AI Algorithm Reduces Sepsis Mortality by 17% in UC San Diego Emergency Departments

In a groundbreaking development, a recent study conducted by researchers at the University of California San Diego School of Medicine has revealed the remarkable impact of artificial intelligence (AI) in identifying patients at risk of sepsis infection. 

The study, published in the January 23, 2024, online edition of npj Digital Medicine, introduced an AI model called COMPOSER, leading to a significant 17% reduction in patient mortality rates.

Continuous surveillance with COMPOSER

The COMPOSER AI model, designed and developed by the research team, operates quietly and effectively behind the scenes in the emergency departments at UC San Diego Health. 

Upon arrival, the algorithm begins monitoring over 150 patient variables, including lab results, vital signs, current medications, demographics, and medical history, which could indicate sepsis risk.

Should the AI identify multiple risk factors, signaling a high potential for sepsis infection, it immediately alerts the nursing staff through the hospital’s electronic health record system. 

The nursing team collaborates with physicians to determine the most suitable treatment plans for these patients, potentially saving lives by initiating early interventions.

Uncovering hidden patterns

One of the key strengths of the COMPOSER AI model is its ability to detect complex and subtle risk patterns that might escape the human eye. By analyzing these risk factors, the AI can provide highly accurate predictions of sepsis. Conversely, if the risk patterns can be attributed to other conditions with higher confidence, no alerts are generated, ensuring a focus on true sepsis cases.

The study conducted by UC San Diego examined the impact of COMPOSER on over 6,000 patient admissions before and after its deployment in the emergency departments at UC San Diego Medical Center in Hillcrest and Jacobs Medical Center in La Jolla. 

This research marks the first instance where an AI deep-learning model utilizing artificial neural networks improved patient outcomes by correctly identifying health concerns.

Dr. Gabriel Wardi, Chief of the Division of Critical Care in the Department of Emergency Medicine at UC San Diego School of Medicine, underscores the significance of COMPOSER in expediting life-saving therapy for patients. 

The AI model has been active since December 2022 and is now expanding its reach to in-patient units in various UC San Diego Health hospitals, including the forthcoming UC San Diego Health East Campus.

UC San Diego health’s commitment to AI in healthcare

UC San Diego Health stands at the forefront of AI integration in healthcare. Recently, the institution appointed its first-ever Chief Health AI Officer and established the Joan and Irwin Jacobs Center for Health Innovation. 

This center develops advanced and sophisticated healthcare solutions that leverage AI technology.

In addition to the COMPOSER AI model, UC San Diego Health is also exploring innovative ways to incorporate AI into healthcare operations. For instance, the institution has initiated a pilot program in collaboration with Epic, a cloud-based electronic health record system, and Microsoft’s generative AI integration. 

This program streamlines the drafting of compassionate message responses through ChatGPT, allowing doctors and caregivers to allocate more time to patient care.