From: A novel reliability-based regression model to analyze and forecast the severity of COVID-19 patients
Author/[Ref.] | Scope | Attributes | Methods | Performance | Size | Country |
---|---|---|---|---|---|---|
Zhang et al. [48] | Severity of COVID-19 | Clinical and laboratory variables | Univariable and multivariable logistic regression models | AUC=0.906 | 80 | China |
Hajiahmadi et al. [16] | ICU and death | CT severity score | Logistic regression model | AUC=0.764 | 192 | Iran |
Homayounieh et al. [18] | ICU and death | Interpretation of radiologists, clinical variables, lung radiomics | Multiple logistic regression model | AUC =0.84 (for ICU admission) | 315 | Iran |
Huang et al. [19] | Severe cases | Clinical and laboratory data | Single-factor and multivariate logistic regression | AUC = 0.985 (95% CI 0.968–1.00) | 125 | China |
Zhou et al. [52] | Severe cases | clinical, laboratory, and CT data | Multivariable logistic regression | AUC =0.952 | 134 | China |
Xiao et al. [44] | Severe illness | Demographic, clinical, laboratory, and radiological data | Univariable and multivariable logistic regression models | AUC= 0.861 (95% CI 0.811–0.902) | 243 | China |
Wei et al. [10] | Common and severe patients | Clinical and CT data | Multivariate logistic regression | AUC=0.95 | 81 | China |
Dong et al. [12] | Survival | Clinical and laboratory findungs | Multivariable Cox regression model | AUC= 0.922 (14 days) AUC= 0.881 (21 days) | 628 | China |
Bai et al. [5] | Severity of disease | Clinical, laboratory, and CT data | Logistic regression model, LDA, SVM, MLP and LSTM | AUC=0.954 | 133 | China |
Al-Najjar and Al-Rousan [2] | Recovered and death cases | Sex, birth year, country, region, group, infection reason, and confirmed date on the outcome | Neural network | Accuracy=0.938 Accuracy=0.995 | 1308 | South Korea |
Li et al. [27] | Severe cases | CT scan data and clinical biochemical attributes | Machine-learning models | AUC =0.93 | 46 | China |
Matos et al. [29] | Mechanical ventilation, death | CT scan and clinical attributes | GLM, PBR, CIT, and SVL | AUC =0.92 | 106 | Italy |
Ning et al. [31] | Negative, mild, and severe cases | CT images and clinical features | CNN, DNNs, and PLR | AUC = 0.944 (negative) AUC = 0.860 (mild) AUC = 0.884 (severe) | 1521 | China |
Zhou et al. [51] | Severe cases | Clinical factors | GA and SVM | Accuracy: over 0.94 Accuracy= 0.80 | 144 25 | China |
Yan et al. [46] | Survival for severe cases | Clinical data | XGBoost algorithm | Accuracy=0.93 | 375 | China |
Shi et al. [37] | Severe cases | Clinical and radiological findings | LASSO logistic regression | AUC= 0.890 | 196 | China |
Bi et al. [6] | Severe illness | Fibrinogen-to-albumin ratio (FAR) and platelet count (PLT) | Multivariate cox analysis | AUC=0.754 | 113 | China |
Zhou et al. [53] | Severe cases | Body temperature, cough, dyspnea, hypertension, cardiovascular disease, chronic liver disease, and chronic kidney disease | Multivariable logistic regression | AUC= 0.862 (95% CI 0.801–0.925) | 366 | China |
Cheng et al. [9] | ICU transfer | Signs, nursing assessments, laboratory features and electrocardiograms | Random forest | AUC= 0.799 (95% CI 0.752–0.846) | 1987 | USA |
McRae et al. [30] | Death | CRP, NT-proBNP, MYO, D-dimer, PCT, CK-MB, cTnI | Logistic regression model | AUC= 0.94 (95% CI 0.89–0.99) | 160 | China |