Development and Validation of a Two-Step Predictive Risk Stratification Model for Coronavirus Disease 2019 In-hospital Mortality: A Multicenter Retrospective Cohort Study

• DOI: 10.3389/fmed.2022.827261
• Publication Date: 2022-04-07T04:50:28Z
• Authors: Yang Li; Yanlei Kong; Mark H. Ebell; Leonardo Martinez; Xinyan Cai; Robert P. Lennon; Derjung M. Tarn; Arch G. Mainous; Aleksandra E. Zgi...; Bruce Barrett; Wen-Jan Tuan; Kevin Maloy; Munish Goyal; Alex H. Krist; Tamas S. Gal; Meng-Hsuan Sung; Changwei Li; Yier Jin; Ye Shen
• Keywords: Triage; Risk of mortality

An accurate prognostic score to predict mortality for adults with COVID-19 infection is needed understand who would benefit most from hospitalizations and more intensive support care. We aimed develop validate a two-step system patient triage, identify patients at relatively low level of risk using easy-to-collect individual information.Multicenter retrospective observational cohort study.Four health centers Virginia Commonwealth University, Georgetown the University Florida, California, Los Angeles.Coronavirus Disease 2019-confirmed hospitalized adult patients.We included 1,673 participants (VCU) as derivation cohort. Risk factors in-hospital death were identified multivariable logistic model variable selection procedures after repeated missing data imputation. A was developed lower, moderate, higher risk. The first step selected increasing age, than one pre-existing comorbidities, heart rate >100 beats/min, respiratory ≥30 breaths/min, SpO2 <93% into predictive model. Besides age SpO2, second used blood urea nitrogen, absolute neutrophil count, C-reactive protein, platelet neutrophil-to-lymphocyte ratio predictors. C-statistics reflected very good discrimination internal validation VCU (0.83, 95% CI 0.79-0.88) external other three systems (range, 0.79-0.85). one-step also derived comparison. Overall, had better performance score.The scoring widely available, point-of-care triage potentially time- cost-saving tool in practice.