A triage sepsis alert indicator set based on logit analysis

Sepsis is a complication of bacterial, mycotic, or viral infections that can lead to serious consequences and even death if not promptly treated. It affects millions of people in the world every year. Early identification is essential for the patient; accordingly, adequate triage management is required during the initial hours. We propose a new set of prognostic factors for the early identification of the severity of sepsis in emergency departments. Our proposal can improve patient management and survival chances with respect to the actual tests used in emergency units. We utilize a logistic regression model based on a study's mortality rate and survival outcomes based on many variables readily available during the triage. We employ a post-stratified balancing procedure to enhance the predictive capacity of our model.