Clinical Triage
Overview
Clinical triage studies how to identify who needs care most urgently, from emergency departments and remote monitoring to specialty referrals and patient-facing advice. Systems combine structured records and clinical text with methods such as machine learning, multimodal models, language models, and rule-based guidelines; some also ask follow-up questions or defer uncertain cases to clinicians. Evaluation tests acuity classification and referral decisions on clinical records, vignettes, conversations, and real workflows, with attention to missed emergencies, unnecessary escalation, fairness, and performance across institutions. A central challenge is that strong results on completed cases or curated benchmarks may not carry over to incomplete, sequential, or changing real-world encounters, where safe decisions depend on uncertainty and information gathering as well as prediction.
Latest papers
48 papers4 papers in the last four weeks, level with the four weeks before. 0.0% of all new papers.