Risk score for mortality due to COVID-19: a prospective temporal validation cohort study in the emergency department of a tertiary care hospital

ABSTRACT

Objective. To validate a previously described hospital emergency department risk model to predict mortality in patients with COVID-19. Methods. Prospective observational noninterventional study. Patients aged over 18 years diagnosed with COVID-19 were included between December 1, 2020, and February 28, 2021. We calculated a risk score for each patient based on age >= 50 years (2 points) plus 1 point each for the presence of the following predictors Barthel index <90 points, altered level of consciousness, ratio of arterial oxygen saturation to fraction of inspired oxygen <400, abnormal breath sounds, platelet concentration <100 x 10(9)/L, C reactive protein level >= 5 mg/dL, and glomerular filtration rate <45 ml/min. The dependent variable was 30-day mortality. We assessed the score's performance with the area under the receiver operating characteristic curve (AUC). Results. The validation cohort included 1223 patients. After a median follow-up of 80 days, 143 patients had died;901 patients were classified as having low risk (score, <= 4 points), 270 as intermediate risk (5-6 points), and 52 as high risk (>= 7 points). Thirty-day mortality rates at each risk level were 2.8%, 22.5%, and 65.4%, respectively. The AUC for the score was 0.883;for risk categorization, the AUC was 0.818. Conclusion. The risk score described is useful for stratifying risk for mortality in patients with COVID-19 who come to a tertiary-care hospital emergency department.