A Clinical Risk Score to Predict In-hospital Mortality from COVID-19 in South Korea
10.3346/jkms.2021.36.e108
- Author:
Ae-Young HER
1
;
Youngjune BHAK
;
Eun Jung JUN
;
Song Lin YUAN
;
Scot GARG
;
Semin LEE
;
Jong BHAK
;
Eun-Seok SHIN
Author Information
1. Division of Cardiology, Department of Internal Medicine, Kangwon National University School of Medicine, Chuncheon, Korea
- Publication Type:Original Article
- From:Journal of Korean Medical Science
2021;36(15):e108-
- CountryRepublic of Korea
- Language:English
-
Abstract:
Background:Early identification of patients with coronavirus disease 2019 (COVID-19) who are at high risk of mortality is of vital importance for appropriate clinical decision making and delivering optimal treatment. We aimed to develop and validate a clinical risk score for predicting mortality at the time of admission of patients hospitalized with COVID-19.
Methods:Collaborating with the Korea Centers for Disease Control and Prevention (KCDC), we established a prospective consecutive cohort of 5,628 patients with confirmed COVID-19 infection who were admitted to 120 hospitals in Korea between January 20, 2020, and April 30, 2020. The cohort was randomly divided using a 7:3 ratio into a development (n = 3,940) and validation (n = 1,688) set. Clinical information and complete blood count (CBC) detected at admission were investigated using Least Absolute Shrinkage and Selection Operator (LASSO) and logistic regression to construct a predictive risk score (COVID-Mortality Score).The discriminative power of the risk model was assessed by calculating the area under the curve (AUC) of the receiver operating characteristic curves.
Results:The incidence of mortality was 4.3% in both the development and validation set.A COVID-Mortality Score consisting of age, sex, body mass index, combined comorbidity, clinical symptoms, and CBC was developed. AUCs of the scoring system were 0.96 (95% confidence interval [CI], 0.85–0.91) and 0.97 (95% CI, 0.84–0.93) in the development and validation set, respectively. If the model was optimized for > 90% sensitivity, accuracies were 81.0% and 80.2% with sensitivities of 91.7% and 86.1% in the development and validation set, respectively. The optimized scoring system has been applied to the public online risk calculator (https://www.diseaseriskscore.com).
Conclusion:This clinically developed and validated COVID-Mortality Score, using clinical data available at the time of admission, will aid clinicians in predicting in-hospital mortality.
