Development and validation of an online model to predict critical COVID-19 with immune-inflammatory parameters

Gao, Yue; Chen, Lingxi; Chi, Jianhua; Zeng, Shaoqing; Feng, Xikang; Li, Huayi; Liu, Dan; Feng, Xinxia; Wang, Siyuan; Wang, Ya; Yu, Ruidi; Yuan, Yuan; Xu, Sen; Li, Chunrui; Zhang, Wei; Li, Shuaicheng; Gao, Qinglei

doi:10.1186/s40560-021-00531-1

Journal of Intensive Care

Table 2 Performance metrics for mortality risk prediction of models in cohorts

From: Development and validation of an online model to predict critical COVID-19 with immune-inflammatory parameters

	AUC (95% CI)	Accuracy (95% CI)	SN (95% CI)	SP (95% CI)	PPV (95% CI)	NPV (95% CI)	Kappa	F1	Brier
Internal validation cohort
LR	0.965 (0.9-1)	0.982 (0.936-0.998)	0.875 (0.617-0.985)	1 (0.962-1)	1 (0.768-1)	0.979 (0.928-0.998)	0.923	0.933	0.018
SVM	0.962 (0.898-1)	0.982 (0.936-0.998)	0.938 (0.698-0.998)	0.990 (0.943-1)	0.938 (0.698-0.998)	0.990 (0.943-1)	0.927	0.938	0.018
GBDT	0.956 (0.891-1)	0.955 (0.898-0.985)	0.875 (0.617-0.985)	0.968 (0.911-0.993)	0.824 (0.566-0.962)	0.979 (0.925-0.997)	0.822	0.849	0.045
KNN	0.964 (0.899-1)	0.973 (0.923-0.994)	0.938 (0.698-0.998)	0.979 (0.926-0.997)	0.882 (0.636-0.985)	0.989 (0.942-1)	0.893	0.909	0.024
NN	0.964 (0.899-1)	0.973 (0.923-0.994)	0.938 (0.698-0.998)	0.979 (0.926-0.997)	0.882 (0.636-0.985)	0.989 (0.942-1)	0.893	0.909	0.027
SPMCIIP	0.991 (0.979-1)	0.964 (0.910-0.990)	0.875 (0.617-0.985)	0.979 (0.926-0.997)	0.875 (0.617-0.985)	0.979 (0.926-0.997)	0.854	0.875	0.025
External validation cohort
LR	0.998 (0.995-1)	0.978 (0.950-0.993)	0.8 (0.593-0.932)	1 (0.982-1)	1 (0.832-1)	0.976 (0.945-0.992)	0.877	0.889	0.022
SVM	0.999 (0.998-1)	0.987 (0.962-0.997)	0.92 (0.740-0.990)	0.995 (0.973-1)	0.958 (0.789-0.999)	0.990 (0.965-0.999)	0.931	0.939	0.011
GBDT	0.998 (0.995-1)	0.987 (0.962-0.997)	0.96 (0.797-0.999)	0.990 (0.965-0.999)	0.923 (0.749-0.991)	0.995 (0.973-1)	0.934	0.941	0.013
KNN	0.978 (0.939-1)	0.987 (0.962-0.997)	0.96 (0.797-0.999)	0.990 (0.965-0.999)	0.923 (0.749-0.991)	0.995 (0.973-1)	0.934	0.941	0.012
NN	0.999 (0.998-1)	0.991 (0.969-0.999)	0.96 (0.797-0.999)	0.995 (0.973-1)	0.96 (0.797-0.999)	0.995 (0.973-1)	0.955	0.960	0.009
SPMCIIP	0.999 (0.998-1)	0.991 (0.969-0.999)	0.96 (0.797-0.999)	0.995 (0.973-1)	0.96 (0.797-0.999)	0.995 (0.973-1)	0.955	0.960	0.007

LR logistic regression, SVM supported vector machine, GBDT gradient boosted decision tree, KNN k-nearest neighbor, NN neural network, SPMCIIP severity prediction model for COVID-19 by immune-inflammatory parameters, AUC area under the curve, SN sensitivity, SP specificity, PPV positive predictive value, NPV negative predictive value, CI confidence interval

Back to article page

ISSN: 2052-0492

Contact us

Submission enquiries: Access here and click Contact Us
General enquiries: info@biomedcentral.com