Overcrowding in emergency departments (EDs) is a major global public health issue and a recognized threat to patient safety [1]. It occurs when an increasing number of patients seeking care leads to a demand for emergency services that exceeds the department's ability to provide high-quality care [2,3]. Over the past two decades, the number of emergency department visits in the USA has increased, with 144 million visits in 2022 [4,5]. This overcrowding may affect the quality of patient care, leading to delayed treatment, increased morbidity, and waiting time [6,7]. This overcrowding also affects emergency nurses, resulting in elevated stress levels, burnout, and decreased job satisfaction [8,9].

Emergency department triage involves prioritizing patients and anticipating resource needs to ensure timely care for those requiring urgent treatment [10,11]. Emergency nurses play a critical role at the frontline of this process, assessing patients and allocating resources appropriately based on the urgency and complexity of their conditions [12,13,14]. The Emergency Severity Index (ESI) is widely used for efficient emergency department patient triage [15,16,17]. It categorizes patients into five categories, including ESI 1 (immediate) and ESI 2 (emergency) groups, which require urgent and critical care. In contrast, patients in ESI 3 (urgent), ESI 4 (semi-urgent), and ESI 5 (non-urgent) are clinically stable and classified according to the number of resources used [18]. Various strategies are currently available to help reduce overcrowding, including triage optimization, dynamic staffing, strategic resource management, the implementation of fast-track systems, and technological integration [19].

In recent years, artificial intelligence (AI) and machine learning have revolutionized healthcare, enabling computers to analyze large datasets, identify patterns, and make decisions that mimic human reasoning, offering new opportunities for streamlining triage in emergency departments [6]. Various studies have found that AI-assisted triage systems can quickly evaluate multiple variables, prioritizing patients based on their symptoms and historical data. This enhances the accuracy, reduces over- and under-triage, and improves the relevance of triage outcomes such as the F1 score (precision and recall), waiting time, time to treatment, and patient workflow [20,21,22,23]. AI-driven triage systems also assist in managing resources more effectively, especially during high-volume periods [6].