New AI approach could cut sepsis deaths and drive efficiency in emergency departments
By explaining which causes led to a patient's death, the team hopes their work will empower hospital managers to make targeted changes.
Researchers in the causal explainability group at King's College London have developed an AI method to explain why patients with sepsis die during their time in emergency departments. The model identifies specific factors, such as delayed treatment, that contribute to patient deaths by analyzing sequences of events, or "traces", that occur during an emergency department visit.
By providing granular, causative data, the team hopes hospital managers can make targeted changes to improve patient outcomes and increase efficiency. Dr. Nathan Blake, a research associate and former emergency department nurse, explained that seeing patterns in patient journeys can be crucial in sepsis cases where time is critical.
The method captures temporal dependencies and incorporates clinical expertise, allowing for a much more accurate analysis of the causes of a patient's deterioration compared to correlation-based approaches. The AI model's explanations are generated without compromising computational efficiency, allowing for timely predictions of patient outcomes.
Written by urgent.news from Medical Xpress's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.