Development and validation of a new prognostic score system for COVID-19

抄録

This study aimed to develop and validate a bedside risk analysis system for predicting the clinical severity and prognosis of patients with COVID-19. A total of 444 COVID-19 patients were included and were randomly assigned into two groups at a ratio of 2:1 as derivation and validation groups. The new scoring system comprised of eight variables, which included history of having malignant diseases, history of having diabetes mellitus, dyspnea, respiratory rate of >24 bpm, C-reactive protein (CRP) of >14 mg/L, white blood cell count of >8×10⁹/L, platelets count of <180×10¹²/L, and lymphocyte count of <1×10⁹/L. The sensitivity analysis revealed that this new score performed better than the sequential organ failure assessment (SOFA) score at the first day of admission. The receiver characteristic curve analysis revealed that this score predicted severe cases of COVID-19 infection at 0.831 (95% confidence interval: 0.783-0.879) and 0.798 (95% confidence interval: 0.727-0.869) of the area under the curve in the derivation and validation group. The proposed risk score system is a fairly reliable and robust tool for evaluating the severity and prognosis of patients with COVID-19. This may help to early identify severe patients with poor prognosis, who may require more intense interventions.