Human-Centered AI Tool Predicts Patient’s Sepsis Risk Within Four Hours
Posted on 28 Aug 2024
Sepsis, a critical and life-threatening response to infection, can quickly lead to organ failure and is highly difficult to diagnose due to its common symptoms such as fever, low blood pressure, and increased heart rate, which mimic many other conditions. Now, an innovative artificial intelligence (AI) tool designed to assist clinicians in making decisions about patients at risk of sepsis introduces a novel feature: it accounts for uncertainties in its predictions and suggests additional information, such as demographic data, vital signs, and lab test results, needed to enhance its accuracy.
The system, called SepsisLab, was developed by scientists at The Ohio State University (Columbus, OH, USA) based on feedback from doctors and nurses in emergency and intensive care settings, where sepsis frequently occurs. These healthcare professionals expressed concerns over existing AI tools that rely solely on electronic health records without incorporating clinical inputs. SepsisLab improves upon this by predicting sepsis risk within a four-hour window while actively identifying and quantifying the importance of missing patient data, visually informing clinicians how certain pieces of information can influence the risk assessment.
This AI system updates its predictions hourly as new patient data is incorporated, continuously refining its accuracy. It also provides clinicians with actionable insights, suggesting which laboratory tests might be most informative and estimating how different clinical interventions could alter the patient's risk of developing sepsis According to the research published Aug. 24 in KDD ’24: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, incorporating just 8% additional data from lab results and other key variables can decrease uncertainty in the predictions by 70%, enhancing the tool’s accuracy in assessing sepsis risk by 11%.
“The existing model represents a more a traditional human-AI competition paradigm, generating numerous annoying false alarms in ICUs and emergency rooms without listening to clinicians,” said senior study author Ping Zhang, associate professor of computer science and engineering and biomedical informatics at Ohio State. “The idea is we need to involve AI in every intermediate step of decision-making by adopting the ‘AI-in-the-human-loop’ concept. We’re not just developing a tool – we also recruited physicians into the project. This is a real collaboration between computer scientists and clinicians to develop a human-centered system that puts the physician in the driver’s seat.”
Related Links:
The Ohio State University
