Predictive Modeling for Sepsis Severity: Enhancing Patient Outcomes Through Early Intervention

Sepsis is a serious medical condition that occurs when the body's response to an infection triggers a chain reaction that can lead to widespread inflammation, organ failure, and, in severe cases, death. It is a life-threatening condition that requires immediate medical attention. Early diagnosis and prompt medical intervention significantly improve the chances of recovery. This study aimed to address critical challenges in sepsis management focusing on creating a predictive model to evaluate the severity of sepsis patients accessing the Emergency Room (ER) at our hospital. 
A set of variables have been considered: age, gender, previous infections, previous hospitalization, co-morbidities, previous treatments and therapies, oxygen supplementation, Glasgow Coma Scale (GCS), Site of sepsis origin, systolic and diastolic blood pressure, heart and respiratory rates, temperature, oxygen saturation, National Early Warning Score 2 (NEWS2), and Triage Early Warning Score (TREWS).
AI-based models were able to identify patients’ deaths within 14 days of ER admission and predict patients’ severity condition upon ER entry. 
We believe such models could support clinical choices, predicting the potential challenges in treatment and facilitating more informed and timely interventions for improved patient outcomes.