Score identifies ED patients with COVID-19 who will end up in ICU
An 11-point score based on confusion, respiratory rate, oxygen saturation, and oxygen supplementation performed better than existing tools in predicting need for intensive care within seven days of ED presentation with COVID-19, a German study found.
A risk score based on confusion, respiratory rate, oxygen saturation, and oxygen supplementation may help triage patients who present to the ED with COVID-19, a recent study found.
Researchers in Germany analyzed nationwide data from a cohort of patients who had COVID-19 in March 2020 to January 2023 to select variables for a risk score to predict ICU admission within seven days of ED presentation. The score was developed using a northern German cohort, evaluated with a southern German cohort, and externally validated on a Colombian cohort. The score's discrimination, calibration, and clinical utility were compared with existing risk prediction tools. The results were published Jan. 10 by Clinical Infectious Diseases.
The 11-point risk score was based on Confusion, Respiratory rate, Oxygen Saturation (with or without concurrent supplemental oxygen), and oxygen Supplementation (CROSS). The derivation cohort included 1,295 patients (median age, 60 years; 38.1% female), the evaluation cohort included 1,123 patients (median age, 58 years; 36.9% female), and the external validation cohort included 780 patients (median age, 57 years; 38.8% female). ICU admission rates within seven days were 30.8%, 28.1%, and 30.3%, respectively, overall.
In the evaluation cohort, ICU admission occurred in 9.5% of the low-risk group, 19.9% of the medium-risk group, and 58.7% of the high-risk group. The CROSS score demonstrated good discrimination, with an area under the curve of 0.77 in the evaluation cohort and 0.69 in the external validation cohort. Calibration was also good, and clinical utility was superior versus 15 existing risk tools. Tools that predicted mortality did not perform well when used to predict ICU admission risk in patients with COVID-19, the analysis found.
The researchers noted that selection bias was possible and that their score was developed using only patients in Germany. They concluded that the CROSS score, which does not require a calculator to use, effectively predicted ICU admission for ED patients with COVID-19, with those scoring zero eligible to receive regular monitoring on the wards and those scoring five or higher considered as candidates for early ICU admission.
“Management of patients with intermediate scores may depend on local conditions and resource availability, potentially prioritizing regular ward admission with intensified monitoring if ICU capacity is limited,” they wrote. “Further studies are needed to determine the generalizability of the CROSS score in other settings.”
