Background: The management of patients with COVID-19 is challenging for front-line healthcare providers given limited validated, evidence-based clinical decision support. Determining patient mortality risk is critical for effective triage, management, and discharge decision making. Numerous COVID-19 risk prediction models have been created, though the robustness of these models varies. The 4C Mortality Score, created by the International Severe Acute Respiratory and emerging Infections Consortium in Great Britain, is one of the largest-scale, high-performing predictive models published to date. The 4C Mortality Score has been externally validated in cohorts from England, Holland, Italy, Brazil and Spain; however, this model has not yet been validated in a United States (U.S.) population. The purpose of this study is to determine whether the 4C Mortality Score is an accurate predictor of COVID-19 mortality in a United States adult inpatient population.

Methods: This retrospective cohort study included adult patients admitted to a single-center, urban hospital with positive PCR results for SARS-CoV-2 from March to June 2020. The variables included in the 4C Mortality Score [age, sex, select comorbidities, respiratory rate, peripheral oxygen saturation, Glasgow Coma Scale, blood urea nitrogen, and C-reactive protein (CRP)] were collected from patient records. The outcome of interest was mortality during hospital admission or within 30 days of discharge. Twenty imputed datasets were created to account for missing data, using predictive mean matching for numerical variables (e.g., CRP) and logistic regression for binary variables (e.g., obesity). The pooled area under the receiver operator curve (AUC), and corresponding 95% confidence interval, was calculated by pooling the 20 results from each imputation using Rubin’s rule. This research was deemed exempt by the Institutional Review Board (#20E.737).

Results: This study included 426 patients; mean age was 64.4 years and 43.4% were female. Data was missing for less than 1.4% of variables. All-cause mortality was observed in 70 patients (16.4%). The area under the receiver operator characteristic curve of the 4C Mortality Score was 0.85 (95% confidence interval, 0.79-0.89). The 4C Score generally over-predicted mortality, particularly in patients with scores < 15 (out of 21 possible points).

Conclusions: The 4C Mortality Score is valid for use in an urban U.S. inpatient population. The 4C Score over-predicts mortality in patients with relatively lower scores (< 15 of 21). This over-prediction is likely further amplified in current patient populations, due to the improvement in COVID-19 management and treatment options since the early pandemic. This validation supports the use of the 4C Score in U.S. populations; clinicians should apply carefully accounting for the mentioned limitations.