Table 3.
Comparisons of the predictive performances of the GBDT model with LR model, SOFA, qSOFA, SAPS II, SIRS, and BISAP scores
| Methods | AUC (95%CI) | Delong test | Accuracy (95%CI) | Sensitivity (95%CI) | Specificity (95%CI) | PPV (95%CI) | NPV (95%CI) |
|---|---|---|---|---|---|---|---|
| In the training set | |||||||
| GBDT model | 0.994(0.988-1.000) | Ref | 0.994(0.988-1.000) | 0.996(0.989-1.000) | 0.992(0.982-1.000) | 0.992(0.982-1.000) | 0.996(0.989-1.000) |
| qSOFA | 0.742(0.703–0.780) | < 0.001 | 0.673(0.633–0.713) | 0.479(0.419–0.539) | 0.869(0.828–0.910) | 0.787(0.724–0.851) | 0.623(0.573–0.672) |
| SAPSII | 0.652(0.605–0.698) | < 0.001 | 0.587(0.545–0.629) | 0.574(0.514–0.634) | 0.600(0.540–0.660) | 0.592(0.532–0.652) | 0.582(0.523–0.641) |
| SIRS | 0.592(0.547–0.637) | < 0.001 | 0.562(0.520–0.605) | 0.452(0.392–0.513) | 0.673(0.616–0.730) | 0.583(0.516–0.651) | 0.549(0.494–0.603) |
| SOFA | 0.761(0.720–0.802) | < 0.001 | 0.713(0.674–0.752) | 0.662(0.604–0.719) | 0.765(0.714–0.817) | 0.740(0.684–0.796) | 0.691(0.638–0.744) |
| BISAP | 0.649(0.605–0.694) | < 0.001 | 0.614(0.572–0.655) | 0.654(0.597–0.711) | 0.573(0.513–0.633) | 0.608(0.551–0.665) | 0.621(0.559–0.682) |
| In the testing set | |||||||
| GBDT | 0.985(0.966-1.000) | Ref | 0.969(0.940–0.999) | 1.000(1.000–1.000) | 0.940(0.884–0.997) | 0.941(0.885–0.997) | 1.000(1.000–1.000) |
| qSOFA | 0.780(0.709–0.852) | < 0.001 | 0.710(0.632–0.788) | 0.516(0.393–0.638) | 0.896(0.822–0.969) | 0.825(0.707–0.943) | 0.659(0.562–0.757) |
| SAPSII | 0.625(0.529–0.720) | < 0.001 | 0.557(0.472–0.642) | 0.500(0.378–0.623) | 0.612(0.495–0.729) | 0.552(0.424–0.680) | 0.562(0.448–0.675) |
| SIRS | 0.552(0.461–0.643) | < 0.001 | 0.527(0.441–0.612) | 0.469(0.346–0.591) | 0.582(0.464-0.700) | 0.517(0.389–0.646) | 0.534(0.420–0.649) |
| SOFA | 0.745(0.660–0.829) | < 0.001 | 0.687(0.608–0.766) | 0.641(0.523–0.758) | 0.731(0.625–0.837) | 0.695(0.577–0.812) | 0.681(0.573–0.788) |
| BISAP | 0.566(0.472–0.660) | < 0.001 | 0.542(0.457–0.627) | 0.562(0.441–0.684) | 0.522(0.403–0.642) | 0.529(0.411–0.648) | 0.556(0.433–0.678) |
Notes: GBDT: gradient boosting decision tree; SOFA: Sequential organ failure assessment score; qSOFA: Quick SOFA; SAPSII: Simplified acute physiology score II; SIRS: Systemic inflammatory response syndrome; BISAP: Bedside index of severity in acute pancreatitis; PPV: Positive predictive values; NPV: Negative predictive values; AUC: Area under curve; CI: confidence interval; Ref: Reference