Skip to main content
. 2020 Oct 9;147(1):99–106.e4. doi: 10.1016/j.jaci.2020.10.001

Fig 3.

Fig 3

CRT analysis algorithms to predict day-30 intubation/mortality in COVID-19 at ED. The algorithms were generated for (A) respiratory rate, (B) sTREM-1, and (C) respiratory rate and sTREM-1. CRT analysis including all biomarkers, vital signs, and clinical scores identified the model including sTREM-1 only (Fig 3, B). CRT was then performed with respiratory rate only (Fig 3, A) and with a combination of respiratory rate and sTREM-1 (Fig 3, C). For all algorithms, the cost of misclassifying a patient who was intubated or died was designated as 10 times the cost of misclassifying a patient who survived without intubation. Cutoff points selected by the CRT analysis are shown between the parent and child nodes. The outcome prediction of the model is indicated below each terminal node.