B14-08 Impact of Prehospital Sepsis Recognition on Timely Antibiotic Administration and Subsequent Adverse Events: The Impress Cluster-randomized, Stepped-wedge Trial

Rationale

Surviving Sepsis Campaign Guidelines strongly recommend standardized sepsis screening in hospitals. The impact of expanding screening to include the prehospital-ambulance setting is unknown. This study evaluated whether an ambulance-based sepsis screening protocol decreases time to antibiotics among patients with sepsis without increasing the use of unnecessary antibiotics in patients without sepsis.

Methods

We performed a cluster-randomized, stepped-wedge implementation trial of a validated prehospital sepsis screening protocol in 3 geographically-distinct United States ambulance systems. Clustering was defined at the level of the ambulance system. Study eligibility criteria included all adult, ambulance patients with a positive sepsis screen that included a combination of abnormal vital signs and patient risk factors. Prehospital providers were trained to screen for sepsis and send a hospital prearrival notification for positive screens. The primary outcome was time from hospital arrival to first antibiotic administration in the emergency department among patients with sepsis. Sepsis was defined using modified Centers for Disease Control Adult Sepsis Event surveillance criteria. Secondary outcomes included the proportion of sepsis cases identified by prehospital personnel, prearrival notifications delivered, and days of hospital antibiotic therapy among false positive sepsis screens.

Results

Among 258,075 ambulance encounters, 984 (0.4%) met study eligibility criteria and were included, of which 201 (20%) had sepsis; 106 in the intervention and 95 in the control. The unadjusted, median time to first antibiotic administration was 104 minutes (IQR, 15-531 minutes) in the intervention group and 101 minutes (IQR, 17-396 minutes) in the control group (P = 0.99). Prespecified modeling analysis showed antibiotics were administered an estimated 6.6 minutes faster per month in the intervention group compared to control (95% CI, 0.51 to 12.73). Sepsis was identified more often by prehospital personnel during the intervention compared to control [50% (53/106) versus 28% (27/95), respectively (P = 0.01)]. Patients with sepsis received more prearrival notifications during the intervention compared to control [43% (45/106) versus 10% (9/95), respectively (P < 0.001)]. Among patients without sepsis, there was no difference in mean days of antibiotic therapy in the intervention compared to control [3.2 days (SD, 1.9) versus 3.3 days (SD, 2.2), respectively (P = 0.99)].

Conclusions

Among acutely-ill adults in ambulances, implementation of a sepsis screening protocol led to higher sepsis recognition rates and more prearrival notifications. While unadjusted time to antibiotics did not change, adjusted analysis showed a decrease in time to antibiotics by an estimated 6.6 minutes per month. There was no difference in antibiotic exposure among patients without sepsis. NCT05502107

This abstract is funded by: Centers for Disease Control