Skip to main content

Table 4 Prediction models fit to UIH test cohort

From: Predicting clinical outcomes among hospitalized COVID-19 patients using both local and published models

Study Cohort origin N (develop) Outcome Method Covariates Performance Test cohort # evaluable Test cohort performance
A] An interpretable mortality prediction model for COVID-19 patients [42] Wuhan, China 351 Mortality* Decision tree (3) CRP, LDH, lymphocyte percentage PPV 96.9%
NPV 98.4%
145 (70%)*** AUC 0.69 (0.60–0.79)
B] Development and validation of a clinical risk score to predict the occurrence of critical illness of hospitalized patients with COVID-19 [19]a China 1590 Criticality** Logistic regression (LR) (10) Age, cancer, direct bilirubin, comorbidities, dyspnea, hemoptysis, LDH, neutrophils/lymphocytes, unconscious, CXR AUC 0.88 (0.84–0.93) 144 (70%) AUC 0.84 (0.78–0.91)
C] Development and validation of prognosis model of mortality risk in patients with COVID-19 [20] Wuhan, China 292 Mortality LR (3) Age, LDH, CRP AUC 0.95 141 (69%) AUC 0.89 (0.82–0.96)
D] Diagnostic performance of initial blood urea nitrogen combined with D-dimer levels for predicting in-hospital mortality in COVID-19 patients [26] Wuhan, China 305 Mortality LR (2) BUN, D-dimer AUC 0.94 (0.90–0.97) 150 (72%) AUC 0.73 (0.62–0.83)
E] Laboratory findings and a combined multifactorial approach to predict death in critically ill patients with COVID-19: a retrospective study [28] Wuhan, China 336 Mortality LR (3) BUN, D-dimer, lymphocyte percentage AUC 0.99 (0.98–1.0) 148 (71%)*** AUC 0.72 (0.61–0.82)
F] Development and external validation of a prognostic multivariable model on admission for hospitalized patients with COVID-19 [44] Wuhan, China 299 Mortality LR (4) Age, LDH, lymphocyte count, O2 Saturation AUC 0.98 (0.96–1.0) 150 (72%) AUC 0.84 (0.73–0.94)
G] Early prediction of mortality risk among severe COVID-19 patients using machine learning [46]b Wuhan, China 183 Mortality LR (4) Age, CRP, D-dimer, lymphocyte count AUC 0.90 142 (69%) AUC 0.68 (0.51–0.81)
H] Risk prediction for poor outcome and death in hospital in-patients with COVID19: derivation in Wuhan, China and external validation in London, UK [(51)]c Wuhan, China 775 Mortality LR (7) Age, CRP, sex, Cr, lymphocytes, neutrophils, platelets count AUC 0.91 165 (80%) AUC 0.72 (0.58–0.86)
UIH mortality model Chicago, USA 309 Mortality Random forest (11) Age, AST, BMI, Cr, CRP, diastolic BP, ferritin, O2 saturation, platelet count, RDW, WBC AUC 0.98 (0.96–1.0) 152 (73%) AUC 0.84 (0.74–0.94)
UIH criticality model Chicago, USA 309 Criticality Random forest (11) Age, ALT, AST, Cr, CRP, ferritin, RDW, neutrophils/lymphocytes, O2 saturation, platelet count, WBC AUC 0.97 (0.94–1.0) 152 (73%) AUC 0.83 (0.76–0.90)
  1. ALT, alanine aminotransferase; AST, aspartate transaminase; BMI, body mass index; BP, blood pressure; BUN, blood urea nitrogen; Cr, creatinine; CRP, C reactive peptide; CXR, chest Xray; LDH, lactate dehydrogenase; O2, oxygen; PC, procalcitonin; RDW, red blood cell distribution width; SAT, oxygen saturation; WBC, white blood cell count
  2. *Mortality defined as death prior to discharge**Criticality defined as mortality or intensive care unit stay
  3. ***For these models only non-pregnant patients were used. The other models either included or did not specify inclusion of pregnant patients
  4. ahttp://118.126.104.170/
  5. bhttps://phenomics.fudan.edu.cn/risk_scores/
  6. chttps://covid.datahelps.life/prediction/