Rule-Based Artificial Intelligence and Workflow to Prompt Early Sepsis Management: A Quality Improvement Project.

Area of Science:

• Healthcare Quality Improvement
• Clinical Informatics
• Artificial Intelligence in Medicine

Background:

• Hospitals face challenges in consistent sepsis screening, impacting patient outcomes.
• The Centers for Disease Control (CDC) established Hospital Sepsis Program Core Elements to guide sepsis management.
• The National Healthcare Safety Network (NHSN) will measure sepsis core elements through an annual survey.

Purpose of the Study:

• To evaluate the impact of rule-based artificial intelligence (AI) for sepsis identification in an emergency department.
• To measure AI sensitivity, sepsis management compliance, length of stay (LOS), and mortality rate post-implementation.
• To assess the effectiveness of AI-driven sepsis workflows in improving patient care.

Main Methods:

• A quality improvement project was implemented in an emergency department, introducing rule-based AI for sepsis identification.
• The study included 895 cases, with 370 preimplementation and 525 postimplementation data points.
• Key metrics measured included AI alert sensitivity, 3-hour sepsis bundle compliance, LOS, and mortality.

Main Results:

• Rule-based AI alerts identified 93.9% of sepsis cases for intervention post-implementation.
• Combined 3-hour compliance for antibiotics, blood cultures, and lactate measurement reached 89.5%.
• Average LOS decreased by 2.3 days (p < .001), and mortality decreased by 22.3% (p = .0998).

Conclusions:

• Implementing rule-based AI software for severe sepsis identification, coupled with a sepsis workflow, reduced LOS.
• The AI tool demonstrated high sensitivity in identifying sepsis cases requiring intervention.
• The study highlights the potential of AI in enhancing sepsis management and improving hospital efficiency.