The Impact of Inpatient Nurse-Triggered Sepsis Alert on Antimicrobial Utilization

ABSTRACT Introduction A nurse-triggered sepsis alert called “Code Sepsis” was implemented for early recognition and management of sepsis. We analyzed its impact on antimicrobial use and identified factors associated with infection as source of Code Sepsis. Methods We reviewed the medical records of hospitalized patients with Code Sepsis between January 1 and June 30, 2018. Patients were classified as “Infection” when probable or definitive infection was identified or “No Infection” when probable or definitive non-infectious source was identified. Patients were categorized as “Escalation” with addition or change to broader-spectrum antimicrobials or “No Escalation” with no change or change to narrower-spectrum antimicrobials. Escalation was classified as “Indicated” with appropriate escalation or “Not Indicated” with inappropriate escalation. Logistic regression model was used to identify factors associated with Infection as Code Sepsis trigger. Results Code Sepsis was activated in 529 patients with Escalation in 246 (47%) and No Escalation in 283 (53%) patients. Escalation was Indicated in 157 (64%) and Not Indicated in 89 (36%) patients. Infection was identified in 356 (67%) and No Infection in 173 (33%) patients. History of HIV (OR 2.75, p=0.03), temperature >38.3°C or 20/minute (OR 1.56, p=0.02) were associated with Infection while surgery within 3 days (OR 0.30, p Conclusion Our Code Sepsis inadvertently identified patients without infections and led to antimicrobial overutilization. By refocusing Code Sepsis on early recognition of severe sepsis and septic shock only, we hope to optimize resource utilization and improve patient outcomes.