AI early-warning system for sepsis improves survival rates

AI-based early warning system for sepsis

Date: 25th August 2021

Article in brief

Sepsis is a life-threatening condition caused by a dysregulated host response to infection which can lead to the damage of multiple organ systems. In 2017, there were an estimated 48.9 million cases of sepsis recorded worldwide and a staggering 11 million sepsis-related deaths. It is extremely challenging to detect and treat as the heterogeneous features of the host and pathogen interactions create a complex and dynamic disease pathology.  Prompt treatment is critical for sepsis, as the condition can rapidly deteriorate, often leaving survivors with long-term damage and disabilities. Now, researchers have developed an AI-based early warning system that alerted both the physicians and pharmacists of flagged patients which then receiving antibiotics significantly faster than standard care patients, resulting in an increase in days alive and out of hospital.

Researchers and physicians at Case Western Reserve University and MetroHealth Medical Center, US, led by the study’s principle investigator Yasir Tarabichi, have reported the first published randomised controlled evaluation of a model-based early warning system in an emergency room setting.

The team randomised a total of 598 patients (aged 18 or over) to standard sepsis care or standard care augmented by the display of a sepsis early warning system–triggered flag in the electronic health record combined with electronic health record–based emergency department pharmacist notification.  Over a 5-month period, time to antibiotic administration, and clinical outcomes were measured as days alive and out of hospital at 28 days.

The team found that those in the augmented cohort were associated with shorter times to antibiotic administration without an increase in undesirable or potentially harmful clinical interventions.  They had, on average, more days alive and out of hospital than the group that received the standard care. Taken together, the increase in survival rates and reduction in hospital stay improved with the implementation of the early warning sepsis system.

Conclusions and future applications

The team here have demonstrated for the first time that an AI-based early warning system for sepsis improved outcomes in a ‘real’ world setting – an emergency room.  The system improved patient outcomes, leading to quicker administration of antibiotics, and more days alive and out of hospital.  It is envisioned that the early warning system’s role will be supportive to the health care’s response to sepsis.

Early warning systems such as these are likely to be a welcome inclusion to help determine sepsis status. Time to treatment is crucial and currently other novel tools are also being developed to alleviate and treat symptoms.  New nanotrap technology has been developed to efficiently capture and attenuate a group of inflammatory mediators for effective sepsis treatment in animal models.  Together, these types of ground breaking technologies could change the way we diagnose and treat sepsis, and offer faster administration of treatments which will attenuate the devastating long lasting and devastating effects that sepsis can have.


For more information please see the press release at Cape Western Reserve University


Tarabichi, Y., Cheng, A., Bar-Shain, D., McCrate, B.M., Reese, L.H., Emerman, C., Siff, J., Wang, C., Kaelber, D.C., Watts, B., et al. (2021). Improving Timeliness of Antibiotic Administration Using a Provider and Pharmacist Facing Sepsis Early Warning System in the Emergency Department Setting: A Randomized Controlled Quality Improvement Initiative. Critical Care Medicine.