Bridging Trust and Efficiency: Benchmark Dataset for Collaborative AI-Human Triage for Emergency Room Decision-Making

ACL ARR 2025 February Submission8186 Authors

16 Feb 2025 (modified: 09 May 2025)ACL ARR 2025 February SubmissionEveryoneRevisionsBibTeXCC BY 4.0
Abstract: AI offers the important potential to enhance Emergency Room (ER) triage efficiency but a lack of trust from healthcare professionals and patients limits its adoption due to concerns over accuracy and reliability. To address this, we introduce the Collaborative Intelligence-based Clinical Triage Dataset (CICTD), a large-scale benchmark containing patient requests and ER doctor annotations for triage decision-making. Along with CICTD, we define key evaluation tasks, including diagnostic question generation, ESI level prediction, triage recommendations, and misdiagnosis prediction. Our approach emphasizes a human-in-the-loop framework, ensuring AI escalates uncertain cases to experts, balancing automation with trust, and improving ER triage efficiency.
Paper Type: Short
Research Area: NLP Applications
Research Area Keywords: Triage, Human-in-the-loop, dataset, benchmark
Contribution Types: Data resources
Languages Studied: English
Submission Number: 8186
Loading

OpenReview is a long-term project to advance science through improved peer review with legal nonprofit status. We gratefully acknowledge the support of the OpenReview Sponsors. © 2025 OpenReview