765: Sepsis Alert Fatigue and Crowding: Impact on Diagnosis and Antibiotic Timing

Introduction: Septic shock is a leading cause of hospital mortality with fatality rates up to 34%. Early recognition and antibiotics are critical, yet responsiveness to electronic health record (EHR) sepsis alerts varies. Understanding how provider experience, workload, and alert engagement affect diagnosis and treatment is essential. Study findings aim to improve septic shock care. Methods: We performed a retrospective chart review of Emergency Department (ED) patients diagnosed with septic shock (ICD-10) during admission from 01/01/2025 - 03/31/2025 who triggered an EHR alert defined by hypotension (SBP < 90 or MAP < 65) or ≥2 SIRS criteria plus end-organ damage. We examined relationships among provider characteristics (training level, years of experience), EHR alert acknowledgment, ED sepsis diagnosis, antibiotic timing, and ED overcrowding measured by the National ED Overcrowding Scale (NEDOCS). NEDOCS >140 indicates a severely busy ED and >180 indicates disaster. Results: Among 116 admitted septic shock patients, 84% of clinicians acknowledged the alert in the EHR. Alert responsiveness declined with resident seniority (PGY-1 100%, PGY-2 93%, PGY-3 78%, PGY-4 76%) but remained stable (~80%) among attendings regardless of experience. The median NEDOCS was 174; higher crowding (NEDOCS 200) corresponded with lower alert responsiveness compared to less busy periods (NEDOCS 173). Busy clinicians were less likely to list sepsis as an ED diagnosis (NEDOCS 191 undiagnosed vs. 166 diagnosed). Diagnosis of sepsis was associated with faster antibiotic ordering (median 42 vs. 96 minutes). Interestingly, antibiotic ordering within 1 hour of sepsis alert occurred with higher NEDOCS scores, while antibiotic infusion did not (ordered: avg. 175 < 1 hr vs. avg. 163 > 1 hr; given: avg. 167 < 1 hr vs. avg. 176 > 1 hr), suggesting overcrowding delays timely antibiotic delivery more than ordering. Conclusions: Crowding in the ED results in a decrease in responsiveness to sepsis alerts and diagnosis rates. While crowding does not delay antibiotic ordering, it is associated with delayed infusion. This hypothesis-generating data underscores the need to address environmental factors in targeted interventions. Future studies will explore ED and ICU teams and settings to improve timely antibiotics in septic shock management.