Development, implementation and evaluation of an early recognition tool for sepsis from triage in an emergency department

Abstract

Introduction: Sepsis detection tools in the triage of Pediatric Emergency Departments (PED) enable timely alerts and treatment.

Objetive: To develop, implement, and evaluate a sepsis screening tool applied during triage in patients with fever or hypothermia in the PED.

Methods: A three-phase study (design, implementation, and evaluation) conducted from 07/01/2023 to 07/31/2024. Variables: pediatric assessment triangle (PAT), parental concern, critical appearance, vital signs, mental status, peripheral perfusion, and risk factors. The tool was classified as positive or negative. Evaluation design: observational, prospective, diagnostic test study using three reference standards: sepsis by intention to treat, sepsis by Phoenix criteria, and sepsis as the final diagnosis. Patients >28 days and <18 years with fever or hypothermia were included. Results were expressed as sensitivity, specificity, predictive values, and likelihood ratios. Positive tools in groups with and without sepsis were compared using logistic regression.

Results: A total of 40,005 patients were seen, 16,771 of whom had fever or hypothermia. The tool was positive in 1.4% (240/16,672). These patients showed higher triage levels I and II (99.6% vs. 4.8%), hospitalization (66% vs. 5%), PICU admission (11% vs. 0.03%), and mortality (3.8% vs. 0.01%). The tool’s performance for sepsis as the final diagnosis was: S: 0.80, E: 0.99, PPV: 0.33, NPV: 1, LR(+): 84, and LR(-): 0.20. The risk of sepsis increased in patients with a positive tool and circulatory alteration in the PAT (OR 2.8, p=0.008), capillary refill (OR 1.9, p=0.045), and a critical risk factor (OR 2.1, p=0.044).

 Conclusions: The sepsis screening tool demonstrated good performance in detecting sepsis in the PED. Circulatory alteration in the PAT, capillary refill, and the presence of a critical risk factor were key.