Early predictors and screening tool developing for severe patients with COVID-19

Background Coronavirus disease 2019 (COVID-19) is a declared global pandemic, causing a lot of death. How to quickly screen risk population for severe patients is essential for decreasing the mortality. Many of the predictors might not be available in all hospitals, so it is necessary to develop a simpler screening tool with predictors which can be easily obtained for wide wise. Methods This retrospective study included all the 813 confirmed cases diagnosed with COVID-19 before March 2nd, 2020 in a city of Hubei Province in China. Data of the COVID-19 patients including clinical and epidemiological features were collected through Chinese Disease Control and Prevention Information System. Predictors were selected by logistic regression, and then categorized to four different level risk factors. A screening tool for severe patient with COVID-19 was developed and tested by ROC curve. Results Seven early predictors for severe patients with COVID-19 were selected, including chronic kidney disease (OR 14.7), age above 60 (OR 5.6), lymphocyte count less than < 0.8 x 10(9) per L (OR 2.5), Neutrophil to Lymphocyte Ratio larger than 4.7 (OR 2.2), high fever with temperature >= 38.5celcius (OR 2.2), male (OR 2.2), cardiovascular related diseases (OR 2.0). The Area Under the ROC Curve of the screening tool developed by above seven predictors was 0.798 (95% CI 0.747-0.849), and its best cut-off value is > 4.5, with sensitivity 72.0% and specificity 75.3%. Conclusions This newly developed screening tool can be a good choice for early prediction and alert for severe case especially in the condition of overload health service.

Automated summary

A screening tool for severe patients with COVID-19 was developed in a retrospective study, with seven early predictors identified such as chronic kidney disease and age above 60, showing high sensitivity and specificity. The tool can be a good choice for early prediction and alert for severe cases, especially in the condition of overload health service.