.088 < 2e-16 ***
## age 0.607024 0.004659 130.292 < 2e-16 ***
## ALBUMIN -0.572964 0.004807 -119.205 < 2e-16 ***
## IRON_SAT_PERCENT -0.196940 0.004315 -45.639 < 2e-16 ***
## HGB -0.202642 0.004593 -44.120 < 2e-16 ***
## CLM_FROM_3year_cat 0.275544 0.004303 64.037 < 2e-16 ***
## race_cat_4Black 0.162264 0.025086 6.468 9.91e-11 ***
## race_cat_4Other 0.070742 0.037778 1.873 0.0611 .
## race_cat_4White 0.605381 0.024095 25.125 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 334679 on 289259 degrees of freedom
## AIC: 334697
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.83 |
1.82, 1.85 |
<0.001 |
ALBUMIN |
0.56 |
0.56, 0.57 |
<0.001 |
IRON_SAT_PERCENT |
0.82 |
0.81, 0.83 |
<0.001 |
HGB |
0.82 |
0.81, 0.82 |
<0.001 |
CLM_FROM_3year_cat |
1.32 |
1.31, 1.33 |
<0.001 |
race_cat_4Black |
1.18 |
1.12, 1.24 |
<0.001 |
race_cat_4Other |
1.07 |
1.00, 1.16 |
0.061 |
race_cat_4White |
1.83 |
1.75, 1.92 |
<0.001 |
model_16 : age ALBUMIN IRON_SAT_PERCENT HGB CLM_FROM_3year_cat
race_cat_4 state_cat
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.709
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.408
## Area under the curve: 0.7569
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 10411 1638
## died 30342 29926
##
## Accuracy : 0.5578
## 95% CI : (0.5542, 0.5614)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 0.9991
##
## Kappa : 0.1846
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9481
## Specificity : 0.2555
## Pos Pred Value : 0.4965
## Neg Pred Value : 0.8641
## Prevalence : 0.4365
## Detection Rate : 0.4138
## Detection Prevalence : 0.8334
## Balanced Accuracy : 0.6018
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.650910 0.025767 -25.261 < 2e-16 ***
## age 0.608547 0.004680 130.023 < 2e-16 ***
## ALBUMIN -0.572683 0.004809 -119.092 < 2e-16 ***
## IRON_SAT_PERCENT -0.196667 0.004319 -45.536 < 2e-16 ***
## HGB -0.200091 0.004603 -43.468 < 2e-16 ***
## CLM_FROM_3year_cat 0.271778 0.004313 63.017 < 2e-16 ***
## race_cat_4Black 0.072134 0.025753 2.801 0.0051 **
## race_cat_4Other 0.058627 0.037838 1.549 0.1213
## race_cat_4White 0.540859 0.024445 22.126 < 2e-16 ***
## state_catFIPSNE -0.109269 0.013878 -7.874 3.45e-15 ***
## state_catFIPSS -0.018164 0.011205 -1.621 0.1050
## state_catFIPSW -0.190822 0.013564 -14.069 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 334404 on 289256 degrees of freedom
## AIC: 334428
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.84 |
1.82, 1.85 |
<0.001 |
ALBUMIN |
0.56 |
0.56, 0.57 |
<0.001 |
IRON_SAT_PERCENT |
0.82 |
0.81, 0.83 |
<0.001 |
HGB |
0.82 |
0.81, 0.83 |
<0.001 |
CLM_FROM_3year_cat |
1.31 |
1.30, 1.32 |
<0.001 |
race_cat_4Black |
1.07 |
1.02, 1.13 |
0.005 |
race_cat_4Other |
1.06 |
0.98, 1.14 |
0.12 |
race_cat_4White |
1.72 |
1.64, 1.80 |
<0.001 |
state_catFIPSNE |
0.90 |
0.87, 0.92 |
<0.001 |
state_catFIPSS |
0.98 |
0.96, 1.00 |
0.11 |
state_catFIPSW |
0.83 |
0.80, 0.85 |
<0.001 |
model_17 : age ALBUMIN IRON_SAT_PERCENT HGB CLM_FROM_3year_cat
CALCIUM_UNCORRECTED
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.703
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.401
## Area under the curve: 0.751
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9711 1594
## died 31042 29970
##
## Accuracy : 0.5487
## 95% CI : (0.5451, 0.5523)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1699
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9495
## Specificity : 0.2383
## Pos Pred Value : 0.4912
## Neg Pred Value : 0.8590
## Prevalence : 0.4365
## Detection Rate : 0.4144
## Detection Prevalence : 0.8437
## Balanced Accuracy : 0.5939
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.323449 0.004199 -77.023 <2e-16 ***
## age 0.628681 0.004632 135.722 <2e-16 ***
## ALBUMIN -0.597701 0.005256 -113.719 <2e-16 ***
## IRON_SAT_PERCENT -0.201675 0.004296 -46.946 <2e-16 ***
## HGB -0.186891 0.004556 -41.021 <2e-16 ***
## CLM_FROM_3year_cat 0.283053 0.004281 66.126 <2e-16 ***
## CALCIUM_UNCORRECTED 0.043700 0.004753 9.193 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 337254 on 289261 degrees of freedom
## AIC: 337268
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.88 |
1.86, 1.89 |
<0.001 |
ALBUMIN |
0.55 |
0.54, 0.56 |
<0.001 |
IRON_SAT_PERCENT |
0.82 |
0.81, 0.82 |
<0.001 |
HGB |
0.83 |
0.82, 0.84 |
<0.001 |
CLM_FROM_3year_cat |
1.33 |
1.32, 1.34 |
<0.001 |
CALCIUM_UNCORRECTED |
1.04 |
1.03, 1.05 |
<0.001 |
model_18 : age ALBUMIN IRON_SAT_PERCENT HGB CLM_FROM_3year_cat
CALCIUM_UNCORRECTED FERRITIN
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.704
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.402
## Area under the curve: 0.7512
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9733 1609
## died 31020 29955
##
## Accuracy : 0.5488
## 95% CI : (0.5452, 0.5524)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.17
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9490
## Specificity : 0.2388
## Pos Pred Value : 0.4913
## Neg Pred Value : 0.8581
## Prevalence : 0.4365
## Detection Rate : 0.4142
## Detection Prevalence : 0.8432
## Balanced Accuracy : 0.5939
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.323213 0.004200 -76.956 < 2e-16 ***
## age 0.625494 0.004645 134.655 < 2e-16 ***
## ALBUMIN -0.593606 0.005275 -112.535 < 2e-16 ***
## IRON_SAT_PERCENT -0.212468 0.004465 -47.590 < 2e-16 ***
## HGB -0.179696 0.004625 -38.850 < 2e-16 ***
## CLM_FROM_3year_cat 0.282728 0.004281 66.040 < 2e-16 ***
## CALCIUM_UNCORRECTED 0.039134 0.004782 8.184 2.74e-16 ***
## FERRITIN 0.040101 0.004486 8.938 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 337173 on 289260 degrees of freedom
## AIC: 337189
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.87 |
1.85, 1.89 |
<0.001 |
ALBUMIN |
0.55 |
0.55, 0.56 |
<0.001 |
IRON_SAT_PERCENT |
0.81 |
0.80, 0.82 |
<0.001 |
HGB |
0.84 |
0.83, 0.84 |
<0.001 |
CLM_FROM_3year_cat |
1.33 |
1.32, 1.34 |
<0.001 |
CALCIUM_UNCORRECTED |
1.04 |
1.03, 1.05 |
<0.001 |
FERRITIN |
1.04 |
1.03, 1.05 |
<0.001 |
model_19 : age ALBUMIN IRON_SAT_PERCENT HGB CLM_FROM_3year_cat
CALCIUM_UNCORRECTED FERRITIN PHOSPHORUS
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.704
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.402
## Area under the curve: 0.7518
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9789 1610
## died 30964 29954
##
## Accuracy : 0.5496
## 95% CI : (0.5459, 0.5532)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1712
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9490
## Specificity : 0.2402
## Pos Pred Value : 0.4917
## Neg Pred Value : 0.8588
## Prevalence : 0.4365
## Detection Rate : 0.4142
## Detection Prevalence : 0.8424
## Balanced Accuracy : 0.5946
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.324500 0.004204 -77.20 <2e-16 ***
## age 0.647402 0.004835 133.91 <2e-16 ***
## ALBUMIN -0.610331 0.005374 -113.58 <2e-16 ***
## IRON_SAT_PERCENT -0.213632 0.004467 -47.82 <2e-16 ***
## HGB -0.177689 0.004629 -38.39 <2e-16 ***
## CLM_FROM_3year_cat 0.281837 0.004283 65.80 <2e-16 ***
## CALCIUM_UNCORRECTED 0.049214 0.004819 10.21 <2e-16 ***
## FERRITIN 0.045267 0.004503 10.05 <2e-16 ***
## PHOSPHORUS 0.077540 0.004526 17.13 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 336880 on 289259 degrees of freedom
## AIC: 336898
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.91 |
1.89, 1.93 |
<0.001 |
ALBUMIN |
0.54 |
0.54, 0.55 |
<0.001 |
IRON_SAT_PERCENT |
0.81 |
0.80, 0.81 |
<0.001 |
HGB |
0.84 |
0.83, 0.84 |
<0.001 |
CLM_FROM_3year_cat |
1.33 |
1.31, 1.34 |
<0.001 |
CALCIUM_UNCORRECTED |
1.05 |
1.04, 1.06 |
<0.001 |
FERRITIN |
1.05 |
1.04, 1.06 |
<0.001 |
PHOSPHORUS |
1.08 |
1.07, 1.09 |
<0.001 |
model_20 : age ALBUMIN IRON_SAT_PERCENT HGB sex_cat
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.698
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.393
## Area under the curve: 0.7416
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 8415 1599
## died 32338 29965
##
## Accuracy : 0.5307
## 95% CI : (0.5271, 0.5344)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1404
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9493
## Specificity : 0.2065
## Pos Pred Value : 0.4810
## Neg Pred Value : 0.8403
## Prevalence : 0.4365
## Detection Rate : 0.4144
## Detection Prevalence : 0.8615
## Balanced Accuracy : 0.5779
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.428060 0.006347 -67.44 <2e-16 ***
## age 0.632889 0.004578 138.24 <2e-16 ***
## ALBUMIN -0.608342 0.004784 -127.17 <2e-16 ***
## IRON_SAT_PERCENT -0.209544 0.004268 -49.10 <2e-16 ***
## HGB -0.201912 0.004544 -44.43 <2e-16 ***
## sex_catMale 0.202388 0.008398 24.10 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 341226 on 289262 degrees of freedom
## AIC: 341238
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.88 |
1.87, 1.90 |
<0.001 |
ALBUMIN |
0.54 |
0.54, 0.55 |
<0.001 |
IRON_SAT_PERCENT |
0.81 |
0.80, 0.82 |
<0.001 |
HGB |
0.82 |
0.81, 0.82 |
<0.001 |
sex_catMale |
1.22 |
1.20, 1.24 |
<0.001 |
model_21 : age ALBUMIN IRON_SAT_PERCENT HGB sex_cat
CLM_FROM_3year_cat
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.705
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.403
## Area under the curve: 0.7519
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9758 1607
## died 30995 29957
##
## Accuracy : 0.5492
## 95% CI : (0.5455, 0.5528)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1706
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9491
## Specificity : 0.2394
## Pos Pred Value : 0.4915
## Neg Pred Value : 0.8586
## Prevalence : 0.4365
## Detection Rate : 0.4142
## Detection Prevalence : 0.8428
## Balanced Accuracy : 0.5943
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.448562 0.006414 -69.93 <2e-16 ***
## age 0.638204 0.004612 138.39 <2e-16 ***
## ALBUMIN -0.586378 0.004809 -121.94 <2e-16 ***
## IRON_SAT_PERCENT -0.202396 0.004300 -47.07 <2e-16 ***
## HGB -0.190764 0.004564 -41.80 <2e-16 ***
## sex_catMale 0.220124 0.008472 25.98 <2e-16 ***
## CLM_FROM_3year_cat 0.286745 0.004288 66.87 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 336660 on 289261 degrees of freedom
## AIC: 336674
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.89 |
1.88, 1.91 |
<0.001 |
ALBUMIN |
0.56 |
0.55, 0.56 |
<0.001 |
IRON_SAT_PERCENT |
0.82 |
0.81, 0.82 |
<0.001 |
HGB |
0.83 |
0.82, 0.83 |
<0.001 |
sex_catMale |
1.25 |
1.23, 1.27 |
<0.001 |
CLM_FROM_3year_cat |
1.33 |
1.32, 1.34 |
<0.001 |
model_22 : age ALBUMIN IRON_SAT_PERCENT HGB sex_cat
CLM_FROM_3year_cat race_cat_4
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.710
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.408
## Area under the curve: 0.7572
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 10380 1663
## died 30373 29901
##
## Accuracy : 0.557
## 95% CI : (0.5534, 0.5606)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 0.9998
##
## Kappa : 0.1832
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9473
## Specificity : 0.2547
## Pos Pred Value : 0.4961
## Neg Pred Value : 0.8619
## Prevalence : 0.4365
## Detection Rate : 0.4135
## Detection Prevalence : 0.8335
## Balanced Accuracy : 0.6010
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.893505 0.024154 -36.992 < 2e-16 ***
## age 0.611713 0.004673 130.907 < 2e-16 ***
## ALBUMIN -0.580583 0.004829 -120.235 < 2e-16 ***
## IRON_SAT_PERCENT -0.196748 0.004320 -45.544 < 2e-16 ***
## HGB -0.207072 0.004604 -44.977 < 2e-16 ***
## CLM_FROM_3year_cat 0.278812 0.004310 64.684 < 2e-16 ***
## sex_catMale 0.193818 0.008530 22.721 < 2e-16 ***
## race_cat_4Black 0.175971 0.025118 7.006 2.46e-12 ***
## race_cat_4Other 0.079055 0.037811 2.091 0.0365 *
## race_cat_4White 0.603252 0.024119 25.012 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 334161 on 289258 degrees of freedom
## AIC: 334181
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.84 |
1.83, 1.86 |
<0.001 |
ALBUMIN |
0.56 |
0.55, 0.56 |
<0.001 |
IRON_SAT_PERCENT |
0.82 |
0.81, 0.83 |
<0.001 |
HGB |
0.81 |
0.81, 0.82 |
<0.001 |
CLM_FROM_3year_cat |
1.32 |
1.31, 1.33 |
<0.001 |
sex_catMale |
1.21 |
1.19, 1.23 |
<0.001 |
race_cat_4Black |
1.19 |
1.14, 1.25 |
<0.001 |
race_cat_4Other |
1.08 |
1.00, 1.17 |
0.037 |
race_cat_4White |
1.83 |
1.74, 1.92 |
<0.001 |
model_23 : age ALBUMIN IRON_SAT_PERCENT HGB CLM_FROM_3year_cat
state_cat
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.705
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.403
## Area under the curve: 0.7524
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9792 1615
## died 30961 29949
##
## Accuracy : 0.5495
## 95% CI : (0.5459, 0.5532)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1712
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9488
## Specificity : 0.2403
## Pos Pred Value : 0.4917
## Neg Pred Value : 0.8584
## Prevalence : 0.4365
## Detection Rate : 0.4141
## Detection Prevalence : 0.8423
## Balanced Accuracy : 0.5946
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.360992 0.010336 -34.924 < 2e-16 ***
## age 0.639659 0.004639 137.883 < 2e-16 ***
## ALBUMIN -0.586866 0.004811 -121.973 < 2e-16 ***
## IRON_SAT_PERCENT -0.201464 0.004304 -46.813 < 2e-16 ***
## HGB -0.187944 0.004578 -41.055 < 2e-16 ***
## CLM_FROM_3year_cat 0.283318 0.004297 65.928 < 2e-16 ***
## sex_catMale 0.223189 0.008481 26.317 < 2e-16 ***
## state_catFIPSNE -0.125837 0.013826 -9.102 < 2e-16 ***
## state_catFIPSS -0.069843 0.011099 -6.293 3.12e-10 ***
## state_catFIPSW -0.200278 0.013267 -15.096 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 336409 on 289258 degrees of freedom
## AIC: 336429
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.90 |
1.88, 1.91 |
<0.001 |
ALBUMIN |
0.56 |
0.55, 0.56 |
<0.001 |
IRON_SAT_PERCENT |
0.82 |
0.81, 0.82 |
<0.001 |
HGB |
0.83 |
0.82, 0.84 |
<0.001 |
CLM_FROM_3year_cat |
1.33 |
1.32, 1.34 |
<0.001 |
sex_catMale |
1.25 |
1.23, 1.27 |
<0.001 |
state_catFIPSNE |
0.88 |
0.86, 0.91 |
<0.001 |
state_catFIPSS |
0.93 |
0.91, 0.95 |
<0.001 |
state_catFIPSW |
0.82 |
0.80, 0.84 |
<0.001 |
model_24 : age ALBUMIN IRON_SAT_PERCENT HGB CLM_FROM_3year_cat
sex_cat CALCIUM_UNCORRECTED FERRITIN PHOSPHORUS
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 pr_auc binary 0.706
## # A tibble: 1 × 3
## .metric .estimator .estimate
## <chr> <chr> <dbl>
## 1 brier_class binary 0.404
## Area under the curve: 0.7531
## Confusion Matrix and Statistics
##
## Reference
## Prediction alive died
## alive 9882 1619
## died 30871 29945
##
## Accuracy : 0.5507
## 95% CI : (0.5471, 0.5544)
## No Information Rate : 0.5635
## P-Value [Acc > NIR] : 1
##
## Kappa : 0.1731
##
## Mcnemar's Test P-Value : <2e-16
##
## Sensitivity : 0.9487
## Specificity : 0.2425
## Pos Pred Value : 0.4924
## Neg Pred Value : 0.8592
## Prevalence : 0.4365
## Detection Rate : 0.4141
## Detection Prevalence : 0.8410
## Balanced Accuracy : 0.5956
##
## 'Positive' Class : died
##
##
## Call:
## NULL
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.463442 0.006471 -71.62 <2e-16 ***
## age 0.649787 0.004847 134.06 <2e-16 ***
## ALBUMIN -0.629260 0.005434 -115.79 <2e-16 ***
## IRON_SAT_PERCENT -0.214488 0.004475 -47.93 <2e-16 ***
## HGB -0.183883 0.004644 -39.60 <2e-16 ***
## CLM_FROM_3year_cat 0.285474 0.004293 66.50 <2e-16 ***
## sex_catMale 0.244582 0.008596 28.45 <2e-16 ***
## CALCIUM_UNCORRECTED 0.069892 0.004885 14.31 <2e-16 ***
## FERRITIN 0.051617 0.004528 11.40 <2e-16 ***
## PHOSPHORUS 0.079955 0.004532 17.64 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 395983 on 289267 degrees of freedom
## Residual deviance: 336066 on 289258 degrees of freedom
## AIC: 336086
##
## Number of Fisher Scoring iterations: 4
Characteristic |
OR |
95% CI |
p-value |
age |
1.92 |
1.90, 1.93 |
<0.001 |
ALBUMIN |
0.53 |
0.53, 0.54 |
<0.001 |
IRON_SAT_PERCENT |
0.81 |
0.80, 0.81 |
<0.001 |
HGB |
0.83 |
0.82, 0.84 |
<0.001 |
CLM_FROM_3year_cat |
1.33 |
1.32, 1.34 |
<0.001 |
sex_catMale |
1.28 |
1.26, 1.30 |
<0.001 |
CALCIUM_UNCORRECTED |
1.07 |
1.06, 1.08 |
<0.001 |
FERRITIN |
1.05 |
1.04, 1.06 |
<0.001 |
PHOSPHORUS |
1.08 |
1.07, 1.09 |
<0.001 |
model_12 vs model_13 Population (all patients) vs Cases Only (people
who actually died) vs Controls Only (people who survived)
## New
## Standard < 0.2 >= 0.2
## < 0.2 8772 843
## >= 0.2 1171 61531
## New
## Standard < 0.2 >= 0.2
## < 0.2 1323 206
## >= 0.2 232 29803
## New
## Standard < 0.2 >= 0.2
## < 0.2 7449 637
## >= 0.2 939 31728

## Estimate
## NRI 0.0065867742
## NRI+ -0.0008237232
## NRI- 0.0074104974
## Pr(Up|Case) 0.0065264225
## Pr(Down|Case) 0.0073501457
## Pr(Down|Ctrl) 0.0230412485
## Pr(Up|Ctrl) 0.0156307511
## Estimate Std.Error Lower Upper
## NRI 0.0065867742 0.0012195046 0.004175042 0.0090472221
## NRI+ -0.0008237232 0.0006700932 -0.002021259 0.0004585375
## NRI- 0.0074104974 0.0009846269 0.005527045 0.0093615897
## Pr(Up|Case) 0.0065264225 0.0004588176 0.005670598 0.0074627984
## Pr(Down|Case) 0.0073501457 0.0004620460 0.006457442 0.0082341151
## Pr(Down|Ctrl) 0.0230412485 0.0007518705 0.021484327 0.0245109829
## Pr(Up|Ctrl) 0.0156307511 0.0006110710 0.014461290 0.0168408739
model_12 vs model_14 Population (all patients) vs Cases Only (people
who actually died) vs Controls Only (people who survived)
## New
## Standard < 0.2 >= 0.2
## < 0.2 7941 1674
## >= 0.2 3369 59333
## New
## Standard < 0.2 >= 0.2
## < 0.2 1105 424
## >= 0.2 487 29548
## New
## Standard < 0.2 >= 0.2
## < 0.2 6836 1250
## >= 0.2 2882 29785

## Estimate
## NRI 0.038050187
## NRI+ -0.001995945
## NRI- 0.040046132
## Pr(Up|Case) 0.013433025
## Pr(Down|Case) 0.015428970
## Pr(Down|Ctrl) 0.070718720
## Pr(Up|Ctrl) 0.030672589
## Estimate Std.Error Lower Upper
## NRI 0.038050187 0.0018802403 0.034421187 0.04165934
## NRI+ -0.001995945 0.0009697532 -0.003691554 0.00000000
## NRI- 0.040046132 0.0016440867 0.036748944 0.04310790
## Pr(Up|Case) 0.013433025 0.0006542492 0.012190694 0.01474144
## Pr(Down|Case) 0.015428970 0.0007077955 0.013949445 0.01690004
## Pr(Down|Ctrl) 0.070718720 0.0012856980 0.068127376 0.07313145
## Pr(Up|Ctrl) 0.030672589 0.0008581252 0.029151769 0.03234952
model_13 vs model_14 Population (all patients) vs Cases Only (people
who actually died) vs Controls Only (people who survived)
## New
## Standard < 0.2 >= 0.2
## < 0.2 8262 1681
## >= 0.2 3048 59326
## New
## Standard < 0.2 >= 0.2
## < 0.2 1165 390
## >= 0.2 427 29582
## New
## Standard < 0.2 >= 0.2
## < 0.2 7097 1291
## >= 0.2 2621 29744

## Estimate
## NRI 0.031463413
## NRI+ -0.001172222
## NRI- 0.032635634
## Pr(Up|Case) 0.012355848
## Pr(Down|Case) 0.013528070
## Pr(Down|Ctrl) 0.064314284
## Pr(Up|Ctrl) 0.031678649
## Estimate Std.Error Lower Upper
## NRI 0.031463413 0.0017960120 0.027928895 0.0350974975
## NRI+ -0.001172222 0.0008969822 -0.002970315 0.0005895468
## NRI- 0.032635634 0.0015525923 0.029608526 0.0358313420
## Pr(Up|Case) 0.012355848 0.0006262851 0.011169870 0.0135993258
## Pr(Down|Case) 0.013528070 0.0006592760 0.012281573 0.0148050971
## Pr(Down|Ctrl) 0.064314284 0.0012066297 0.061953604 0.0667864928
## Pr(Up|Ctrl) 0.031678649 0.0008682444 0.029957245 0.0332039333
hoslem.test model_12
## cutyhat Var2 Freq_observed
## 1 [0.00966,0.245] y0 11886
## 2 (0.245,0.362] y0 10130
## 3 (0.362,0.479] y0 8521
## 4 (0.479,0.619] y0 6652
## 5 (0.619,0.984] y0 3564
## 6 [0.00966,0.245] y1 2578
## 7 (0.245,0.362] y1 4333
## 8 (0.362,0.479] y1 5942
## 9 (0.479,0.619] y1 7811
## 10 (0.619,0.984] y1 10900
## cutyhat Var2 Freq_expected
## 1 [0.00966,0.245] yhat0 12061.238
## 2 (0.245,0.362] yhat0 10042.897
## 3 (0.362,0.479] yhat0 8400.802
## 4 (0.479,0.619] yhat0 6593.864
## 5 (0.619,0.984] yhat0 3816.059
## 6 [0.00966,0.245] yhat1 2402.762
## 7 (0.245,0.362] yhat1 4420.103
## 8 (0.362,0.479] yhat1 6062.198
## 9 (0.479,0.619] yhat1 7869.136
## 10 (0.619,0.984] yhat1 10647.941
hoslem.test model_13
## cutyhat Var2 Freq_observed
## 1 [0.008,0.243] y0 11892
## 2 (0.243,0.36] y0 10175
## 3 (0.36,0.476] y0 8566
## 4 (0.476,0.621] y0 6657
## 5 (0.621,0.986] y0 3463
## 6 [0.008,0.243] y1 2572
## 7 (0.243,0.36] y1 4288
## 8 (0.36,0.476] y1 5897
## 9 (0.476,0.621] y1 7806
## 10 (0.621,0.986] y1 11001
## cutyhat Var2 Freq_expected
## 1 [0.008,0.243] yhat0 12099.079
## 2 (0.243,0.36] yhat0 10091.139
## 3 (0.36,0.476] yhat0 8428.969
## 4 (0.476,0.621] yhat0 6580.445
## 5 (0.621,0.986] yhat0 3716.937
## 6 [0.008,0.243] yhat1 2364.921
## 7 (0.243,0.36] yhat1 4371.861
## 8 (0.36,0.476] yhat1 6034.031
## 9 (0.476,0.621] yhat1 7882.555
## 10 (0.621,0.986] yhat1 10747.063
hoslem.test model_14
## cutyhat Var2 Freq_observed
## 1 [0.00565,0.23] y0 12199
## 2 (0.23,0.357] y0 10197
## 3 (0.357,0.482] y0 8442
## 4 (0.482,0.634] y0 6516
## 5 (0.634,0.986] y0 3399
## 6 [0.00565,0.23] y1 2265
## 7 (0.23,0.357] y1 4266
## 8 (0.357,0.482] y1 6021
## 9 (0.482,0.634] y1 7947
## 10 (0.634,0.986] y1 11065
## cutyhat Var2 Freq_expected
## 1 [0.00565,0.23] yhat0 12280.831
## 2 (0.23,0.357] yhat0 10207.233
## 3 (0.357,0.482] yhat0 8404.331
## 4 (0.482,0.634] yhat0 6436.620
## 5 (0.634,0.986] yhat0 3588.504
## 6 [0.00565,0.23] yhat1 2183.169
## 7 (0.23,0.357] yhat1 4255.767
## 8 (0.357,0.482] yhat1 6058.669
## 9 (0.482,0.634] yhat1 8026.380
## 10 (0.634,0.986] yhat1 10875.496