Westmead Hospital in New South Wales is piloting a new AI-powered clinical decision support tool for assessing the risk of sepsis in ED waiting rooms.

WHAT IT'S ABOUT

The Sepsis Risk Tool Dashboard provides a risk score of patients with sepsis using an algorithm trained on historic patient data. It also combines a patient's age, gender, and vitals as they are entered into the EMR. 

The AI tool was developed by eHealth NSW in collaboration with the Clinical Excellence Commission (CEC), Sydney Health Partners, the University of Sydney and NSW Health Pathology. 

It is being introduced to complement the existing Sepsis Kills programme at Westmead, which seeks to support clinicians in recognising and treating the infection.

The clinical pilot will run for six months after which eHealth NSW will decide if the support tool can be scaled statewide.

WHY IT MATTERS

According to eHealth NSW, a sepsis diagnosis for an estimated one in three patients is delayed or missed in the ED. "Sepsis can be difficult to diagnose in ED and even harder for junior doctors. Delays in diagnosing and treating sepsis can result in patient deaths," explained Dr Amith Shetty, senior staff specialist at Westmead Hospital and clinical director of NSW Health.

He also noted that while there are existing sepsis detection algorithms, there is none which focus on the ED waiting room where sepsis is most likely undiscovered.

The early detection of deterioration due to sepsis can be "lifesaving" if treatments are also administered earlier in the patient journey, claimed Dr Margaret Murphy, a clinical nurse consultant at Westmead Hospital's Emergency Services.

THE LARGER TREND

Last year, Queensland Health also tested eHealth NSW's AI tool for detecting sepsis at Townsville Hospital and Health Service. In partnership with the CEC, the project used the predictive algorithm to help clinicians identify potential sepsis cases from EMR data.

In 2020, eHealth NSW piloted an EMR module, the Adult Sepsis Pathway, at Prince of Wales Hospital. This tool also assists clinicians in the early detection, escalation and treatment of patients with suspected sepsis.