Explainable, Generalizable and Responsible AI Model to Triage Emergency Patients
Abstract
Triage helps to deliver the right level of emergency healthcare at the right time for the right person using the right resources. However, triage is vulnerable to mis-triage which causes delayed treatment, poor healthcare outcomes and ED overcrowding. This study, hence, aimed to develop an explainable, generalizable and responsible AI model that assists triage nurses. We identify the most important predictors, measure the order, direction, and effects of important predictors across triage levels, and quantify the minimum information required to develop a generalizable triage model.
Cite
Text
Fulli et al. "Explainable, Generalizable and Responsible AI Model to Triage Emergency Patients." NeurIPS 2024 Workshops: MusIML, 2024.Markdown
[Fulli et al. "Explainable, Generalizable and Responsible AI Model to Triage Emergency Patients." NeurIPS 2024 Workshops: MusIML, 2024.](https://mlanthology.org/neuripsw/2024/fulli2024neuripsw-explainable/)BibTeX
@inproceedings{fulli2024neuripsw-explainable,
title = {{Explainable, Generalizable and Responsible AI Model to Triage Emergency Patients}},
author = {Fulli, Jemal A and Alemayehu, Berihun T and Yasin, Omer K and Ahmed, Abubeker S and Sualih, Muhammed A},
booktitle = {NeurIPS 2024 Workshops: MusIML},
year = {2024},
url = {https://mlanthology.org/neuripsw/2024/fulli2024neuripsw-explainable/}
}