Large language models (LLMs) have shown potential in the field of emergency department (ED) triage. However, a comprehensive evaluation of their diagnostic performance and a comparison with triage nurses have not been extensively explored. To systematically evaluate the comparative triage accuracy of LLMs versus triage nurses in ED, assess their performance in identifying patients assigned to the highest-acuity triage category by the reference standard. A systematic search was conducted in CNKI, Wanfang, VIP, CBM, PubMed, Embase, Web of Science, and Cochrane Library databases from inception to March 2025. Two researchers independently screened and evaluated the quality of the selected studies and conducted meta-analyses using RevMan 5.4 and Stata 16.0. A total of 11 studies were included, involving 3088 real or simulated ED cases. The pooled odds ratio (OR) for LLM triage accuracy was 0.65 (95% CI: 0.32-1.34; I² = 97%). For identifying patients assigned to the highest-acuity triage category by the reference standard, LLMs demonstrated a pooled sensitivity of 61% (95% CI: 48%-73%) and specificity of 97% (95% CI: 88%-99%), with statistically significant differences (p < 0.05). The pooled area under the curve (AUC) was 0.82 (95% CI: 0.78-0.85). These estimates reflect concordance with triage category by the reference standard rather than discrimination of outcome-anchored physiological acuity. No significant publication bias was detected by Deeks' funnel plot asymmetry test. Current LLMs demonstrate moderate discriminatory performance for identifying patients assigned to the highest-acuity triage category by the reference standard, but have limited sensitivity for these cases. LLMs may have potential as adjunctive support tools, but further prospective evaluation with patient-centered and safety-critical outcomes is required before implementation can be considered.

BackgroundEmergency Department (ED) crowding has escalated globally and has been associated with delayed treatment and increased short-term mortality. In western Sweden, several EDs transitioned from the Rapid Emergency Triage and Treatment System (RETTS©) to the West coast system for triage (WEST), based on principles of the South African Triage Scale. While WEST was introduced and spread rapidly and may have potential benefits, reports suggested challenges during its implementation.ObjectiveThis study examined the implementation of WEST at the first two EDs (ED1 and ED2) and identified barriers and facilitators. The study focused on the transition from RETTS© to WEST and explored staff experiences across implementation stages.MethodsThe study used a qualitative descriptive design with a mixed-methods approach and included semi-structured individual interviews with 36 triage nurses from two hospitals (ED1 n = 17; ED2 n = 19). Interviews explored experiences with WEST, implementation challenges, and comparisons with RETTS©, using Rogers’ “Diffusion of Innovations” as a sensitizing framework and treating each ED as a distinct social system. A deductive qualitative content analysis guided by Rogers’ Diffusion of Innovations was conducted, with inductive coding of unexpected themes from the final open question. Descriptive analyses of Likert-scale items complemented the qualitative analysis. The study received ethical approval, and written informed consent was obtained from all participants before interviews.ResultsED1 participants expressed great enthusiasm and perceived WEST as a notable improvement. ED2 responses were more varied, with reservations commonly targeting the implementation process and concurrent organizational changes. Across both sites, most nurses (30/36) perceived WEST as more accurate than RETTS© in identifying critically ill patients.ConclusionsImplementing WEST was feasible but context dependent. Information, motivation, and consensus-building, education, and organizational readiness emerged as important factors in the implementation process. Viewing each ED as a distinct social system, with its own culture and readiness for change, helped explain why adoption of WEST varied between sites.

Comparison of the performance of ChatGPT Plus, Gemini Pro, and experienced nurses in pediatric emergency triage using physician reference assessment: A prospective observational study.

BackgroundRemote nurse triage services are increasingly delivered through call centers to manage demand for urgent and unscheduled care. These encounters typically rely on decision-support tools, but safe and efficient outcomes also depend on how nurses interpret patient information and make decisions about the most appropriate level of care. Despite the growth of tele-triage, nurse decision-making in this context remains underexplored.ObjectiveTo examine factors shaping triage decisions and outcomes in the U.S. Department of Veterans Affairs' (VA) national tele-triage and urgent care program.MethodsWe conducted semi-structured interviews with 39 participants, including triage nurses, nurse managers, and virtual urgent care providers, across six VA regions. Interviews explored decision-making processes, perceived challenges, and coordination between nurses and providers. Transcripts were coded and analyzed thematically.ResultsFour categories of factors influenced triage decisions: (1) technology-decision-support tools structured encounters but were limited by usability issues and risk-averse design; (2) professional training and judgment-nurses varied in critical thinking and remote assessment skills; (3) organizational priorities-policies emphasizing safety and liability restricted decision-making discretion; and (4) patient factors-preferences, beliefs, and situational constraints shaped acceptance of virtual care. Participants described tensions between efficiency and safety, and providers thought some Veterans' symptoms were more appropriately treated through primary care, reflecting institutional pressures.ConclusionsVA's tele-triage process is shaped by decision-support tools, nurse expertise, organizational priorities, and patient preferences. Algorithms provide structure, but effective triage depends on nurses' ability to interpret recommendations in context. As call center-based tele-triage expands, ensuring technology supports rather than constrains clinician judgment will be essential for safe, efficient, and patient-centered care.

A Prospective, Observational Study Evaluating the Impact and Safety of an Emergency Physician Collaboration With a Nurse Triage Phone Line

How triage nurses generate initial hypotheses during the first patient encounter: A focused ethnographic study.

To explore emergency department triage nurses' scope of practice and activities related to their triage role and management of patients located in emergency department waiting areas. Exploratory, descriptive, observational study using naturalistic decision making. Data were collected using semi-structured non-participant observation: researchers recorded their observations using a lapel microphone and recorder from 8 January to 7 May 2025. Fifteen triage nurses from three emergency departments in Melbourne, Australia were observed for 2 to 2.5 h each. Audio-recordings were transcribed verbatim and analysed using deductive content analysis. The HIRAID emergency nursing framework (History including Infection risk, Red flags, Assessment, Interventions, Diagnostics, reassessment and communication) was used as the coding frame. Participants had a median of 10 (interquartile range 7.5-21) years nursing and 5 (interquartile range 3.2-13) years triage experience. During a total of 33 observation hours, there were 303 interactions, including 237 interactions with 169 different patients; the remainder were with carers or other clinicians. In total, 1183 tasks were coded ranging from 12 to 128 tasks per triage nurse. The most common tasks were: taking a history (n = 475); post-triage communication with patients, carers and other clinicians (n = 288); patient assessment including vital signs and focussed assessments (n = 165); and interventions including medications and psychological care (n = 134). All elements of taking a history, identifying red flags and assessment were more common during the triage process, with few instances during post-triage care. Interventions (medication administration, psychological care), diagnostics (pathology, imaging, urinalysis) and communication with patients, carers and other clinicians occurred during both triage and post-triage care. Triage nurse practice is complex, multifaceted and extends beyond triage category allocation. The traditional perspective that triage and post-triage care are linear and clearly divided is not fit for purpose in contemporary triage practice. The scope of triage nurses' practice both in the context of triage of incoming patients and care of patients in ED waiting areas is poorly understood. The role of triage nurses extends far beyond rapid assessment and triage category allocation and triage nurses use a breadth of expertise and skills to maintain safety, promote comfort and expedite emergency care for patients in ED waiting areas. History, red flags and assessment were more likely during the triage process than post-triage, but interventions, diagnostics and communication spanned both triage and post-triage care. Future triage and waiting area models of care and educational preparation of triage nurses should be co-designed with triage nurses and consumers, so they reflect care as delivered rather than care as imagined. Consolidated criteria for reporting qualitative research (COREQ). No patient or public contribution.

Incidence of Cervical Spine Fracture or Other Urgently Actionable Injury in Trauma Patients Who Present to the Emergency Department by Private Vehicle.

Escalating emergency department overcrowding poses a serious threat to healthcare delivery, leading to reduced quality of care, increased risk of infection transmission, and heightened tension between medical staff and patients. This study aimed to investigate the status of emergency department visits by young and middle-aged patients triaged as non-urgent and to explore the influencing factors. This study retrospectively analyzed clinical data from 55,578 young and middle-aged patients who visited the emergency department of a tertiary class A hospital in Guangdong Province between January 1 and December 31, 2024. Patients were grouped into urgent and non-urgent based on triage classification. The differences in the clinical indicators between the two groups were compared. Logistic regression analysis was conducted to identify the factors associated with non-urgent triage classification. Among the included young and middle-aged patients, 35,314 (63.5%) were classified as non-urgent cases, while 20,264 (36.5%) were classified as urgent. Binary logistic regression analysis revealed that non-urgent triage classification was significantly associated with the following factors (p < 0.05): visiting Emergency Fever (OR = 1.513) and Emergency Gastroenterology (OR = 4.278) compared to Emergency Internal Medicine and Surgery; arriving on foot or by other modes (OR = 6.073) compared to ambulance transport; visiting during the afternoon shift (OR = 1.110) compared to morning shift, on weekends (OR = 1.111), and during the autumn (OR = 1.140) and winter seasons (OR = 1.083) compared to spring. The rate of non-urgent triage among young and middle-aged patients is high. Triage nurses should be particularly vigilant in assessing patients presenting to departments with a higher likelihood of non-urgent classification, such as the Emergency Fever and Emergency Gastroenterology, as well as those arriving without ambulance transport, during non-working hours and peak flu seasons. These findings can inform nurse-led strategies, such as dynamic staffing and guiding appropriate patients to alternative services, such as internet hospitals, thereby mitigating emergency department overcrowding. It should be noted that in this study, non-urgent status was defined solely by the triage system and was not validated against clinical outcomes; therefore, the findings describe patterns of triage classification rather than objective patient acuity.

BackgroundEmergency department nurses’ attitudes toward early disease deterioration affect their assessment behavior and triage ability. A negative attitude delays assessment, impacting patient prognosis. Analyzing attitude characteristics, cultivating positive attitude and enhancing triage abilities are crucial. This study aims to identify distinct profiles of attitudes toward the early recognition of clinical deterioration among emergency department (ED) nurses in China, analyze demographic and statistical factors associated with these profiles, and explore the relationship between these attitude profiles and ED nurses' triage competence.MethodsFrom June to July 2024, a questionnaire survey was conducted among ED nurses from 31 general hospitals across seven provinces and municipalities in China (Anhui Province, Hubei Province, Hunan Province, Henan Province, Zhejiang Province, Jiangsu Province, Shanghai). The Attitudes Toward Recognition of Early Nursing Deterioration (ATREND) questionnaire and the Triage Nurse Professional Competency Questionnaire for Emergency Departments (TNPCQ-ED) were used as research tools. Data were statistically analyzed using SPSS 28.0 and Mplus 8.7. Latent profile analysis was employed to identify attitude characteristics, while multivariate logistic regression was applied to explore influencing factors of attitude profiles and regression analysis was used to examine the relationship between attitude profiles and triage competence.ResultsThe Chinese versions of the ATREND and TNPCQ-ED questionnaires demonstrated good reliability and validity among ED nurses. A total of 942 valid questionnaires were collected. The overall mean score for the ATREND questionnaire was (39.50 ± 7.34) points. Among its subscales, the Beliefs About Monitoring Importance scored the highest, while the other two subscales followed. The overall mean score for the TNPCQ-ED was (149.19 ± 21.01) points. Among its subscales, the professional commitment dimension had the highest score, followed by the clinical comprehensive capability and psychological empowerment dimensions. Latent profile analysis revealed four distinct attitude profiles: “Skill-Proficient & Belief-Mismatch Profile” (11.25%); “Belief-Dominant & Skill-Developing Profile” (53.72%); “Neutral & Balanced Profile” (15.92%); “Belief-Advanced & Skill-Developing Profile” (19.11%). Univariate analysis showed significant differences in age, years of work experience, job title, trained in triage, years at triage station and hospital level among the four profiles (p < 0.05). Multivariate logistic regression revealed the following: 1. Nurses in the "Belief-Advanced & Skill-Developing Profile" were more likely to have received systematic triage training. 2. Nurses in the “Belief-Dominant & Skill-Developing Profile” were more likely to have <5 years of triage desk experience. 3. The “Belief-Advanced & Skill-Developing Profile” were also more likely to work in provincial tertiary hospitals (p < 0.05). There was a statistically significant association (p < 0.001) between attitude profiles and dimensions of triage competence.ConclusionsED nurses’ attitudes toward early recognition of clinical deterioration can be categorized into four latent profiles: “Skill-Proficient & Belief-Mismatch Profile”, “Belief-Dominant & Skill-Developing Profile”, “Neutral & Balanced Profile” and “Belief-Advanced & Skill-Developing Profile”. Multivariate logistic regression indicated that factors such as systematic triage training, <5 years of triage desk experience, and employment in provincial tertiary hospitals influenced attitude profiles. Moreover, nurses’ attitudes significantly impacted their triage competence. Nurse managers should design targeted interventions tailored to the characteristics of these profiles to enhance ED nurses’ triage competence and foster positive attitudes toward the early recognition of clinical deterioration. This approach could improve proactive assessment capabilities, prevent further patient deterioration, and ultimately enhance the quality of emergency nursing care.

Structural empowerment of triage nurses in the redirection of low-acuity patients in a Swiss emergency department: a mixed-methods convergent pilot study

Older adults account for a disproportionate share of emergency department (ED) visits and are often present with complex needs. Frailty is a key determinant of adverse outcomes in this population. We previously developed the Emergency Department Frailty Scale (ED-FraS), a novel clinician-judgment-based instrument. This study prospectively validated the ED-FraS when used by triage nurses and evaluated its predictive performance for hospital admission and ED length of stay (EDLOS). This prospective observational study was conducted at a tertiary academic medical center in Taiwan between February and July 2025. Triage nurses assessed patients aged ≥ 65 years using the ED-FraS during routine triage encounters. We evaluated the association between ED-FraS levels (1-5) and hospital admission and EDLOS. We compared the predictive performance of ED-FraS, the standard Taiwan Triage and Acuity Scale (TTAS), and a modified TTAS (mTTAS), which integrated frailty scores. Discriminative ability was measured using the area under the receiver operating characteristic curve (AUROC). A total of 550 older adults were enrolled (mean age 77.2 years). Higher ED-FraS levels were significantly associated with increased admission rates (25.0% in level 1 vs. 59.4% in level 5) and prolonged median EDLOS (2.6 vs. 27.3 h). The mTTAS demonstrated superior discriminatory ability for hospital admission (AUROC 0.720) compared to TTAS (0.657) or ED-FraS alone (0.638). Nurses reported the tool was feasible, taking < 30 s to complete. The ED-FraS is a feasible and valid tool for identifying older adults at risk of adverse outcomes during triage. Integrating frailty assessment into standard triage systems enhances risk stratification and may improve resource allocation for vulnerable older adults.

Redirection of non-urgent patient visits from the pediatric emergency department (ED) to community healthcare professionals may reduce crowding. As there is no consensus about who should be eligible for redirection, this qualitative study explored how health professionals conceptualize eligibility for pediatric redirection. We completed one-on-one semi-structured interviews with pediatric ED triage nurses and physicians, office-based pediatricians and family physicians, and healthcare leaders who oversee pediatric and ED operations. Interviews were recorded, transcribed, and analyzed using reflexive thematic analysis within an interpretive description framework to generate clinically relevant insights. From our 24 interviews (6 community professionals, 11 pediatric ED professionals, and 7 healthcare leaders), we identified three guiding principles that are important to consider when developing eligibility criteria: 1. assessing the impact of presenting condition on the patient and caregivers; 2. medical complexity of the clinical condition, patient, and health system; and 3. process-related considerations. This study captures referring and receiving healthcare professional perspectives about eligibility for pediatric ED redirection. The resulting guiding principles support prior work that focuses on patient characteristics and medical complexity, and also emphasizes the importance of considering family capacity, patient history, and shared decision making when creating pediatric redirection eligibility criteria. Through interviews with referring and receiving healthcare professionals and health leaders, we identified guiding principles that can support identifying patients suitable for redirection from the pediatric ED. The guiding principles we developed provide a structured, healthcare professional-informed foundation to determine which children may be suitable for ED redirection and highlight the importance of caregiver, health system, and process factors often overlooked in prior approaches. Our guiding principles reflect health professionals' values surrounding redirection and if applied can enhance the acceptability and sustainability of pediatric ED redirection programs.

The Virtual Urgent Nova Scotia (VUNS) program was introduced to address emergency department (ED) overcrowding and improve access to urgent care through virtual health services. This study captures the perspectives, sentiment, and attitudes toward the implementation of VUNS from rural ED providers. VUNS staff from seven different sites across the province of Nova Scotia were invited to focus groups to explore the implementation of VUNS in their local context. Focus groups were facilitated by members of the Implementation Science Team at Nova Scotia Health from November 2024 to January 2025. Focus group transcripts were coded deductively, and results were presented in a narrative and descriptive fashion. Key findings highlight recurrent issues, including staff confusion over inclusion and exclusion criteria, inconsistent VUNS physician decision-making, workflow disruptions influenced by staffing shortages, and patient bounce-backs from the VUNS triage nurses or physician. While initial implementation demonstrated potential for reducing lower acuity patients' wait times and enhancing care delivery, operational challenges limited its effectiveness for some sites. Despite these barriers, success stories from specific sites emphasize the program's value, particularly in underserved rural areas. Recommendations for improvement focus on stabilizing criteria, enhancing staff training, optimizing workflows, and fostering consistent communication between stakeholders. Findings indicate that VUNS should be scaled cautiously within current sites, addressing identified challenges before spreading to new locations. Lessons learned stress the importance of staff engagement, clear workflows, and patient education. These findings provide a roadmap for refining VUNS operations and contribute to the broader discourse on scaling digital healthcare innovations and streamlining patient flow in EDs.

We evaluated how emergency department (ED) staff perceives the practice known as redirection, whereby triage nurses guide low-acuity patients to alternative care settings without evaluation by a physician. Our aim was to evaluate its use across Canada and to identify its key drivers and barriers to its implementation. We conducted a cross-sectional survey of Canadian ED physicians, nurses and managers from September to December 2023. The survey tool was developed in French and English following a modified Dillman's tailored design method, including: (1) literature review to identify key themes on redirection; (2) semi-structured interviews with experts on redirection; (3) the development of a tool prototype; (4) scientific and linguistic revisions; and (5) pre-testing. The survey was distributed through the mailing list of the Canadian emergency medicine and nursing associations. Of the 719 respondents recruited, 47.0% were nurses, 44.2% were physicians and 5% were managers. The overall response rate was 10.2%. Most respondents endorsed redirection as safe, with this endorsement ranging from 75.5% in Ontario to 94.3% in Manitoba. Similarly, the view that first-line physicians can adequately manage redirected patients was supported by most respondents, with proportions ranging from 78.1% in Ontario to 92.1% in Québec. Redirection strategies reported by the majority of respondents were based on the Canadian Triage and Acuity Scale (65.2%). Insufficient opening hours of clinics (87.2%) and those with a CTAS score of 3 (62.7%) were identified as the main challenges. Professionals most suggested to receive redirected patients were family physicians (90.9%), nurse practitioners (86.4%), dentists (83.8%), social workers (71.9%), pharmacists (63.9%), and physiotherapists (58.0%). In this pan-Canadian survey of ED personnel, the majority of respondents expressed support for redirecting low-acuity patients. These findings indicate an opportunity for further research on the development of redirection tools.

Atopic dermatitis (AD) is a prevalent, chronic inflammatory skin disease with diverse clinical presentations, often overlapping with other dermatoses. Its diagnosis remains largely dependent on clinical expertise, leading to variability and limited diagnostic accuracy, particularly among general practitioners. This study aimed to develop and evaluate a multimodal artificial intelligence (AI) model that integrates lesion image analysis and structured anamnesis to improve AD diagnosis. This diagnostic study was conducted in two phases: Phase 1 used retrospective data from 2021-2024, and Phase 2 involved prospective external validation from multiple hospitals in 2025. Patients with AD or related skin conditions were included, with diagnoses based on AAD 2014 criteria. Multimodal fusion combined ResNet50-extracted image features and MPNet-based anamnesis text features using a late fusion model. This approach mimics clinical reasoning by integrating visual and contextual clinical information to classify cases as AD or non-AD. The multimodal AI model integrating ResNet50 (image) and MPNet (anamnesis) achieved 98.28% accuracy in classifying AD vs non-AD, outperforming image-or text-only models. It offers clinical advantages by mimicking physician reasoning, improving diagnostic consistency, reducing subjectivity, and enabling mass triage. However, real-world generalizability remains a challenge due to limited training diversity, potential language constraints (Bahasa Indonesia), and narrow differential diagnoses. External validation and explainable AI (XAI) are critical for broader application. Despite limitations, the model aligns with emerging literature, showing multimodal AI can approach or surpass expert-level performance in dermatological diagnosis when rigorously validated. The multimodal ResNet50-MPNet model shows near-perfect accuracy in diagnosing AD by mimicking clinician reasoning. It offers consistent, holistic assessment but requires external validation and improved interpretability for clinical adoption. Continued AI-clinician collaboration is vital to translating this promising technology into real-world dermatological care.

Background: The activation of Emergency Department (ED) Code Stroke at our primary stroke center was not consistently meeting metric goals set by the hospital system for Door to Needle (DTN) time of &lt;45 minutes. The process utilized Provider-only activation, with delays from limited providers and multiple relays of communication. Purpose: This practice change aimed to discover if nurse-activation of a Code Stroke in triage influenced the DTN metrics of: Door to Activation, Door to CT initiation, and DTN times. Methods: A collaborative approach was used to identify points of delay, develop innovative workflow, and education to the new process. A “Code Stroke Button” was added to the emergency call system terminal to minimize communication relays; pressing the button sends an alert directly to the Code Stroke response team. Education on the updated workflow was provided to the ED team by a multidisciplinary team (neurologist, ED provider, Nurse Educator and RN Stroke Coordinator). Training and education included: recognition of stroke symptoms, criteria for Code Stroke activation, and feedback. The Plan-Do-Check-Act framework was used to evaluate process performance monthly for 6 months and data was analyzed at one year post implementation. Results: By Quarter 4, six months from change, median times for all metrics saw meaningful reduction. Door to Activation time was reduced by 16 minutes and Door to CT initiation by 11 minutes. In thrombolytic cases, median DTN decreased by 12 minutes. At one year, the median DTN time improved by 20.8% (53 minutes to 42 minutes), meeting the DTN goal of &lt;45 minutes. Concerns for overactivation were dispelled with data at one year showing 91.5% of nurse-activated Code Stroke met criteria. Conclusions: The activation of Code Stroke by nurses in triage led to a meaningful reduction in specific metrics. A collaborative approach was imperative to identify process issues and led our team to make innovative changes. Constructive feedback allowed staff to confidently put changes into practice. Sharing monthly data results with ED nurses, technicians, management, providers, Stroke Program management and neurologists allowed for evaluation of the changes to meet the DTN goals.

Emergency Nurses' Knowledge, Attitudes, and Behaviors Related to Patients Presenting to Triage With Pain: A Mixed Methods Study.

Interventions to reduce overcrowding in emergency departments: An umbrella review.

Emergency Nurses' Experiences of Triaging Geriatric Trauma Patients: A Qualitative Study.