From: Comparing machine learning algorithms for predicting COVID-19 mortality
Classes | Number of suggested features | Delphi round | Final features | Included features | Excluded features | |
---|---|---|---|---|---|---|
< 75% | 75% < | |||||
Demographic | 9 | 6 | 3 | 3 | Gender, age, length of hospitalization | Body mass index, blood group, marital status, ethnicity, place of birth, level of education |
Risk factors | 10 | 2 | 7 | 7 | Smoking, ICU admission, hypertension, pneumonia, diabetes, cardiac disease, another underline disease | Recent travel, exposure type |
Clinical manifestations | 23 | 9 | 14 | 14 | Dyspnea, sore throat, runny nose, loss of taste, loss of smell, contusion, muscular pain, chill, fever, cough, nausea/ vomiting, chest pain and pressure, headache, gastrointestinal symptoms | Weakness, sneezing, exudative pharyngitis, mucus or phlegm, conjunctivitis, hemoptysis, anorexia, dry mouth, decrease consciousness |
Laboratory results | 24 | 11 | 13 | 13 | Blood urea nitrogen, white cell count, C-reactive protein, hypersensitive troponin, glucose, erythrocyte sedimentation rate, creatinine, alkaline phosphatase, aspartate aminotransferase, alanine aminotransferase, absolute lymphocyte count, absolute neutrophil count, | Hematocrit, red cell count, hemoglobin, total bilirubin, thromboplastin time, prothrombin time, albumin calcium, phosphorus, magnesium, sodium, potassium |
Therapeutic plan | 1 | 0 | 1 | 1 | Oxygen therapy | Â |