Implementing AI-Driven Clinical Decision Support Tools for Sepsis

Tuesday, April 18 at 2:45 PM - 3:15 PM CT
South Building, Level 1 | S104
Sepsis is the leading cause of mortality among hospitalized patients in our healthcare system. Understandably, sepsis has been the target of multiple major national initiatives such as Surviving Sepsis led by the SCCM and Get Ahead of Sepsis led by the CDC. Our institution strives to improve outcomes for patients with sepsis by implementing a novel suite of clinical decision support tools driven by a predictive learning algorithm. These tools were built with a focus on human centered design and multi-disciplinary care to improve usability and interaction of the tools by providers. These tools also allow the implementation of a floor and ceiling approach to managing sepsis at an institution. Virtual screening of patients at risk of sepsis by nurses helps to ensure patients are not missed and improved tools to rapidly implement complete care for the septic patient with a timer helps drive down time from alert to action. Robust analysis of outcomes and provider interaction with our tools has allowed us to improve the median time to antibiotic administration and increased the number of patients screened for sepsis at our hospitals.

Learning Objectives

  • Construct a user centered design for alerts with tiering of actions to customize decision support
  • Recommend how layered alerts with multidisciplinary interactivity creates mutual accountability and transparency
  • Illustrate how dashboards meaningfully track decision support alert volumes and how alert to action ratios play a role in thoughtful alert design
Chief Quality Officer and Chief Clinical Transformation Officer, Clinical Informaticist, CMIO/CMO


Christopher Girardo, DO
Fellow, Clinical Informatics / PRN Staff, Pathology
LSU Health Shreveport and Ochsner Health
Marshall Frieden, MD
Resident Physician
LCMC Health