B-245 Effect of early-stage protocolized implementation of a novel host-response test on discharges, hospital free days, and returns at four eds in a health system

Abstract Background There is a major unmet need in emergency departments for diagnostics that aid in rapid sepsis risk assessment of potentially infected patients [1-3]. In this study, we evaluated the results of implementation of a novel host response test (HR) as part of the protocolized screening and guideline-directed treatment pathway process for patients at risk of sepsis, in four geographically distinct EDs within a health system. Methods The HR test, which generates an Index based on the state of immune activation stratified into 3 interpretation bands (Band 1- 3) of increasing sepsis likelihood [4], was integrated into the protocolized response to suspected infection at four EDs in Baton Rouge, LA (S1, 9/24/2024 :1/24/2025), Lafayette, LA (S2, 9/24/2024 :1/24/2025), (Jackson, MS (S3, 8/27:12/27/2024) and Monroe, LA (S4, 8/27: 12/27/2024). At all sites, screening and ordering was achieved through 2 mechanisms, a triage-based nurse-driven protocol, and a later physician-driven protocol. Guideline-directed treatment pathways based on the results of the HR test were adopted by the health system. Patients with missing data for discharge disposition were omitted from this analysis. ED returns were ruled as those who had an ED encounter within 30 days of their original encounter, excluding encounters with discharge ICD-10 diagnosis codes related to mental health (F1-99), substance abuse (F10-19), chronic pain (G89), sickle cell disease (D57), end of life care (Z51.5), injury/poisoning (S1-99,T1-88), or external morbidity (V1-99, Y1-99). Return-adjusted hospital free days was calculated by further subtracting a patient’s length of stay for ed returns within 30 days from the original encounter, or by setting it to zero for mortalities on such returns. Results This study consisted of 4650 patients with 758, 764, 1930, and 1198 from sites 1-4, respectively. Of these 2546 (54.8%), 1194 (25.7%), and 910 (19.6%) returned HR results of Bands 1-3, respectively. From month 1 (M1) to month 4 (M4), a 12.3% and 8.2% increase in rates of ED discharge was observed in Band 1 (M1 22.0%, M4 34.3%, p < 0.01) and Band 2 (M1 10.5%, M4 18.7%, p < 0.05, Fig1A), respectively. During this period, no significant change to rates of ED return was observed across any Band. Additionally, an overall 1 day increase in median return-adjusted hospital free days was observed across the entire cohort from M1 to M4 (M1 26.0, M4 27.0, p < 10-4), with a significant 1-day increase (M1 27.0, M4 28.0, p < 10-4) and 2-day increase (M1 25.0, M4 27.0, p < 0.05, Fig1B) in Bands 1-2, respectively. Conclusion These findings suggest that the addition of a host response test as an aid in risk stratification to protocolized screening and treatment of those presenting to the ED suspected of infection may lead to improved rates of ED discharge and hospital-free days in geographically diverse EDs.