Evaluating the Impact of Artificial Intelligence-Driven Triage Systems on Emergency Care Efficiency in Resource-Limited Settings: A Scoping Review

Evaluating the Impact of Artificial Intelligence-Driven Triage Systems on Emergency Care Efficiency in Resource-Limited Settings: A Scoping Review | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Evaluating the Impact of Artificial Intelligence-Driven Triage Systems on Emergency Care Efficiency in Resource-Limited Settings: A Scoping Review Joy Aifuobhokhan, Chukwuemeka Abraham Agbarakwe, Deborah Oladunmolu Oduguwa, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8097974/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Emergency departments (EDs) in resource-limited settings face persistent challenges including overcrowding, delayed triage, and workforce shortages. Conventional triage systems often struggle under these conditions, contributing to inefficiencies and preventable morbidity. Artificial intelligence (AI)-driven triage systems have emerged as innovative tools to enhance patient prioritization, clinical decision-making, and resource allocation. However, their real-world impact on emergency care efficiency, particularly within low- and middle-income countries (LMICs), remains insufficiently defined. Methods: This scoping review followed the Joanna Briggs Institute (JBI) methodology and adhered to PRISMA-ScR reporting standards. This review was registered with PROSPERO (1208548). Comprehensive searches were performed across PubMed, Scopus, Web of Science, IEEE Xplore, Cochrane Library, and Embase, supplemented by grey literature sources such as Google Scholar, WHO, and World Bank reports. Eligible studies examined AI-driven triage applications within emergency care settings, with relevance to LMICs or comparable resource-constrained environments. Data were charted and analyzed thematically to synthesize trends in design, implementation, and outcomes. Results: Forty-three studies met inclusion criteria, comprising 21 quantitative, 12 qualitative or mixed-methods, and 10 review articles. Most research originated from high-income countries, though studies from LMICs are increasing. AI triage systems consistently improved patient flow, reduced waiting times, and enhanced alignment between triage levels and clinical urgency. In resource-limited contexts, AI supported overburdened clinicians, optimized staffing, and improved patient safety. Key challenges included algorithmic bias, limited data infrastructure, insufficient external validation, and ethical concerns surrounding transparency and accountability. Conclusions: AI-driven triage systems demonstrate strong potential to enhance emergency care efficiency in resource-limited settings. However, context-specific evidence from LMICs remains limited. Future research should emphasize prospective validation, cost-effectiveness analysis, and ethical governance to ensure equitable and sustainable AI integration in emergency care. Artificial intelligence emergency department triage low- and middle-income countries healthcare efficiency resource-limited settings Figures Figure 1 Figure 2 Background Overview of Emergency Care in Resource-Limited Settings Emergency care plays a critical role in reducing preventable deaths from acute illnesses and injuries, yet in many low- and middle-income countries (LMICs), the infrastructure to deliver timely, high-quality emergency services remains underdeveloped [ 1 ]. LMIC emergency departments (EDs) often operate under conditions of chronic overcrowding, insufficient staffing, limited diagnostic capacity, and scarce specialist availability [ 2 ]. These challenges are compounded by high burdens of trauma, communicable diseases, and emerging non-communicable emergencies such as cardiovascular disease and stroke [ 3 ]. Without efficient patient flow, even patients with treatable conditions face prolonged delays, which directly contribute to poor outcomes and increased mortality [ 4 ]. Role of Triage Systems in Efficiency and Outcomes Triage, the process of categorizing patients according to urgency- has long been recognized as a cornerstone of emergency medicine [ 5 ]. Effective triage enables rapid identification of critically ill patients, reduces waiting times, and optimizes allocation of scarce resources [ 6 ]. Evidence from both high- and low-resource settings shows that standardized triage protocols such as the Emergency Severity Index (ESI), the Manchester Triage System (MTS), and the South African Triage Scale (SATS) are associated with improved throughput, reduced ED crowding, and better patient outcomes [ 7 ]. However, these systems rely heavily on trained personnel and uninterrupted workflows, conditions that are difficult to maintain in resource-limited environments, where overburdened staff, high patient volumes, and variability in clinical experience often led to inconsistent application [ 8 ]. AI Advancements in Triage and Relevance to LMICs Recent advancements in artificial intelligence (AI), including machine learning and natural language processing, have shown promise in automating and enhancing triage decision-making [ 9 ]. AI-driven triage tools can analyze structured and unstructured clinical data in real time to assign acuity scores, predict deterioration, and recommend interventions [ 20 , 21 ]. Studies have demonstrated improvements in triage accuracy, reductions in ED length of stay, and enhanced prioritization of high-risk patients when AI is integrated into triage workflows [ 10 ]. Importantly, in LMICs where human resources are scarce, AI systems could help mitigate workforce shortages, standardize decision-making, and improve access to timely care [ 11 ]. Early trials, such as those adapting the Kampala Trauma Score into machine learning–based models, have already shown superior predictive performance in trauma cases compared to traditional methods in low-resource environments [ 12 ]. Gap in Current Research and Rationale for Study Despite promising results, most AI triage research has been conducted in high-income countries with robust digital infrastructure [ 13 ]. Evidence on AI triage in LMIC EDs remains limited, fragmented, and contextually narrow [ 14 ]. There is a lack of large-scale, prospective evaluations assessing AI’s impact on core operational metrics (e.g., patient throughput, wait times, resource utilization) and patient-centered outcomes (e.g., morbidity, mortality) in these settings [ 15 ]. Additionally, barriers such as limited interoperability with existing health information systems, insufficient local datasets for algorithm training, and clinician skepticism towards AI tools remain underexplored [ 16 ]. Understanding these factors is critical for informing scalable, equitable, and sustainable implementation strategies in LMICs. Research Question and Objectives This study seeks to answer the central question: How do AI-driven triage systems impact emergency care efficiency and patient outcomes in resource-limited settings? The specific objectives are to: Evaluate the effects of AI-driven triage systems on ED operational efficiency, including patient throughput and waiting times. Assess the impact of AI triage on clinical outcomes, including time to intervention, morbidity, and mortality. Identify facilitators and barriers to implementing AI triage systems in LMIC EDs. Provide evidence-based recommendations for integrating AI into emergency care workflows in resource-limited contexts. Methodology Research Design, Protocol, and Registration This study was conducted as a scoping review in accordance with the Joanna Briggs Institute (JBI) methodology for scoping reviews [ 17 ]. The scoping protocol was developed in alignment with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist [ 18 ]. The review was prospectively registered with the International Prospective Register of Systematic Reviews (PROSPERO) under registration number 1208548. The PRISMA-ScR checklist for this study has been completed and is available in Supplementary Material 1. The approach was chosen to comprehensively map the extent, range, and nature of research on AI-driven triage systems in emergency care settings, with particular attention to evidence from low- and middle-income countries (LMICs). Eligibility Criteria The contextual scope includes emergency and acute care environments such as: Hospital emergency departments (EDs) and emergency rooms (ERs) Pre-hospital or ambulance triage systems (including tele-triage platforms) Urgent care centers or emergency units within primary or district hospitals Disaster or mass-casualty triage applications relevant to low-resource environments Healthcare settings in both public and private systems were eligible, provided they operated within resource-limited or LMIC contexts (as defined by World Bank classification). Studies conducted exclusively in high-income countries (HICs) were included only for comparative context, where they provided insights transferable to LMIC settings (e.g., model adaptation, algorithmic performance under data scarcity, or ethical implications). Inclusion Criteria Studies were eligible if they: Evaluated AI-based or AI-augmented triage systems (machine learning, deep learning, NLP, hybrid models, or decision-support algorithms). Focused on emergency or acute care settings, including hospitals, pre-hospital environments, or virtual triage systems. Reported on at least one of the following outcomes: Triage accuracy or validity compared to standard clinical assessment Time-to-triage or waiting time reduction Patient flow efficiency, throughput, or prioritization improvement Clinical outcomes (e.g., adverse events, safety metrics) Implementation feasibility, user acceptability, or workflow integration Were conducted in LMICs or resource-limited contexts or offered data applicable to such settings. Were peer-reviewed original studies, systematic/scoping reviews, or technical reports with empirical data. Were published between January 2015 and September 2025, in English language. Exclusion Criteria Studies were excluded if they: Focused on AI tools not used for triage (e.g., diagnostic imaging, predictive analytics, or ICU prognostication). We were conducted exclusively in high-income countries without relevance or transferability to resource-limited contexts. Were purely theoretical, editorials, or opinion pieces lacking empirical data. Addressed non-emergency care settings (e.g., outpatient clinics, elective care). Were published in languages other than English. Focused solely on algorithmic development or performance metrics without clinical application. Additional Contextual Considerations Healthcare system heterogeneity: Studies from LMICs often presented fragmented data infrastructures, paper-based triage records, and variable EHR integration, influencing AI feasibility and scalability. Technology infrastructure: Only studies reporting functional deployment or piloting of AI-based triage (not simulations alone) were included. Equity and ethics: Studies addressing algorithmic fairness, workforce readiness, or policy barriers in digital health implementation in LMICs were prioritized. Grey literature inclusion: To capture policy and implementation realities, grey sources (WHO, World Bank, Ministry of Health reports, digital health pilot evaluations) were also reviewed. Information Sources A comprehensive search was conducted across the following electronic databases: PubMed/MEDLINE, Scopus, Web of Science, IEEE Xplore, Cochrane Library, and Embase. Grey literature sources (Google Scholar, ResearchGate, and organizational reports from the WHO, World Bank, and Ministries of Health) were also searched to identify unpublished or policy-relevant studies [ 19 ]. Search Strategy The search strategy combined controlled vocabulary terms (e.g., MeSH) and free-text keywords related to artificial intelligence, triage, emergency care, and resource-limited settings. An example PubMed search string was: ("artificial intelligence"[MeSH Terms] OR "machine learning" OR "deep learning" OR "natural language processing") AND ("triage"[MeSH Terms] OR "emergency department triage" OR "emergency severity index" OR "acuity assessment") AND ("emergency department" OR "emergency care" OR "emergency services") AND ("low- and middle-income countries" OR "resource-limited" OR "developing countries") Search strategies were adapted for each database. All searches were conducted between June 1 and September 15, 2025. Selection of Sources of Evidence All records identified from the searches were imported into EndNote X9 for deduplication, then uploaded into Rayyan for screening. Two reviewers independently screened titles and abstracts, followed by a full-text review for eligibility. Disagreements were resolved through discussion or third-party adjudication [ 20 ]. Critical Appraisal for Risk of Bias/Quality Assessment Individual Sources The methodological quality and risk of bias of the included studies were evaluated using validated critical appraisal tools appropriate to each study design. Quantitative studies were assessed with the Joanna Briggs Institute (JBI) Critical Appraisal Checklists, which evaluate aspects such as sampling adequacy, outcome measurement validity, and confounding control. Qualitative and mixed-methods studies were appraised using the CASP (Critical Appraisal Skills Programmer) qualitative checklist, focusing on credibility, transferability, and methodological rigor. 22]. Each study was independently appraised by two reviewers, with discrepancies resolved through consensus. While no studies were excluded based solely on quality, the appraisal findings informed the interpretation of evidence strength and confidence in synthesized outcomes. Although scoping reviews do not require exclusion based on quality, an appraisal was undertaken to contextualize findings. Strategy for Data Synthesis Data from included studies were synthesized using both quantitative descriptive and qualitative thematic approaches in accordance with the PRISMA 2020 guidelines and the Joanna Briggs Institute (JBI) framework for scoping reviews. Extracted data were tabulated to summarize study characteristics, AI model types, triage outcomes, and contextual factors relevant to emergency care in resource-limited settings. Quantitative data (e.g., triage accuracy, waiting time reduction, efficiency metrics) were summarized using descriptive statistics, including frequencies, percentages, and mean or median values, to identify trends across studies. When comparable outcome measures were available, effect sizes (e.g., mean differences, relative risks) were reported narratively rather than pooled due to heterogeneity in study designs and metrics. Qualitative and mixed-methods findings were analyzed through inductive thematic synthesis, identifying recurring patterns related to feasibility, clinical acceptance, ethical concerns, and implementation barriers. Data synthesis focused on mapping evidence to key review objectives, efficiency, safety, feasibility, and equity, while highlighting research gaps specific to low- and middle-income countries (LMICs). Findings were integrated into a narrative summary supported by evidence tables and visual figures (e.g., PRISMA flow diagram, outcome distribution charts) to ensure transparency and reproducibility of synthesis methods, consistent with PRISMA and JBI reporting standards. Data Charting and Extraction Data Extraction and Management A structured data extraction process was employed to ensure consistency and transparency across all included studies. A standardized data charting form was developed in Microsoft Excel, adapted from the Joanna Briggs Institute (JBI) Scoping Review Manual, to systematically capture relevant study information. The extraction form was pilot tested on a subset of five studies and refined iteratively to improve clarity and comprehensiveness. Data Items Extracted: For each included study, the following details were extracted: Bibliographic information: author(s), year of publication, journal, and country of study. Study characteristics: design, setting (e.g., tertiary hospital, urban/rural emergency department), and population. AI system attributes: algorithm type (machine learning, deep learning, NLP, hybrid), data source, and training/validation methods. Triage context: acuity scales used (e.g., ESI, CTAS, MTS), comparator (e.g., human or standard triage), and deployment environment. Outcomes reported: efficiency metrics (e.g., waiting time, patient flow), diagnostic accuracy (e.g., AUROC, sensitivity, specificity), feasibility, ethical considerations, and equity dimensions. Key findings and conclusions: including summary statistics and main thematic insights. Two reviewers independently extracted data and cross-verified entries to ensure accuracy. Discrepancies were resolved through consensus or consultation with a third reviewer. Where information was incomplete, corresponding authors were contacted where possible for clarification. All extracted data were stored in a shared, version-controlled dataset, with metadata tags to facilitate sorting by study design, region, or outcome domain. This database formed the basis for descriptive tabulation and thematic synthesis. Data management procedures followed the FAIR (Findable, Accessible, Interoperable, and Reusable) data principles, ensuring traceability and reproducibility of the synthesis process. Analysis Approach Due to heterogeneity in study designs, populations, and outcome measures, a thematic synthesis approach was used for qualitative and implementation outcomes, while quantitative findings were summarized descriptively. Meta-analysis was not feasible due to variations in outcome definitions, statistical reporting, and study contexts. Results Study Selection A total of 1,323 records were identified through databases and supplementary searches. After removing 254 duplicates, 1,069 records remained for title and abstract screening. Of these, 874 were excluded for being irrelevant to AI triage in emergency care or not situated in LMIC contexts. 195 full-text articles were assessed for eligibility, and 152 were excluded (67 not AI-driven, 39 not emergency care, 31 not LMIC-relevant, 15 editorials/reviews without primary data). Ultimately, 43 studies were included in the final synthesis. The study selection process is illustrated in the PRISMA 2020 flow diagram (Fig. 1). PRISMA 2020 Flow Diagram Figure 1. PRISMA 2020 flow diagram illustrating the study selection process for a scoping review on AI-driven triage systems in emergency care in resource-limited settings. Numbers reflect identification, screening, eligibility, and inclusion stages, with a total of 43 studies included in the final synthesis. Characteristics of Included Studies Table 1 summarizes the characteristics of the 43 included studies. These comprised: 10 systematic/scoping/narrative reviews synthesizing global evidence on AI triage in EDs. 21 quantitative studies, including RCTs, cross-sectional analyses, validation studies, and pilot projects. 12 qualitative or mixed-methods studies, often exploring staff, patient, and system-level perspectives on implementation in LMIC contexts. Table 1 Characteristics of Included Studies (n = 43) Characteristic Categories n (%) of studies Geographical region HIC (USA, UK, Korea, Germany, etc.) 31 (72%) LMIC (India, Kenya, Nigeria, etc.) 12 (28%) Study design Systematic/Scoping Reviews 10 (23%) Quantitative (RCT, cross-sectional, pilots) 21 (49%) Qualitative/Mixed methods 12 (28%) AI model Machine Learning (ML) 25 (58%) Deep Learning (DL) 8 (19%) Natural Language Processing (NLP) 5 (12%) Hybrid/Rule-based 5 (12%) Outcomes Efficiency gains 26 (60%) Accuracy & safety 19 (44%) Feasibility/acceptability 12 (28%) Ethical/legal/bias 11 (26%) Table 1. Characteristics of Included Studies (n = 43). Summary of study design, setting, geographical distribution, population, AI models used, and primary focus of included studies on AI-driven triage systems in emergency care across resource-limited settings. Geographical Distribution The majority of studies (n = 31, ~ 72%) were conducted in high-income countries (HICs) such as the United States, South Korea, Germany, Spain, and the United Kingdom. Only 12 studies (28%) were based in low- and middle-income countries (LMICs), including India, Kenya, Nigeria, Ghana, Pakistan, Bangladesh, Peru, and Brazil. This underscores the significant evidence gap in LMIC contexts, where the need for efficiency-enhancing triage tools is arguably greatest. Study Designs and Methodological Rigor Among quantitative studies (n = 21), a minority were randomized controlled trials (RCTs) or prospective validation studies (e.g., in China, South Korea, and India), while the majority comprised retrospective or pilot evaluations. Systematic reviews (n = 10) provided a comprehensive synthesis but frequently highlighted methodological heterogeneity and limited external validity. Qualitative studies (n = 12), primarily from LMICs, offered crucial insights into barriers such as staff skepticism, digital illiteracy, workflow integration issues, and infrastructural limitations. AI Models Applied The AI approaches varied considerably: Machine learning (ML) classifiers were most frequently applied (n ≈ 25 studies), demonstrating improved triage accuracy and reduced misclassification. Deep learning (DL) approaches, often applied in large hospital datasets (China, Korea, Brazil), achieved high predictive accuracy for sepsis detection, trauma severity, and triage acuity. Natural language processing (NLP) models (USA, Germany, Nigeria) were leveraged to analyze unstructured triage notes and enhance acuity prediction. Hybrid ML + rule-based approaches (Kenya, India, Pakistan) were more common in LMIC pilots, as they required fewer computational resources and were easier to adapt to constrained settings. Outcomes Reported Across the included studies, four major outcome themes emerged and have been summarized in Table 2: Efficiency and throughput : Approximately 60% of quantitative studies reported reductions in patient wait times, improved throughput, and mitigation of ED overcrowding through AI-enabled triage support (e.g., Simbo.ai pilots, Mednition NLP, India/China ML trials). Accuracy and safety : Several DL and ML models (China, South Korea, Brazil) demonstrated enhanced predictive accuracy for triage prioritization, particularly in early detection of sepsis, trauma, and critical illness. These improvements suggest potential gains in patient safety, though external validation was limited. Feasibility and acceptability : Qualitative studies in LMICs (Kenya, India, Ghana, Nigeria) highlighted moderate feasibility but mixed provider acceptance. Barriers included a lack of digital literacy, fear of algorithmic replacement, and infrastructure gaps. Facilitators included perceived improvements in decision confidence and workload reduction. Ethical, legal, and equity concerns : Reviews and qualitative studies emphasized the risks of algorithmic bias, lack of transparency in proprietary AI tools, and equity concerns when models trained in HIC datasets were applied in LMIC contexts. Table 2 Outcomes of AI-Driven Triage Systems in Emergency Care (n = 43) Outcome Category Number of Studies (n) Examples of Findings Efficiency & Patient Flow 25 Reduced wait times, improved patient streaming (Yi et al., 2024; Simbo.ai, 2025) Clinical Outcomes 14 Faster time-to-treatment, improved morbidity proxies (Abdalhalim et al., 2025) Accuracy & Safety 18 Higher sensitivity/specificity than manual triage, fewer misclassifications (El Arab & Al Moosa, 2025; Porto, 2024) Implementation Barriers 12 Infrastructure gaps, data bias, ethical concerns (Božić, 2023; Friedman et al., 2024) Feasibility in LMICs 16 Hybrid/rule-based AI is more feasible in low-resource settings (Tahernejad et al., 2024) Table 2. Outcomes of AI-Driven Triage Systems in Emergency Care (n = 43). Overview of key reported outcomes, including effects on triage accuracy, patient flow, waiting times, clinical decision support, feasibility in resource-limited settings, and identified challenges or barriers to implementation. Quantitative Impact on Emergency Care Efficiency Across multiple systematic reviews and observational studies, AI-driven triage systems consistently demonstrated measurable improvements in emergency department (ED) efficiency, even in resource-limited settings. Reduced patient wait times were reported in both academic and industry analyses (Tyler et al., 2025; Simbo.ai Blog, 2025; Mednition, 2024), with reductions ranging from 15–40% in high-volume EDs. Shorter length of stay and faster time-to-treatment initiation were found in studies by Abdalhalim et al. (2025), Porto (2024), and Yi et al. (2024). Triage accuracy improvements were observed when AI was combined with natural language processing (NLP) for free-text data (Porto, 2024) and when models were trained on large, diverse datasets (Varughese et al., 2024). In mass-casualty and disaster contexts, AI-enhanced triage significantly improved resource allocation and prioritization speed (Tahernejad et al., 2024), a critical factor in low-resource environments. Clinical Outcomes and Patient Safety Several studies linked efficiency gains to better patient outcomes: Enhanced early detection of critical cases and reduced over-triage rates were documented in El Arab & Al Moosa (2025) and Yi et al. (2024). In military and austere environments, AI-driven autonomous trauma care systems improved survival during the “golden hour” (Philip & Harikumar, 2025). However, Ilick (2022) and Friedman et al. (2024) emphasized that accuracy varied considerably depending on data quality and warned against over-reliance without adequate clinical oversight. Implementation of Barriers and Facilitators in Resource-Limited Settings Barriers : Infrastructure constraints such as unreliable power supply, limited internet connectivity, and the absence of EHR integration tools (Philip & Harikumar, 2025). Data-related issues, including a lack of high-quality annotated datasets for specific populations (Ilick, 2022; Tahernejad et al., 2024). Human factors such as variable clinician trust in AI recommendations (Varughese et al., 2024) and the need for targeted training (El Arab & Al Moosa, 2025). Regulatory and ethical considerations, especially around patient data privacy (Yi et al., 2024; Božić, 2023). Facilitators : Workflow adaptability, several studies (Tyler et al., 2025; Abdalhalim et al., 2025) found that AI triage tools could integrate into existing ED processes with minimal disruption. Portability of some AI platforms, enabling deployment in mobile or rural clinics (Philip & Harikumar, 2025). Commercial readiness of certain industry tools, allowing for faster adoption in facilities with budget flexibility (Simbo.ai Blog, 2025; Mednition, 2024). Selected Study Characteristics A total of 43 studies were included in the final synthesis and captured in Fig. 2. Geographically, the majority were conducted in high-income countries (HICs, n = 31; 72%), while low- and middle-income countries (LMICs, n = 12; 28%) contributed a smaller but growing proportion (Fig. 2, top-left). In terms of study design, nearly half employed quantitative approaches (n = 21; 49%), followed by qualitative or mixed-methods studies (n = 12; 28%), and systematic/scoping reviews (n = 10; 23%) (Fig. 2, top-right). Across the included literature, the most frequently applied AI approaches were machine learning (ML, n = 25; 58%), followed by deep learning (DL, n = 8; 19%), natural language processing (NLP, n = 5; 12%), and hybrid or rule-based models (n = 5; 12%) (Figure X, bottom-left). Reported outcomes clustered around four major domains: efficiency gains such as reduced wait times and improved triage accuracy (n = 26; 60%), accuracy and patient safety improvements (n = 19; 44%), feasibility and clinical acceptability (n = 12; 28%), and ethical or equity-related considerations including algorithmic bias (n = 11; 26%) (Fig. 2, bottom-right). Together, these findings highlight the predominance of HIC-driven research and machine learning applications, while emphasizing critical evidence gaps in LMIC contexts, particularly regarding implementation feasibility, equity, and long-term impact. (Top-left) Geographic distribution of studies, with most conducted in high-income countries (HICs, 72%) and a smaller proportion in low- and middle-income countries (LMICs, 28%). (Top-right) Study designs represented: quantitative (49%), qualitative/mixed-methods (28%), and systematic/scoping reviews (23%). (Bottom-left) AI approaches applied: machine learning (58%), deep learning (19%), natural language processing (12%), and hybrid/rule-based models (12%). (Bottom-right) Reported outcomes clustered into efficiency and workflow gains (60%), accuracy and patient safety (44%), feasibility and clinical acceptability (28%), and ethical/equity considerations (26%). Discussion This scoping review synthesized evidence from 43 studies on AI-driven triage systems in emergency care. Overall, findings suggest that AI offers significant potential for efficiency, accuracy, and workflow optimization in triage, but challenges persist in feasibility, ethical governance, and the applicability of evidence to low- and middle-income countries (LMICs). Interpretation of Findings in Resource-Limited Settings Emergency departments in LMICs face chronic shortages of staff, overcrowding, and inadequate infrastructure [ 23 ]. AI-driven triage tools can theoretically automate early decision-making, reduce delays, and compensate for workforce gaps [ 24 ]. However, the majority of studies reviewed originated from high-income settings, with only a minority evaluating AI in resource-limited contexts [ 25 ]. Even where LMIC-related studies exist, they are often retrospective or modeling-based, rather than prospective evaluations in real-world environments [ 26 ]. This underlines a major evidence gap: while AI’s theoretical utility in LMICs is high, empirical validation remains scarce. Comparison with High-Resource Settings In high-income countries, AI triage has been shown to reduce wait times, improve patient flow, and enhance the prediction of critical outcomes such as ICU admission or mortality [ 27 ]. These benefits are facilitated by robust digital infrastructure, interoperability with electronic health records, and strong regulatory frameworks [ 28 ]. By contrast, LMICs lack comparable infrastructure and face challenges such as unreliable internet connectivity and limited health data digitization [ 29 ]. This disparity suggests that successful AI triage adoption in LMICs requires adaptation to local realities rather than direct transfer of systems from high-resource settings. Efficiency and Workflow Optimization Many studies demonstrated that AI-driven triage could streamline workflow, improve throughput, and prioritize critical patients more effectively than manual triage [ 30 ]. This is particularly relevant in disaster or pandemic contexts, where rapid prioritization of cases is essential [ 31 ]. Yet in LMICs, infrastructure bottlenecks (e.g., insufficient resuscitation capacity) may limit the extent to which efficiency gains translate into improved outcomes [ 32 ]. The most consistently reported outcome across studies was improved efficiency, including reductions in waiting times, faster allocation of acuity levels, and enhanced throughput. Machine learning, based models demonstrated the ability to process large volumes of patient data in real time, outperforming traditional manual triage in several high-income settings. Accuracy and Patient Safety Improved acuity assignment accuracy was a consistent finding, with several AI models outperforming human judgment in predicting adverse outcomes [ 33 , 46 ]. This has direct implications for patient safety, especially in crowded emergency departments. However, concerns remain about algorithmic generalizability: most models were trained on patient populations from high-income settings, raising the risk of mis-triage when applied to LMIC populations with different epidemiological profiles [ 34 ]. Feasibility and Clinical Acceptability Only a minority of studies assessed feasibility, clinician acceptability, or integration into workflows [ 35 ]. Clinicians voiced concerns about interpretability and liability in decision-making, echoing broader debates on trust in AI [ 36 ]. In LMICs, additional barriers such as lack of training, limited IT support, and costs of deployment exacerbate feasibility challenges [ 37 ]. Without addressing these, the real-world impact of AI triage in resource-limited settings may remain minimal. Ethical and Legal Considerations Ethical issues were consistently underexplored. Studies highlighted risks of bias, inequity, and data privacy breaches [ 38 ]. In LMICs, reliance on algorithms trained in high-resource contexts may exacerbate inequities if local population characteristics are not represented [ 39 ]. Furthermore, weak regulatory and legal safeguards in some LMICs heighten risks of misuse of sensitive health data [ 40 ]. Addressing these issues requires transparent model development, inclusion of diverse datasets, and creation of robust governance frameworks. Limitations of Included Studies This review revealed substantial heterogeneity in study designs, populations, and outcomes [ 41 ]. Many studies were retrospective, with few prospective or randomized evaluations, limiting the strength of causal inferences [ 42 ]. Additionally, reporting was often inconsistent, particularly regarding cost-effectiveness and unintended harm. The overwhelming dominance of high-income settings raises concerns about generalizability, especially in LMICs [ 43 ]. Implications for Clinical Practice and Policy For clinicians, AI-driven triage offers the promise of earlier identification of critically ill patients and reduced cognitive burden [ 44 ]. For policymakers, successful adoption could enhance health system resilience, particularly during crises. However, policies must ensure investments in infrastructure, training, and regulation before large-scale implementation in LMICs [ 46 ]. Otherwise, AI triage risks widen rather than narrowing global disparities in emergency care. Recommendations for Future Research To bridge the evidence gaps, future research should: Conduct prospective, context-specific studies in LMICs. Include cost-effectiveness and implementation of science analyses. Address clinician acceptability and patient trust. Incorporate ethically grounded governance frameworks addressing bias, accountability, and transparency. Foster cross-disciplinary collaboration between AI developers, clinicians, policymakers, and communities. Conclusion This scoping review demonstrates that AI-driven triage systems hold considerable promises for enhancing efficiency, accuracy, and patient safety in emergency care. However, the current evidence base is heavily skewed toward high-income settings, with only limited empirical evaluation in resource-limited and LMIC contexts. While AI has the potential to mitigate workforce shortages and streamline patient flow in these settings, its successful adoption requires adaptation to local infrastructures, rigorous validation in diverse populations, and governance frameworks that ensure equity, transparency, and accountability. The findings highlight both an opportunity and a caution: without context-sensitive implementation, AI triage may risk reinforcing global health inequities rather than alleviating them. Policymakers, clinicians, and researchers must therefore collaborate to design, evaluate, and regulate AI tools that are fit-for-purpose in low-resource environments. Future research should prioritize prospective evaluations, feasibility studies, and cost-effectiveness analyses in LMICs, ensuring that innovations in AI triage translate into meaningful and equitable improvements in emergency care outcomes worldwide. Abbreviations Abbreviation Full Term / Definition AI Artificial Intelligence ML Machine Learning DL Deep Learning NLP Natural Language Processing ANN Artificial Neural Network CNN Convolutional Neural Network MLP Multilayer Perceptron LR Logistic Regression SVM Support Vector Machine RF Random Forest ED Emergency Department ER Emergency Room LMIC Low- and Middle-Income Country HIC High-Income Country EHR Electronic Health Record AI-ED Artificial Intelligence–Enhanced Emergency Department (used contextually) WHO World Health Organization WB World Bank CASP Critical Appraisal Skills Programme JBI Joanna Briggs Institute PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses PRISMA-ScR Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews PROSPERO International Prospective Register of Systematic Reviews OSF Open Science Framework AI-Triage Artificial Intelligence–Driven Triage System QoC Quality of Care AI-ML System Artificial Intelligence and Machine Learning System HRQoL Health-Related Quality of Life NCD Noncommunicable Disease AIH Artificial Intelligence in Healthcare CINAHL Cumulative Index to Nursing and Allied Health Literature (referenced in method description) PRISMA 2020 Updated version of PRISMA guideline (2020 update) IEEE Institute of Electrical and Electronics Engineers HIS Health Information System MoH Ministry of Health AI Bias Algorithmic Bias or Artificial Intelligence Bias HFE Human Factors Engineering SDG Sustainable Development Goal ICT Information and Communication Technology RCT Randomized Controlled Trial CAD Computer-Aided Decision System AI-CDSS Artificial Intelligence–Based Clinical Decision Support System BMJ British Medical Journal (referenced in style and format discussions) ED Triage Emergency Department Triage AI-Triage Model Artificial Intelligence–Driven Clinical Triage Model PHC Primary Health Care AI Readiness The preparedness of healthcare systems for integrating AI tools AI Governance Ethical, regulatory, and policy structures governing AI use AI Equity Equitable deployment and access to AI technologies across income settings AI-Enabled Workflow Integration of AI tools into clinical operational workflows AI-Driven Decision Support Automated or semi-automated assistance for clinical prioritization and triage Declarations Ethics approval and consent to participate This study does not involve human participants, and ethical approval was not required. Consent for publication Not applicable Data Availability Statement All data extracted during this scoping review are included in the published article (and its supplementary information files). The search strategies, data extraction forms, and worksheets are available from the corresponding author upon reasonable request. Conflicts of Interest The authors declare no conflicts of interest Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Author Contributions JA conceptualized the study. C.A.A and D.O.O contributed to the design of the study. O.S.A and C.E contributed to the acquisition and analysis, while B.O.J and P.O.O contributed to the interpretation of data. J.A, C.A.A, and D.O.O drafted the study and O.S.A, C.E. and B.O.J substantively revised it. J.A developed the search strategy with consultation from P.O.O and C.A.A. D.O.O and O.S.A screened, assessed the eligibility, and assessed the quality of the included studies with consultation from J.A and P.O.O, analyzed the data and created the figures with consultation from C.E and O.S.A. J.A is responsible for the data management and storage. All authors reviewed the final manuscript and approved the final version for submission. All authors have agreed both to be personally accountable for the author's own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which they were not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature. Acknowledgments The authors acknowledge the contributions of AwaDoc for assistance with search strategy development, members of the advisory panel for guidance throughout the review process, and all researchers whose work was synthesized in this review for their contributions to advancing knowledge regarding AI applications for triage system in emergency care. Authors Information Joy Aifuobhokhan. MD,* 1 0009-0005-9521-905X, Digital Health and Research, Lakeshore Cancer Centre, Lagos, Nigeria, Chukwuemeka Abraham Agbarakwe. MD 2 , 0000-0002-7267-3699, Emergency Medicine, Calvary Specialist Hospital, Port Harcourt, Nigeria Deborah Oladunmolu Oduguwa. MD 3 , 0009-0003-8422-6153, Family Medicine, Babcock University Teaching Hospital, Ilishan-Remo, Nigeria Oseni, Olukayode A. MD 4 , 0009-0000-0148-5650, Urology, Lagos University Teaching Hospital Chimdi Eleweke 5 , 0009-0003-1907-3023, Emergency Medicine, Babcock University Teaching Hospital, Ilishan-Remo, Nigeria Boluwatife Oluwafayoyimika Johnson. MD 6 , 0009-0004-2767-4547, Internal Medicine, Babcock University Teaching Hospital, Ilishan-Remo, Nigeria Precious Oluwadamilola Opawoye 7 0009-0003-0626-6299, Research, Lakeshore Cancer Center, Lagos, Nigeria Corresponding Author: Joy Aifuobhokhan.MD [email protected] References Hsia RY, Razzak J, Tsai AC, Hirshon JM. Placing emergency care on the global agenda. Ann Emerg Med. 2010;56(2):142–9. 10.1016/j.annemergmed.2010.03.028 . Obermeyer Z, Abujaber S, Makar M, Stoll S, Kayden S, Wallis LA, Reynolds T. Emergency care in 59 low- and middle-income countries: a systematic review. Bull World Health Organ. 2015;93(8):577–86. 10.2471/BLT.14.148338 . Reynolds TA, Sawe HR, Rubiano AM, Shin SD, Wallis L, Mock CN. Strengthening health systems to provide emergency care. Disease Control Priorities, Third Edition (Volume 9): Injury Prevention and Environmental Health. The World Bank; 2017. 10.1596/978-1-4648-0522-6_ch13 . Razzak J, Kellermann AL. Emergency medical care in developing countries: is it worthwhile? Bull World Health Organ. 2002;80(11):900–5. URL: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2567679/ Gilboy N, Tanabe T, Travers D, Rosenau AM. Emergency Severity Index (ESI): A Triage Tool for Emergency Department Care, Version 4. Implementation Handbook 2012 Edition. Agency for Healthcare Research and Quality (US); 2012. URL: https://www.ahrq.gov/esindex.html Van der Wulp I, van Baar ME, Schrijvers AJP. Reliability and validity of the Manchester Triage System in a general emergency department patient population in the Netherlands: results of a simulation study. Emerg Med J. 2008;25(7):431–4. 10.1136/emj.2007.055228 . Joy Aifuobhokhan, Tobiloba PO, Chijioke CE. Artificial Intelligence in Cancer Diagnosis: A Scoping Review of Global Innovation and African Implementation. World J Adv Res Reviews. 2025;28(01):449–68. https://doi.org/10.30574/wjarr.2025.28.1.3444 . Article DOI. Joy Aifuovhokhan, Olajire OO, Emmanuel TP, Artificial Intelligence - Clinical Decision Support Systems. From brain drain to health system gain. World J Adv Res Reviews. 2025;28(01):1482–93. https://doi.org/10.30574/wjarr.2025.28.1.3486 . Article DOI. Aifuobhokhan J, Hussain AK, Ekechi CC, Olanrewaju AO, Afuadajo E, Bowale DA, Inioluwa OM. Predicting Thyroid Cancer Recurrence Using Machine Learning: An Artificial Intelligence Approach to Clinical Oncology. Int J Adv Artif Intell Mach Learn. 2025;2(3):135–48. https://doi.org/10.58723/ijaaiml.v2i3.469 . Fernandes CM, Tanabe P, Gilboy N, Johnson LA, McNair RS, Rosenau AM, Sawchuk P, Thompson DA. Five-level triage: a report from the ACEP/ENA Five-level Triage Task Force. J Emerg Nurs. 2005;31(1):39–50. 10.1016/j.jen.2004.11.002 . Tyler S, Olis M, Aust N, Patel L, Simon L, Triantafyllidis C, Patel V, Lee DW, Ginsberg B, Ahmad H, Jacobs RJ. Use of Artificial Intelligence in Triage in Hospital Emergency Departments: A Scoping Review. Cureus. 2024;16(6):e248163. PMID: 38854295; PMCID: PMC11158416. Ahmed AAZ, Ahmed SNNA, Ezzelarab AMD, Mustafa M, Albasheer MGA, Ahmed REA, Elsayed MBGE. Clinical Impact of Artificial Intelligence-Based Triage Systems in Emergency Departments: A Systematic Review. Cureus. 2025;17(2):e85667. 10.7759/cureus.85667 . El Arab RA, Al Moosa OA. The role of AI in emergency department triage: An integrative systematic review. Intensive Crit Care Nurs. 2025;79:104058. 10.1016/j.iccn.2025.104058 . Tahernejad A, Sahebi A, Abadi ASS, et al. Application of artificial intelligence in triage in emergencies and disasters: a systematic review. BMC Public Health. 2024;24:3203. 10.1186/s12889-024-20447-3 . Porto BM. Improving triage performance in emergency departments using machine learning and natural language processing: a systematic review. BMC Emerg Med. 2024;24:219. 10.1186/s12873-024-01135-2 . Yi N, Baik D, Baek G. The effects of applying artificial intelligence to triage in the emergency department: A systematic review of prospective studies. J Nurs Scholarsh. 2024;56(4):765–77. 10.1111/jnu.13024 . Friedman AB, Delgado MK, Weissman GE. Artificial Intelligence for Emergency Care Triage—Much Promise, but Still Much to Learn. JAMA Netw Open. 2024;7(5):e248857. 10.1001/jamanetworkopen.2024.8857 . Ilick J. Challenges in evaluating the accuracy of AI-containing digital triage systems: A systematic review. PLoS ONE. 2023;18(1):e0279636. 10.1371/journal.pone.0279636 . Božić V. Using Artificial Intelligence in Triage Process: Benefits, Challenges, and Considerations. Res Square Preprint. 2023. URL. Simbo.ai. The Impact of Artificial Intelligence on Triage Systems in Emergency Departments: Enhancing Efficiency and Patient Care. Simbo Blog. 2023. Mednition. Emergency Triage: Revolutionizing Patient Care with AI for Better Outcomes. Mednition Blog. 2023. URL: https://mednition.com/blog/emergency-triage-revolutionizing-patient-care-with-ai-for-better-outcomes/ Peters MDJ, Godfrey CM, McInerney P, Baldini Soares C, Khalil H, Parker D. Methodology for JBI scoping reviews. In: Aromataris E, Munn Z, editors. Joanna Briggs Institute Reviewer’s Manual. The Joanna Briggs Institute; 2017. Available from: https://reviewersmanual.joannabriggs.org/ Munn Z, Peters MDJ, Stern C, Tufanaru C, McArthur A, Aromataris E. Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Med Res Methodol. 2018;18:143. 10.1186/s12874-018-0611-x . Tricco AC, Lillie E, Zarin W, O’Brien KK, Colquhoun H, Levac D, et al. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Ann Intern Med. 2018;169(7):467–73. 10.7326/M18-0850 . Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. 10.1136/bmj.n71 . Critical Appraisal Skills Programme (CASP). CASP Checklists. Oxford: CASP. 2023. Available from Joanna Briggs Institute (JBI). JBI Critical Appraisal Tools. Adelaide: JBI; 2023. Available from: https://jbi.global/critical-appraisal-tools Bramer WM, Rethlefsen ML, Kleijnen J, Franco OH. Optimal database combinations for literature searches in systematic reviews: a prospective exploratory study. Syst Rev. 2017;6:245. 10.1186/s13643-017-0644-y . Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editors. Cochrane Handbook for Systematic Reviews of Interventions. Version 6.3 (updated February 2022). Cochrane; 2022. Available from. https://training.cochrane.org/handbook Porto BM. Improving triage performance in emergency departments using machine learning and natural language processing: a systematic review. BMC Emerg Med. 2024;24:219. 10.1186/s12873-024-01135-2 . Yi N, Baik D, Baek G, et al. The effects of applying artificial intelligence to triage in the emergency department: A systematic review of prospective studies. J Nurs Scholarsh. 2024;56(5):123–34. 10.1111/jnu.13024 . Friedman AB, Delgado MK, Weissman GE. Artificial Intelligence for Emergency Care Triage—Much Promise, but Still Much to Learn. JAMA Netw Open. 2024;7(5):e248857. 10.1001/jamanetworkopen.2024.8857 . Ahmed AAZ, Ahmed SNNA, Ezzelarab AMD, Mustafa M, Albasheer MGA, Ahmed REA, Elsayed MBGE. Clinical Impact of Artificial Intelligence-Based Triage Systems in Emergency Departments: A Systematic Review. Cureus. 2025;17(3):e85667. 10.7759/cureus.85667 . Tyler S, Olis M, Aust N, Patel L, Simon L, Triantafyllidis C et al. Use of Artificial Intelligence in Triage in Hospital Emergency Departments: A Scoping Review. Cureus. 2025;17(4):e11158416. PMID:38854295; PMCID:PMC11158416. Available from: https://www.cureus.com/articles/248163 El Arab RA, Al Moosa OA. The role of AI in emergency department triage: An integrative systematic review. Intensive Crit Care Nurs. 2025;79:104058. 10.1016/j.iccn.2025.104058 . Tahernejad A, Sahebi A, Abadi ASS, et al. Application of artificial intelligence in triage in emergencies and disasters: a systematic review. BMC Public Health. 2024;24:3203. 10.1186/s12889-024-20447-3 . Božić V. Using Artificial Intelligence in Triage Process: Benefits, Challenges, and Considerations. 2023. Available from: https://www.researchgate.net/publication/371911400_Using_Artifical_Intelligence_in_Triage_Process_Benefits_Challenges_and_Considerations Ilick J. Challenges in evaluating the accuracy of AI-containing digital triage systems: A systematic review. PLoS ONE. 2022;17(12):e0279636. 10.1371/journal.pone.0279636 . Arab RAE, Omayma Abdulaziz Al Moosa. The role of AI in emergency department triage: an integrative systematic review. Intensive Crit Care Nurs. 2025;79:104058. 10.1016/j.iccn.2025.104058 . Mednition. Emergency Triage: Revolutionizing Patient Care with AI for Better Outcomes. 2024. Available from: https://mednition.com/blog/emergency-triage-revolutionizing-patient-care-with-ai-for-better-outcomes/ Simbo.AI. The Impact of Artificial Intelligence on Triage Systems in Emergency Departments: Enhancing Efficiency and Patient Care. 2024. Available from: https://www.simbo.ai/blog/the-impact-of-artificial-intelligence-on-triage-systems-in-emergency-departments-enhancing-efficiency-and-patient-care-2687317/ Simbo.AI. The Impact of AI-Driven Triage Systems on Reducing Patient Wait Times during Critical Healthcare Scenarios. 2024. Available from: https://www.simbo.ai/blog/the-impact-of-ai-driven-triage-systems-on-reducing-patient-wait-times-during-critical-healthcare-scenarios-2096665/ Hsia R, Mbembati NA, Macfarlane S, Kruk ME. Access to emergency and surgical care in sub-Saharan Africa: the infrastructure gap. Health Policy Plan. 2012;27(3):234–44. 10.1093/heapol/czr023 . Alami H, Rivard L, Lehoux P, Hoffman SJ, Cadeddu SB, Savoldelli M, et al. Artificial intelligence in health care: laying the foundation for responsible, sustainable, and inclusive innovation in low- and middle-income countries. Global Health. 2020;16:52. 10.1186/s12992-020-00584-1 . Leslie D, Mazumder A, Peppin A, Wolters MK, Hagerty A. Does AI stand for augmenting inequality in the era of COVID-19 healthcare? BMJ. 2021;372:n242. 10.1136/bmj.n242 . WHO. Ethics and Governance of Artificial Intelligence for Health: WHO Guidance. World Health Organization. 2021. Available from: European Commission. Ethics Guidelines for Trustworthy AI. High-Level Expert Group on AI. 2019. Available from: https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai O’Sullivan ED, Schofield SJ. Cognitive bias in clinical medicine. J R Coll Physicians Edinb. 2018;48(3):225–32. 10.4997/JRCPE.2018.306 . Jobin A, Ienca M, Vayena E. The global landscape of AI ethics guidelines. Nat Mach Intell. 2019;1:389–99. 10.1038/s42256-019-0088-2 . Additional Declarations No competing interests reported. Supplementary Files EmergencySupplementaryFile1.xlsx EmergencySupplementaryMaterial2SearchStrategy.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8097974","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":549493494,"identity":"c870aa9c-d6bd-4b9e-a651-5528d43b7332","order_by":0,"name":"Joy Aifuobhokhan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCUlEQVRIiWNgGAWjYBACxgYgkWDDwGAgwcDMDGTLgUQPPCCoJQ2hxRisJYGgXUhaEkFmMODTwtzeY/jhQcJhOXPp5sPGBTV30ueHHX4ItMVOTrcBh8N6zhhLJCQcNraccyw5ecaxZ7kbb6cZALUkG5sdwKFlRo6BROKPw4kbbuQYH+ZhO5y7cXYCSMuBxG24tRj/ANoC1fLvcLrh7PQPhLSYScC0JPO2HU6Ql84hYEvPsTKLhIR0Y8sZacnGvH2HDTdI5xQcSDDA7RfD9ubNN38kWMuZSyQflub5dlhefnb65g8fKuzkcGpp4DBAFTEAqzTAohYK5BnYH6CJNOBWPQpGwSgYBSMTAAAnLWdAUm6m3AAAAABJRU5ErkJggg==","orcid":"","institution":"Lakeshore Cancer Centre","correspondingAuthor":true,"prefix":"","firstName":"Joy","middleName":"","lastName":"Aifuobhokhan","suffix":""},{"id":549493495,"identity":"065290b5-cbcf-475e-a804-97dd2f3810c2","order_by":1,"name":"Chukwuemeka Abraham Agbarakwe","email":"","orcid":"","institution":"Calvary Specialist Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chukwuemeka","middleName":"Abraham","lastName":"Agbarakwe","suffix":""},{"id":549493496,"identity":"1acd627c-3733-4c25-af97-44f7f7a10965","order_by":2,"name":"Deborah Oladunmolu Oduguwa","email":"","orcid":"","institution":"Babcock University Teaching Hospital","correspondingAuthor":false,"prefix":"","firstName":"Deborah","middleName":"Oladunmolu","lastName":"Oduguwa","suffix":""},{"id":549493497,"identity":"b65042fc-05d8-4517-b254-67d0472af848","order_by":3,"name":"Oseni Olukayode A","email":"","orcid":"","institution":"Lagos University Teaching Hospital","correspondingAuthor":false,"prefix":"","firstName":"Oseni","middleName":"Olukayode","lastName":"A","suffix":""},{"id":549493498,"identity":"9590c764-4651-4dee-b806-81346725d173","order_by":4,"name":"Chimdi Eleweke","email":"","orcid":"","institution":"Babcock University Teaching Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chimdi","middleName":"","lastName":"Eleweke","suffix":""},{"id":549493499,"identity":"d3d0d985-915e-4661-bfe0-8d43578d789d","order_by":5,"name":"Boluwatife Oluwafayoyimika Johnson","email":"","orcid":"","institution":"Babcock University Teaching Hospital","correspondingAuthor":false,"prefix":"","firstName":"Boluwatife","middleName":"Oluwafayoyimika","lastName":"Johnson","suffix":""},{"id":549493500,"identity":"adae33cb-c7de-42c7-8c3f-b735c33fee1b","order_by":6,"name":"Precious Oluwadamilola Opawoye","email":"","orcid":"","institution":"Lakeshore Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Precious","middleName":"Oluwadamilola","lastName":"Opawoye","suffix":""}],"badges":[],"createdAt":"2025-11-12 15:38:38","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8097974/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8097974/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":96968975,"identity":"192f4ac0-97c8-4c78-9c16-bbcd8afb3017","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":242695,"visible":true,"origin":"","legend":"","description":"","filename":"EvaluatingtheImpactofArtificialIntelligence.docx","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/e2445bc82424545e87367563.docx"},{"id":97138560,"identity":"4db99cb1-f70c-4f20-b9bb-5ba134c7470b","added_by":"auto","created_at":"2025-12-01 09:59:03","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10307,"visible":true,"origin":"","legend":"","description":"","filename":"56d2a67d6c2b49498cca112f94c30a04.json","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/3a0248d5bad80e34e6e1a0e9.json"},{"id":96968973,"identity":"b6c392d2-bbef-4596-bf0a-67c6bcbc9530","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":19893,"visible":true,"origin":"","legend":"","description":"","filename":"EmergencySupplementaryFile1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/249600eb2d26c13b59ce27bf.xlsx"},{"id":96968981,"identity":"0733e647-fa10-449d-aad8-ff06c4d35ee1","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":26227,"visible":true,"origin":"","legend":"","description":"","filename":"EmergencySupplementaryMaterial2SearchStrategy.docx","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/e7559f34861070b3a9f2fcee.docx"},{"id":96968983,"identity":"b3c2fda6-3f99-4531-8494-bde1117ac792","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"xml","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":134521,"visible":true,"origin":"","legend":"","description":"","filename":"56d2a67d6c2b49498cca112f94c30a041enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/ab3ae00b1c9c98a9f3003ff2.xml"},{"id":96968984,"identity":"569a17a7-3f1a-4422-97b7-2e75fe205ec2","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":28792,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/da4fefead6c4ab3c692f2716.png"},{"id":96968979,"identity":"b137e24b-748a-4339-a687-f0b72c94154b","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13343,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/0c77bbd42fe97ee675915642.png"},{"id":96968980,"identity":"956f7d63-94ab-4434-a96c-d8b6224d508e","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"xml","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":131688,"visible":true,"origin":"","legend":"","description":"","filename":"56d2a67d6c2b49498cca112f94c30a041structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/a6ebd535e24490341b9ba70a.xml"},{"id":96968982,"identity":"b028b326-45ec-44ed-971e-2ee2b84d2c80","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"html","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":147502,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/5a94151d72727f360edda44c.html"},{"id":96968978,"identity":"c0d04fb9-9152-4161-b622-369342a17a6f","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":95527,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003ePRISMA 2020 Flow Diagram\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ePRISMA 2020 flow diagram illustrating the study selection process for a scoping review on AI-driven triage systems in emergency care in resource-limited settings. Numbers reflect identification, screening, eligibility, and inclusion stages, with a total of 43 studies included in the final synthesis.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/4219e9bebd9bcec2ceadbdc9.png"},{"id":96968976,"identity":"0422b308-3055-4fb1-b88f-7ade2586a725","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":81054,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution and characteristics of included studies (n = 43).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e(Top-left) Geographic distribution of studies, with most conducted in high-income countries (HICs, 72%) and a smaller proportion in low- and middle-income countries (LMICs, 28%). (Top-right) Study designs represented: quantitative (49%), qualitative/mixed-methods (28%), and systematic/scoping reviews (23%). (Bottom-left) AI approaches applied: machine learning (58%), deep learning (19%), natural language processing (12%), and hybrid/rule-based models (12%). (Bottom-right) Reported outcomes clustered into efficiency and workflow gains (60%), accuracy and patient safety (44%), feasibility and clinical acceptability (28%), and ethical/equity considerations (26%).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/05a6d948c580846c0487af1e.png"},{"id":101188591,"identity":"f9d7397d-953a-474c-b706-041438dde8de","added_by":"auto","created_at":"2026-01-27 06:41:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1669933,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/97720c41-fc58-4ac3-9415-8fb7e61b1c57.pdf"},{"id":96968972,"identity":"1e7032ed-6da1-4ca9-a9ca-cd5714cda3f5","added_by":"auto","created_at":"2025-11-28 07:07:24","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":19893,"visible":true,"origin":"","legend":"","description":"","filename":"EmergencySupplementaryFile1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/ab73c07fdbe847e3fc4d16da.xlsx"},{"id":97136957,"identity":"cd13a923-f8a4-4c36-80ad-0cddcd07bedc","added_by":"auto","created_at":"2025-12-01 09:57:12","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":26227,"visible":true,"origin":"","legend":"","description":"","filename":"EmergencySupplementaryMaterial2SearchStrategy.docx","url":"https://assets-eu.researchsquare.com/files/rs-8097974/v1/a3c28030c86f942c4438f322.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Evaluating the Impact of Artificial Intelligence-Driven Triage Systems on Emergency Care Efficiency in Resource-Limited Settings: A Scoping Review","fulltext":[{"header":"Background","content":"\u003cdiv id=\"Sec2\" class=\"Section2\"\u003e\u003ch2\u003eOverview of Emergency Care in Resource-Limited Settings\u003c/h2\u003e\u003cp\u003eEmergency care plays a critical role in reducing preventable deaths from acute illnesses and injuries, yet in many low- and middle-income countries (LMICs), the infrastructure to deliver timely, high-quality emergency services remains underdeveloped [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. LMIC emergency departments (EDs) often operate under conditions of chronic overcrowding, insufficient staffing, limited diagnostic capacity, and scarce specialist availability [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. These challenges are compounded by high burdens of trauma, communicable diseases, and emerging non-communicable emergencies such as cardiovascular disease and stroke [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Without efficient patient flow, even patients with treatable conditions face prolonged delays, which directly contribute to poor outcomes and increased mortality [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eRole of Triage Systems in Efficiency and Outcomes\u003c/h2\u003e\u003cp\u003eTriage, the process of categorizing patients according to urgency- has long been recognized as a cornerstone of emergency medicine [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Effective triage enables rapid identification of critically ill patients, reduces waiting times, and optimizes allocation of scarce resources [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Evidence from both high- and low-resource settings shows that standardized triage protocols such as the Emergency Severity Index (ESI), the Manchester Triage System (MTS), and the South African Triage Scale (SATS) are associated with improved throughput, reduced ED crowding, and better patient outcomes [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. However, these systems rely heavily on trained personnel and uninterrupted workflows, conditions that are difficult to maintain in resource-limited environments, where overburdened staff, high patient volumes, and variability in clinical experience often led to inconsistent application [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eAI Advancements in Triage and Relevance to LMICs\u003c/h3\u003e\n\u003cp\u003eRecent advancements in artificial intelligence (AI), including machine learning and natural language processing, have shown promise in automating and enhancing triage decision-making [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. AI-driven triage tools can analyze structured and unstructured clinical data in real time to assign acuity scores, predict deterioration, and recommend interventions [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Studies have demonstrated improvements in triage accuracy, reductions in ED length of stay, and enhanced prioritization of high-risk patients when AI is integrated into triage workflows [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Importantly, in LMICs where human resources are scarce, AI systems could help mitigate workforce shortages, standardize decision-making, and improve access to timely care [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Early trials, such as those adapting the Kampala Trauma Score into machine learning\u0026ndash;based models, have already shown superior predictive performance in trauma cases compared to traditional methods in low-resource environments [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e\n\u003ch3\u003eGap in Current Research and Rationale for Study\u003c/h3\u003e\n\u003cp\u003eDespite promising results, most AI triage research has been conducted in high-income countries with robust digital infrastructure [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Evidence on AI triage in LMIC EDs remains limited, fragmented, and contextually narrow [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. There is a lack of large-scale, prospective evaluations assessing AI\u0026rsquo;s impact on core operational metrics (e.g., patient throughput, wait times, resource utilization) and patient-centered outcomes (e.g., morbidity, mortality) in these settings [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Additionally, barriers such as limited interoperability with existing health information systems, insufficient local datasets for algorithm training, and clinician skepticism towards AI tools remain underexplored [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Understanding these factors is critical for informing scalable, equitable, and sustainable implementation strategies in LMICs.\u003c/p\u003e\n\u003ch3\u003eResearch Question and Objectives\u003c/h3\u003e\n\u003cp\u003eThis study seeks to answer the central question:\u003c/p\u003e\u003cp\u003eHow do AI-driven triage systems impact emergency care efficiency and patient outcomes in resource-limited settings?\u003c/p\u003e\u003cp\u003eThe specific objectives are to:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eEvaluate the effects of AI-driven triage systems on ED operational efficiency, including patient throughput and waiting times.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eAssess the impact of AI triage on clinical outcomes, including time to intervention, morbidity, and mortality.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eIdentify facilitators and barriers to implementing AI triage systems in LMIC EDs.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eProvide evidence-based recommendations for integrating AI into emergency care workflows in resource-limited contexts.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e"},{"header":"Methodology","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eResearch Design, Protocol, and Registration\u003c/h2\u003e\u003cp\u003eThis study was conducted as a scoping review in accordance with the Joanna Briggs Institute (JBI) methodology for scoping reviews [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. The scoping protocol was developed in alignment with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. The review was prospectively registered with the International Prospective Register of Systematic Reviews (PROSPERO) under registration number 1208548. The PRISMA-ScR checklist for this study has been completed and is available in Supplementary Material 1.\u003c/p\u003e\u003cp\u003eThe approach was chosen to comprehensively map the extent, range, and nature of research on AI-driven triage systems in emergency care settings, with particular attention to evidence from low- and middle-income countries (LMICs).\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eEligibility Criteria\u003c/h3\u003e\n\u003cp\u003eThe contextual scope includes emergency and acute care environments such as:\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eHospital emergency departments (EDs) and emergency rooms (ERs)\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003ePre-hospital or ambulance triage systems (including tele-triage platforms)\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eUrgent care centers or emergency units within primary or district hospitals\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eDisaster or mass-casualty triage applications relevant to low-resource environments\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eHealthcare settings in both public and private systems were eligible, provided they operated within resource-limited or LMIC contexts (as defined by World Bank classification). Studies conducted exclusively in high-income countries (HICs) were included only for comparative context, where they provided insights transferable to LMIC settings (e.g., model adaptation, algorithmic performance under data scarcity, or ethical implications).\u003c/p\u003e\n\u003ch3\u003eInclusion Criteria\u003c/h3\u003e\n\u003cp\u003eStudies were eligible if they:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eEvaluated AI-based or AI-augmented triage systems (machine learning, deep learning, NLP, hybrid models, or decision-support algorithms).\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eFocused on emergency or acute care settings, including hospitals, pre-hospital environments, or virtual triage systems.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eReported on at least one of the following outcomes:\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eTriage accuracy or validity compared to standard clinical assessment\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eTime-to-triage or waiting time reduction\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003ePatient flow efficiency, throughput, or prioritization improvement\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eClinical outcomes (e.g., adverse events, safety metrics)\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eImplementation feasibility, user acceptability, or workflow integration\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eWere conducted in LMICs or resource-limited contexts or offered data applicable to such settings.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eWere peer-reviewed original studies, systematic/scoping reviews, or technical reports with empirical data.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eWere published between January 2015 and September 2025, in English language.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003eExclusion Criteria\u003c/p\u003e\u003cp\u003eStudies were excluded if they:\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eFocused on AI tools not used for triage (e.g., diagnostic imaging, predictive analytics, or ICU prognostication).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eWe were conducted exclusively in high-income countries without relevance or transferability to resource-limited contexts.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eWere purely theoretical, editorials, or opinion pieces lacking empirical data.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eAddressed non-emergency care settings (e.g., outpatient clinics, elective care).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eWere published in languages other than English.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eFocused solely on algorithmic development or performance metrics without clinical application.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eAdditional Contextual Considerations\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eHealthcare system heterogeneity: Studies from LMICs often presented fragmented data infrastructures, paper-based triage records, and variable EHR integration, influencing AI feasibility and scalability.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eTechnology infrastructure: Only studies reporting functional deployment or piloting of AI-based triage (not simulations alone) were included.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eEquity and ethics: Studies addressing algorithmic fairness, workforce readiness, or policy barriers in digital health implementation in LMICs were prioritized.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eGrey literature inclusion: To capture policy and implementation realities, grey sources (WHO, World Bank, Ministry of Health reports, digital health pilot evaluations) were also reviewed.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eInformation Sources\u003c/h2\u003e\u003cp\u003eA comprehensive search was conducted across the following electronic databases: PubMed/MEDLINE, Scopus, Web of Science, IEEE Xplore, Cochrane Library, and Embase. Grey literature sources (Google Scholar, ResearchGate, and organizational reports from the WHO, World Bank, and Ministries of Health) were also searched to identify unpublished or policy-relevant studies [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eSearch Strategy\u003c/h2\u003e\u003cp\u003eThe search strategy combined controlled vocabulary terms (e.g., MeSH) and free-text keywords related to artificial intelligence, triage, emergency care, and resource-limited settings. An example PubMed search string was:\u003c/p\u003e\u003cp\u003e(\"artificial intelligence\"[MeSH Terms] OR \"machine learning\" OR \"deep learning\" OR \"natural language processing\")\u003c/p\u003e\u003cp\u003eAND (\"triage\"[MeSH Terms] OR \"emergency department triage\" OR \"emergency severity index\" OR \"acuity assessment\")\u003c/p\u003e\u003cp\u003eAND (\"emergency department\" OR \"emergency care\" OR \"emergency services\")\u003c/p\u003e\u003cp\u003eAND (\"low- and middle-income countries\" OR \"resource-limited\" OR \"developing countries\")\u003c/p\u003e\u003cp\u003eSearch strategies were adapted for each database. All searches were conducted between June 1 and September 15, 2025.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eSelection of Sources of Evidence\u003c/h2\u003e\u003cp\u003eAll records identified from the searches were imported into EndNote X9 for deduplication, then uploaded into Rayyan for screening. Two reviewers independently screened titles and abstracts, followed by a full-text review for eligibility. Disagreements were resolved through discussion or third-party adjudication [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eCritical Appraisal for Risk of Bias/Quality Assessment Individual Sources\u003c/h2\u003e\u003cp\u003eThe methodological quality and risk of bias of the included studies were evaluated using validated critical appraisal tools appropriate to each study design. Quantitative studies were assessed with the Joanna Briggs Institute (JBI) Critical Appraisal Checklists, which evaluate aspects such as sampling adequacy, outcome measurement validity, and confounding control. Qualitative and mixed-methods studies were appraised using the CASP (Critical Appraisal Skills Programmer) qualitative checklist, focusing on credibility, transferability, and methodological rigor. 22]. Each study was independently appraised by two reviewers, with discrepancies resolved through consensus. While no studies were excluded based solely on quality, the appraisal findings informed the interpretation of evidence strength and confidence in synthesized outcomes. Although scoping reviews do not require exclusion based on quality, an appraisal was undertaken to contextualize findings.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eStrategy for Data Synthesis\u003c/h2\u003e\u003cp\u003eData from included studies were synthesized using both quantitative descriptive and qualitative thematic approaches in accordance with the PRISMA 2020 guidelines and the Joanna Briggs Institute (JBI) framework for scoping reviews. Extracted data were tabulated to summarize study characteristics, AI model types, triage outcomes, and contextual factors relevant to emergency care in resource-limited settings.\u003c/p\u003e\u003cp\u003eQuantitative data (e.g., triage accuracy, waiting time reduction, efficiency metrics) were summarized using descriptive statistics, including frequencies, percentages, and mean or median values, to identify trends across studies. When comparable outcome measures were available, effect sizes (e.g., mean differences, relative risks) were reported narratively rather than pooled due to heterogeneity in study designs and metrics.\u003c/p\u003e\u003cp\u003eQualitative and mixed-methods findings were analyzed through inductive thematic synthesis, identifying recurring patterns related to feasibility, clinical acceptance, ethical concerns, and implementation barriers. Data synthesis focused on mapping evidence to key review objectives, efficiency, safety, feasibility, and equity, while highlighting research gaps specific to low- and middle-income countries (LMICs).\u003c/p\u003e\u003cp\u003eFindings were integrated into a narrative summary supported by evidence tables and visual figures (e.g., PRISMA flow diagram, outcome distribution charts) to ensure transparency and reproducibility of synthesis methods, consistent with PRISMA and JBI reporting standards.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eData Charting and Extraction\u003c/h2\u003e\u003cdiv id=\"Sec17\" class=\"Section3\"\u003e\u003ch2\u003eData Extraction and Management\u003c/h2\u003e\u003cp\u003eA structured data extraction process was employed to ensure consistency and transparency across all included studies. A standardized data charting form was developed in Microsoft Excel, adapted from the Joanna Briggs Institute (JBI) Scoping Review Manual, to systematically capture relevant study information. The extraction form was pilot tested on a subset of five studies and refined iteratively to improve clarity and comprehensiveness.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003eData Items Extracted:\u003c/h2\u003e\u003cp\u003eFor each included study, the following details were extracted:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eBibliographic information: author(s), year of publication, journal, and country of study.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eStudy characteristics: design, setting (e.g., tertiary hospital, urban/rural emergency department), and population.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eAI system attributes: algorithm type (machine learning, deep learning, NLP, hybrid), data source, and training/validation methods.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eTriage context: acuity scales used (e.g., ESI, CTAS, MTS), comparator (e.g., human or standard triage), and deployment environment.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eOutcomes reported: efficiency metrics (e.g., waiting time, patient flow), diagnostic accuracy (e.g., AUROC, sensitivity, specificity), feasibility, ethical considerations, and equity dimensions.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eKey findings and conclusions: including summary statistics and main thematic insights.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003eTwo reviewers independently extracted data and cross-verified entries to ensure accuracy. Discrepancies were resolved through consensus or consultation with a third reviewer. Where information was incomplete, corresponding authors were contacted where possible for clarification.\u003c/p\u003e\u003cp\u003eAll extracted data were stored in a shared, version-controlled dataset, with metadata tags to facilitate sorting by study design, region, or outcome domain. This database formed the basis for descriptive tabulation and thematic synthesis. Data management procedures followed the FAIR (Findable, Accessible, Interoperable, and Reusable) data principles, ensuring traceability and reproducibility of the synthesis process.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003eAnalysis Approach\u003c/h2\u003e\u003cp\u003eDue to heterogeneity in study designs, populations, and outcome measures, a thematic synthesis approach was used for qualitative and implementation outcomes, while quantitative findings were summarized descriptively. Meta-analysis was not feasible due to variations in outcome definitions, statistical reporting, and study contexts.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003cdiv id=\"Sec21\" class=\"Section3\"\u003e\u003ch2\u003eStudy Selection\u003c/h2\u003e\u003cp\u003eA total of 1,323 records were identified through databases and supplementary searches. After removing 254 duplicates, 1,069 records remained for title and abstract screening. Of these, 874 were excluded for being irrelevant to AI triage in emergency care or not situated in LMIC contexts. 195 full-text articles were assessed for eligibility, and 152 were excluded (67 not AI-driven, 39 not emergency care, 31 not LMIC-relevant, 15 editorials/reviews without primary data). Ultimately, 43 studies were included in the final synthesis. The study selection process is illustrated in the PRISMA 2020 flow diagram (Fig.\u0026nbsp;1).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e\u003ch2\u003ePRISMA 2020 Flow Diagram\u003c/h2\u003e\u003cp\u003e\u003cem\u003eFigure 1. PRISMA 2020 flow diagram illustrating the study selection process for a scoping review on AI-driven triage systems in emergency care in resource-limited settings. Numbers reflect identification, screening, eligibility, and inclusion stages, with a total of 43 studies included in the final synthesis.\u003c/em\u003e\u003c/p\u003e\u003cdiv id=\"Sec23\" class=\"Section3\"\u003e\u003ch2\u003eCharacteristics of Included Studies\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;1 summarizes the characteristics of the 43 included studies. These comprised:\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003e10 systematic/scoping/narrative reviews synthesizing global evidence on AI triage in EDs.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e21 quantitative studies, including RCTs, cross-sectional analyses, validation studies, and pilot projects.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e12 qualitative or mixed-methods studies, often exploring staff, patient, and system-level perspectives on implementation in LMIC contexts.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eCharacteristics of Included Studies (n\u0026thinsp;=\u0026thinsp;43)\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCharacteristic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCategories\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003en (%) of studies\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eGeographical region\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHIC (USA, UK, Korea, Germany, etc.)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e31 (72%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLMIC (India, Kenya, Nigeria, etc.)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e12 (28%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eStudy design\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSystematic/Scoping Reviews\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e10 (23%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eQuantitative (RCT, cross-sectional, pilots)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e21 (49%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eQualitative/Mixed methods\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e12 (28%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eAI model\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMachine Learning (ML)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e25 (58%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDeep Learning (DL)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e8 (19%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNatural Language Processing (NLP)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5 (12%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHybrid/Rule-based\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5 (12%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eOutcomes\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEfficiency gains\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e26 (60%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAccuracy \u0026amp; safety\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e19 (44%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFeasibility/acceptability\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e12 (28%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEthical/legal/bias\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e11 (26%)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eTable\u0026nbsp;1.\u003c/b\u003e Characteristics of Included Studies (n\u0026thinsp;=\u0026thinsp;43).\u003c/p\u003e\u003cp\u003e\u003cem\u003eSummary of study design, setting, geographical distribution, population, AI models used, and primary focus of included studies on AI-driven triage systems in emergency care across resource-limited settings.\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec24\" class=\"Section2\"\u003e\u003ch2\u003eGeographical Distribution\u003c/h2\u003e\u003cp\u003eThe majority of studies (n\u0026thinsp;=\u0026thinsp;31, ~\u0026thinsp;72%) were conducted in high-income countries (HICs) such as the United States, South Korea, Germany, Spain, and the United Kingdom. Only 12 studies (28%) were based in low- and middle-income countries (LMICs), including India, Kenya, Nigeria, Ghana, Pakistan, Bangladesh, Peru, and Brazil. This underscores the significant evidence gap in LMIC contexts, where the need for efficiency-enhancing triage tools is arguably greatest.\u003c/p\u003e\u003cdiv id=\"Sec25\" class=\"Section3\"\u003e\u003ch2\u003eStudy Designs and Methodological Rigor\u003c/h2\u003e\u003cp\u003eAmong quantitative studies (n\u0026thinsp;=\u0026thinsp;21), a minority were randomized controlled trials (RCTs) or prospective validation studies (e.g., in China, South Korea, and India), while the majority comprised retrospective or pilot evaluations. Systematic reviews (n\u0026thinsp;=\u0026thinsp;10) provided a comprehensive synthesis but frequently highlighted methodological heterogeneity and limited external validity. Qualitative studies (n\u0026thinsp;=\u0026thinsp;12), primarily from LMICs, offered crucial insights into barriers such as staff skepticism, digital illiteracy, workflow integration issues, and infrastructural limitations.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec26\" class=\"Section3\"\u003e\u003ch2\u003eAI Models Applied\u003c/h2\u003e\u003cp\u003eThe AI approaches varied considerably:\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eMachine learning (ML) classifiers were most frequently applied (n\u0026thinsp;\u0026asymp;\u0026thinsp;25 studies), demonstrating improved triage accuracy and reduced misclassification.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eDeep learning (DL) approaches, often applied in large hospital datasets (China, Korea, Brazil), achieved high predictive accuracy for sepsis detection, trauma severity, and triage acuity.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eNatural language processing (NLP) models (USA, Germany, Nigeria) were leveraged to analyze unstructured triage notes and enhance acuity prediction.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eHybrid ML\u0026thinsp;+\u0026thinsp;rule-based approaches (Kenya, India, Pakistan) were more common in LMIC pilots, as they required fewer computational resources and were easier to adapt to constrained settings.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec27\" class=\"Section3\"\u003e\u003ch2\u003eOutcomes Reported\u003c/h2\u003e\u003cp\u003eAcross the included studies, four major outcome themes emerged and have been summarized in Table\u0026nbsp;2:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e\u003cb\u003eEfficiency and throughput\u003c/b\u003e: Approximately 60% of quantitative studies reported reductions in patient wait times, improved throughput, and mitigation of ED overcrowding through AI-enabled triage support (e.g., Simbo.ai pilots, Mednition NLP, India/China ML trials).\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e\u003cb\u003eAccuracy and safety\u003c/b\u003e: Several DL and ML models (China, South Korea, Brazil) demonstrated enhanced predictive accuracy for triage prioritization, particularly in early detection of sepsis, trauma, and critical illness. These improvements suggest potential gains in patient safety, though external validation was limited.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e\u003cb\u003eFeasibility and acceptability\u003c/b\u003e: Qualitative studies in LMICs (Kenya, India, Ghana, Nigeria) highlighted moderate feasibility but mixed provider acceptance. Barriers included a lack of digital literacy, fear of algorithmic replacement, and infrastructure gaps. Facilitators included perceived improvements in decision confidence and workload reduction.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e\u003cb\u003eEthical, legal, and equity concerns\u003c/b\u003e: Reviews and qualitative studies emphasized the risks of algorithmic bias, lack of transparency in proprietary AI tools, and equity concerns when models trained in HIC datasets were applied in LMIC contexts.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eOutcomes of AI-Driven Triage Systems in Emergency Care (n\u0026thinsp;=\u0026thinsp;43)\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOutcome Category\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNumber of Studies (n)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eExamples of Findings\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eEfficiency \u0026amp; Patient Flow\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eReduced wait times, improved patient streaming (Yi et al., 2024; Simbo.ai, 2025)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eClinical Outcomes\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eFaster time-to-treatment, improved morbidity proxies (Abdalhalim et al., 2025)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eAccuracy \u0026amp; Safety\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e18\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHigher sensitivity/specificity than manual triage, fewer misclassifications (El Arab \u0026amp; Al Moosa, 2025; Porto, 2024)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eImplementation Barriers\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eInfrastructure gaps, data bias, ethical concerns (Božić, 2023; Friedman et al., 2024)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eFeasibility in LMICs\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e16\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHybrid/rule-based AI is more feasible in low-resource settings (Tahernejad et al., 2024)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eTable\u0026nbsp;2.\u003c/b\u003e Outcomes of AI-Driven Triage Systems in Emergency Care (n\u0026thinsp;=\u0026thinsp;43).\u003c/p\u003e\u003cp\u003e\u003cem\u003eOverview of key reported outcomes, including effects on triage accuracy, patient flow, waiting times, clinical decision support, feasibility in resource-limited settings, and identified challenges or barriers to implementation.\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec28\" class=\"Section2\"\u003e\u003ch2\u003eQuantitative Impact on Emergency Care Efficiency\u003c/h2\u003e\u003cp\u003eAcross multiple systematic reviews and observational studies, AI-driven triage systems consistently demonstrated measurable improvements in emergency department (ED) efficiency, even in resource-limited settings.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eReduced patient wait times were reported in both academic and industry analyses (Tyler et al., 2025; Simbo.ai Blog, 2025; Mednition, 2024), with reductions ranging from 15\u0026ndash;40% in high-volume EDs.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eShorter length of stay and faster time-to-treatment initiation were found in studies by Abdalhalim et al. (2025), Porto (2024), and Yi et al. (2024).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eTriage accuracy improvements were observed when AI was combined with natural language processing (NLP) for free-text data (Porto, 2024) and when models were trained on large, diverse datasets (Varughese et al., 2024).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eIn mass-casualty and disaster contexts, AI-enhanced triage significantly improved resource allocation and prioritization speed (Tahernejad et al., 2024), a critical factor in low-resource environments.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec29\" class=\"Section2\"\u003e\u003ch2\u003eClinical Outcomes and Patient Safety\u003c/h2\u003e\u003cp\u003eSeveral studies linked efficiency gains to better patient outcomes:\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eEnhanced early detection of critical cases and reduced over-triage rates were documented in El Arab \u0026amp; Al Moosa (2025) and Yi et al. (2024).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eIn military and austere environments, AI-driven autonomous trauma care systems improved survival during the \u0026ldquo;golden hour\u0026rdquo; (Philip \u0026amp; Harikumar, 2025).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eHowever, Ilick (2022) and Friedman et al. (2024) emphasized that accuracy varied considerably depending on data quality and warned against over-reliance without adequate clinical oversight.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eImplementation of Barriers and Facilitators in Resource-Limited Settings\u003c/h3\u003e\n\u003cdiv id=\"Sec31\" class=\"Section2\"\u003e\u003ch2\u003e\u003cb\u003eBarriers\u003c/b\u003e:\u003c/h2\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eInfrastructure constraints such as unreliable power supply, limited internet connectivity, and the absence of EHR integration tools (Philip \u0026amp; Harikumar, 2025).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eData-related issues, including a lack of high-quality annotated datasets for specific populations (Ilick, 2022; Tahernejad et al., 2024).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eHuman factors such as variable clinician trust in AI recommendations (Varughese et al., 2024) and the need for targeted training (El Arab \u0026amp; Al Moosa, 2025).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eRegulatory and ethical considerations, especially around patient data privacy (Yi et al., 2024; Božić, 2023).\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec32\" class=\"Section2\"\u003e\u003ch2\u003e\u003cb\u003eFacilitators\u003c/b\u003e:\u003c/h2\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eWorkflow adaptability, several studies (Tyler et al., 2025; Abdalhalim et al., 2025) found that AI triage tools could integrate into existing ED processes with minimal disruption.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003ePortability of some AI platforms, enabling deployment in mobile or rural clinics (Philip \u0026amp; Harikumar, 2025).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eCommercial readiness of certain industry tools, allowing for faster adoption in facilities with budget flexibility (Simbo.ai Blog, 2025; Mednition, 2024).\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cdiv id=\"Sec33\" class=\"Section3\"\u003e\u003ch2\u003eSelected Study Characteristics\u003c/h2\u003e\u003cp\u003eA total of 43 studies were included in the final synthesis and captured in Fig.\u0026nbsp;2. Geographically, the majority were conducted in high-income countries (HICs, n\u0026thinsp;=\u0026thinsp;31; 72%), while low- and middle-income countries (LMICs, n\u0026thinsp;=\u0026thinsp;12; 28%) contributed a smaller but growing proportion (Fig.\u0026nbsp;2, top-left). In terms of study design, nearly half employed quantitative approaches (n\u0026thinsp;=\u0026thinsp;21; 49%), followed by qualitative or mixed-methods studies (n\u0026thinsp;=\u0026thinsp;12; 28%), and systematic/scoping reviews (n\u0026thinsp;=\u0026thinsp;10; 23%) (Fig.\u0026nbsp;2, top-right).\u003c/p\u003e\u003cp\u003eAcross the included literature, the most frequently applied AI approaches were machine learning (ML, n\u0026thinsp;=\u0026thinsp;25; 58%), followed by deep learning (DL, n\u0026thinsp;=\u0026thinsp;8; 19%), natural language processing (NLP, n\u0026thinsp;=\u0026thinsp;5; 12%), and hybrid or rule-based models (n\u0026thinsp;=\u0026thinsp;5; 12%) (Figure X, bottom-left). Reported outcomes clustered around four major domains: efficiency gains such as reduced wait times and improved triage accuracy (n\u0026thinsp;=\u0026thinsp;26; 60%), accuracy and patient safety improvements (n\u0026thinsp;=\u0026thinsp;19; 44%), feasibility and clinical acceptability (n\u0026thinsp;=\u0026thinsp;12; 28%), and ethical or equity-related considerations including algorithmic bias (n\u0026thinsp;=\u0026thinsp;11; 26%) (Fig.\u0026nbsp;2, bottom-right).\u003c/p\u003e\u003cp\u003eTogether, these findings highlight the predominance of HIC-driven research and machine learning applications, while emphasizing critical evidence gaps in LMIC contexts, particularly regarding implementation feasibility, equity, and long-term impact.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003e(Top-left) Geographic distribution of studies, with most conducted in high-income countries (HICs, 72%) and a smaller proportion in low- and middle-income countries (LMICs, 28%). (Top-right) Study designs represented: quantitative (49%), qualitative/mixed-methods (28%), and systematic/scoping reviews (23%). (Bottom-left) AI approaches applied: machine learning (58%), deep learning (19%), natural language processing (12%), and hybrid/rule-based models (12%). (Bottom-right) Reported outcomes clustered into efficiency and workflow gains (60%), accuracy and patient safety (44%), feasibility and clinical acceptability (28%), and ethical/equity considerations (26%).\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis scoping review synthesized evidence from 43 studies on AI-driven triage systems in emergency care. Overall, findings suggest that AI offers significant potential for efficiency, accuracy, and workflow optimization in triage, but challenges persist in feasibility, ethical governance, and the applicability of evidence to low- and middle-income countries (LMICs).\u003c/p\u003e\n\u003ch3\u003eInterpretation of Findings in Resource-Limited Settings\u003c/h3\u003e\n\u003cp\u003eEmergency departments in LMICs face chronic shortages of staff, overcrowding, and inadequate infrastructure [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. AI-driven triage tools can theoretically automate early decision-making, reduce delays, and compensate for workforce gaps [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. However, the majority of studies reviewed originated from high-income settings, with only a minority evaluating AI in resource-limited contexts [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Even where LMIC-related studies exist, they are often retrospective or modeling-based, rather than prospective evaluations in real-world environments [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. This underlines a major evidence gap: while AI\u0026rsquo;s theoretical utility in LMICs is high, empirical validation remains scarce.\u003c/p\u003e\n\u003ch3\u003eComparison with High-Resource Settings\u003c/h3\u003e\n\u003cp\u003eIn high-income countries, AI triage has been shown to reduce wait times, improve patient flow, and enhance the prediction of critical outcomes such as ICU admission or mortality [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. These benefits are facilitated by robust digital infrastructure, interoperability with electronic health records, and strong regulatory frameworks [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. By contrast, LMICs lack comparable infrastructure and face challenges such as unreliable internet connectivity and limited health data digitization [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. This disparity suggests that successful AI triage adoption in LMICs requires adaptation to local realities rather than direct transfer of systems from high-resource settings.\u003c/p\u003e\u003cdiv id=\"Sec37\" class=\"Section2\"\u003e\u003ch2\u003eEfficiency and Workflow Optimization\u003c/h2\u003e\u003cp\u003eMany studies demonstrated that AI-driven triage could streamline workflow, improve throughput, and prioritize critical patients more effectively than manual triage [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. This is particularly relevant in disaster or pandemic contexts, where rapid prioritization of cases is essential [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Yet in LMICs, infrastructure bottlenecks (e.g., insufficient resuscitation capacity) may limit the extent to which efficiency gains translate into improved outcomes [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe most consistently reported outcome across studies was improved efficiency, including reductions in waiting times, faster allocation of acuity levels, and enhanced throughput. Machine learning, based models demonstrated the ability to process large volumes of patient data in real time, outperforming traditional manual triage in several high-income settings.\u003c/p\u003e\u003cdiv id=\"Sec38\" class=\"Section3\"\u003e\u003ch2\u003eAccuracy and Patient Safety\u003c/h2\u003e\u003cp\u003eImproved acuity assignment accuracy was a consistent finding, with several AI models outperforming human judgment in predicting adverse outcomes [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. This has direct implications for patient safety, especially in crowded emergency departments. However, concerns remain about algorithmic generalizability: most models were trained on patient populations from high-income settings, raising the risk of mis-triage when applied to LMIC populations with different epidemiological profiles [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec39\" class=\"Section2\"\u003e\u003ch2\u003eFeasibility and Clinical Acceptability\u003c/h2\u003e\u003cp\u003eOnly a minority of studies assessed feasibility, clinician acceptability, or integration into workflows [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. Clinicians voiced concerns about interpretability and liability in decision-making, echoing broader debates on trust in AI [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. In LMICs, additional barriers such as lack of training, limited IT support, and costs of deployment exacerbate feasibility challenges [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. Without addressing these, the real-world impact of AI triage in resource-limited settings may remain minimal.\u003c/p\u003e\u003cdiv id=\"Sec40\" class=\"Section3\"\u003e\u003ch2\u003eEthical and Legal Considerations\u003c/h2\u003e\u003cp\u003eEthical issues were consistently underexplored. Studies highlighted risks of bias, inequity, and data privacy breaches [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. In LMICs, reliance on algorithms trained in high-resource contexts may exacerbate inequities if local population characteristics are not represented [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. Furthermore, weak regulatory and legal safeguards in some LMICs heighten risks of misuse of sensitive health data [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. Addressing these issues requires transparent model development, inclusion of diverse datasets, and creation of robust governance frameworks.\u003c/p\u003e\u003cp\u003e\u003cb\u003eLimitations of Included Studies\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThis review revealed substantial heterogeneity in study designs, populations, and outcomes [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. Many studies were retrospective, with few prospective or randomized evaluations, limiting the strength of causal inferences [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. Additionally, reporting was often inconsistent, particularly regarding cost-effectiveness and unintended harm. The overwhelming dominance of high-income settings raises concerns about generalizability, especially in LMICs [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e].\u003c/p\u003e\u003cp\u003e\u003cb\u003eImplications for Clinical Practice and Policy\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFor clinicians, AI-driven triage offers the promise of earlier identification of critically ill patients and reduced cognitive burden [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]. For policymakers, successful adoption could enhance health system resilience, particularly during crises. However, policies must ensure investments in infrastructure, training, and regulation before large-scale implementation in LMICs [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. Otherwise, AI triage risks widen rather than narrowing global disparities in emergency care.\u003c/p\u003e\u003cp\u003e\u003cb\u003eRecommendations for Future Research\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo bridge the evidence gaps, future research should:\u003c/p\u003e\u003cp\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eConduct prospective, context-specific studies in LMICs.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eInclude cost-effectiveness and implementation of science analyses.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eAddress clinician acceptability and patient trust.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eIncorporate ethically grounded governance frameworks addressing bias, accountability, and transparency.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eFoster cross-disciplinary collaboration between AI developers, clinicians, policymakers, and communities.\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis scoping review demonstrates that AI-driven triage systems hold considerable promises for enhancing efficiency, accuracy, and patient safety in emergency care. However, the current evidence base is heavily skewed toward high-income settings, with only limited empirical evaluation in resource-limited and LMIC contexts. While AI has the potential to mitigate workforce shortages and streamline patient flow in these settings, its successful adoption requires adaptation to local infrastructures, rigorous validation in diverse populations, and governance frameworks that ensure equity, transparency, and accountability.\u003c/p\u003e\u003cp\u003eThe findings highlight both an opportunity and a caution: without context-sensitive implementation, AI triage may risk reinforcing global health inequities rather than alleviating them. Policymakers, clinicians, and researchers must therefore collaborate to design, evaluate, and regulate AI tools that are fit-for-purpose in low-resource environments. Future research should prioritize prospective evaluations, feasibility studies, and cost-effectiveness analyses in LMICs, ensuring that innovations in AI triage translate into meaningful and equitable improvements in emergency care outcomes worldwide.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAbbreviation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFull Term / Definition\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Intelligence\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eML\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMachine Learning\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDeep Learning\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNLP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNatural Language Processing\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eANN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Neural Network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eConvolutional Neural Network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMLP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMultilayer Perceptron\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLogistic Regression\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSupport Vector Machine\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRandom Forest\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eED\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEmergency Department\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eER\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEmergency Room\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLMIC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLow- and Middle-Income Country\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHIC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHigh-Income Country\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEHR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eElectronic Health Record\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI-ED\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Intelligence\u0026ndash;Enhanced Emergency Department (used contextually)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWHO\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWorld Health Organization\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWorld Bank\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCASP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCritical Appraisal Skills Programme\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJBI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJoanna Briggs Institute\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePRISMA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePreferred Reporting Items for Systematic Reviews and Meta-Analyses\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePRISMA-ScR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePreferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePROSPERO\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInternational Prospective Register of Systematic Reviews\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOSF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOpen Science Framework\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI-Triage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Intelligence\u0026ndash;Driven Triage System\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eQoC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eQuality of Care\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI-ML System\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Intelligence and Machine Learning System\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHRQoL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHealth-Related Quality of Life\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNCD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNoncommunicable Disease\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAIH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Intelligence in Healthcare\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCINAHL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCumulative Index to Nursing and Allied Health Literature (referenced in method description)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePRISMA 2020\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUpdated version of PRISMA guideline (2020 update)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIEEE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInstitute of Electrical and Electronics Engineers\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHIS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHealth Information System\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMoH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMinistry of Health\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI Bias\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAlgorithmic Bias or Artificial Intelligence Bias\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHFE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHuman Factors Engineering\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSDG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSustainable Development Goal\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eICT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInformation and Communication Technology\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRCT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRandomized Controlled Trial\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCAD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eComputer-Aided Decision System\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI-CDSS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Intelligence\u0026ndash;Based Clinical Decision Support System\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBMJ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBritish Medical Journal (referenced in style and format discussions)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eED Triage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEmergency Department Triage\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI-Triage Model\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArtificial Intelligence\u0026ndash;Driven Clinical Triage Model\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePHC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePrimary Health Care\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI Readiness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eThe preparedness of healthcare systems for integrating AI tools\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI Governance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEthical, regulatory, and policy structures governing AI use\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI Equity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEquitable deployment and access to AI technologies across income settings\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI-Enabled Workflow\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIntegration of AI tools into clinical operational workflows\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI-Driven Decision Support\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAutomated or semi-automated assistance for clinical prioritization and triage\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study does not involve human participants, and ethical approval was not required.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data extracted during this scoping review are included in the published article (and its supplementary information files). The search strategies, data extraction forms, and worksheets are available from the corresponding author upon reasonable request.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflicts of interest\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eJA conceptualized the study. C.A.A and D.O.O contributed to the design of the study. O.S.A and C.E contributed to the acquisition and analysis, while B.O.J and P.O.O contributed to the interpretation of data. J.A, C.A.A, and D.O.O drafted the study and O.S.A, C.E. and B.O.J substantively revised it. \u0026nbsp;J.A developed the search strategy with consultation from P.O.O and C.A.A. D.O.O and O.S.A screened, assessed the eligibility, and assessed the quality of the included studies with consultation from J.A and P.O.O, analyzed the data and created the figures with consultation from C.E and O.S.A. J.A is responsible for the data management and storage. All authors reviewed the final manuscript and approved the final version for submission. All authors have agreed both to be personally accountable for the author\u0026apos;s own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which they were not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors acknowledge the contributions of AwaDoc for assistance with search strategy development, members of the advisory panel for guidance throughout the review process, and all researchers whose work was synthesized in this review for their contributions to advancing knowledge regarding AI applications for triage system in emergency care.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors Information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eJoy Aifuobhokhan. MD,*\u003csup\u003e1\u003c/sup\u003e 0009-0005-9521-905X, Digital Health and Research, Lakeshore Cancer Centre, Lagos, Nigeria,\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eChukwuemeka Abraham Agbarakwe. MD\u003csup\u003e2\u003c/sup\u003e, 0000-0002-7267-3699, Emergency Medicine, Calvary Specialist Hospital, Port Harcourt, Nigeria\u003c/p\u003e\n\u003cp\u003eDeborah Oladunmolu Oduguwa. MD\u003csup\u003e3\u003c/sup\u003e, 0009-0003-8422-6153, Family Medicine, Babcock University Teaching Hospital, Ilishan-Remo, Nigeria\u003c/p\u003e\n\u003cp\u003eOseni, Olukayode A. MD\u003csup\u003e4\u003c/sup\u003e, 0009-0000-0148-5650, Urology, Lagos University Teaching Hospital\u003c/p\u003e\n\u003cp\u003eChimdi Eleweke\u003csup\u003e5\u003c/sup\u003e, 0009-0003-1907-3023, Emergency Medicine, Babcock University Teaching Hospital, Ilishan-Remo, Nigeria\u003c/p\u003e\n\u003cp\u003eBoluwatife Oluwafayoyimika Johnson. MD\u003csup\u003e6\u003c/sup\u003e, 0009-0004-2767-4547, Internal Medicine, Babcock University Teaching Hospital, Ilishan-Remo, Nigeria\u003c/p\u003e\n\u003cp\u003ePrecious Oluwadamilola Opawoye\u003csup\u003e7\u003c/sup\u003e 0009-0003-0626-6299, Research, Lakeshore Cancer Center, Lagos, Nigeria\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCorresponding Author:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eJoy Aifuobhokhan.MD\u003c/p\u003e\n\u003cp\[email protected]\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eHsia RY, Razzak J, Tsai AC, Hirshon JM. Placing emergency care on the global agenda. Ann Emerg Med. 2010;56(2):142\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.annemergmed.2010.03.028\u003c/span\u003e\u003cspan address=\"10.1016/j.annemergmed.2010.03.028\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eObermeyer Z, Abujaber S, Makar M, Stoll S, Kayden S, Wallis LA, Reynolds T. Emergency care in 59 low- and middle-income countries: a systematic review. Bull World Health Organ. 2015;93(8):577\u0026ndash;86. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2471/BLT.14.148338\u003c/span\u003e\u003cspan address=\"10.2471/BLT.14.148338\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eReynolds TA, Sawe HR, Rubiano AM, Shin SD, Wallis L, Mock CN. Strengthening health systems to provide emergency care. Disease Control Priorities, Third Edition (Volume 9): Injury Prevention and Environmental Health. The World Bank; 2017. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1596/978-1-4648-0522-6_ch13\u003c/span\u003e\u003cspan address=\"10.1596/978-1-4648-0522-6_ch13\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRazzak J, Kellermann AL. Emergency medical care in developing countries: is it worthwhile? Bull World Health Organ. 2002;80(11):900\u0026ndash;5. URL: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/pmc/articles/PMC2567679/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2567679/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGilboy N, Tanabe T, Travers D, Rosenau AM. Emergency Severity Index (ESI): A Triage Tool for Emergency Department Care, Version 4. Implementation Handbook 2012 Edition. Agency for Healthcare Research and Quality (US); 2012. URL: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ahrq.gov/esindex.html\u003c/span\u003e\u003cspan address=\"https://www.ahrq.gov/esindex.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVan der Wulp I, van Baar ME, Schrijvers AJP. Reliability and validity of the Manchester Triage System in a general emergency department patient population in the Netherlands: results of a simulation study. Emerg Med J. 2008;25(7):431\u0026ndash;4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/emj.2007.055228\u003c/span\u003e\u003cspan address=\"10.1136/emj.2007.055228\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJoy Aifuobhokhan, Tobiloba PO, Chijioke CE. Artificial Intelligence in Cancer Diagnosis: A Scoping Review of Global Innovation and African Implementation. World J Adv Res Reviews. 2025;28(01):449\u0026ndash;68. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.30574/wjarr.2025.28.1.3444\u003c/span\u003e\u003cspan address=\"10.30574/wjarr.2025.28.1.3444\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Article DOI.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJoy Aifuovhokhan, Olajire OO, Emmanuel TP, Artificial Intelligence - Clinical Decision Support Systems. From brain drain to health system gain. World J Adv Res Reviews. 2025;28(01):1482\u0026ndash;93. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.30574/wjarr.2025.28.1.3486\u003c/span\u003e\u003cspan address=\"10.30574/wjarr.2025.28.1.3486\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Article DOI.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAifuobhokhan J, Hussain AK, Ekechi CC, Olanrewaju AO, Afuadajo E, Bowale DA, Inioluwa OM. Predicting Thyroid Cancer Recurrence Using Machine Learning: An Artificial Intelligence Approach to Clinical Oncology. Int J Adv Artif Intell Mach Learn. 2025;2(3):135\u0026ndash;48. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.58723/ijaaiml.v2i3.469\u003c/span\u003e\u003cspan address=\"10.58723/ijaaiml.v2i3.469\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFernandes CM, Tanabe P, Gilboy N, Johnson LA, McNair RS, Rosenau AM, Sawchuk P, Thompson DA. Five-level triage: a report from the ACEP/ENA Five-level Triage Task Force. J Emerg Nurs. 2005;31(1):39\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jen.2004.11.002\u003c/span\u003e\u003cspan address=\"10.1016/j.jen.2004.11.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTyler S, Olis M, Aust N, Patel L, Simon L, Triantafyllidis C, Patel V, Lee DW, Ginsberg B, Ahmad H, Jacobs RJ. Use of Artificial Intelligence in Triage in Hospital Emergency Departments: A Scoping Review. Cureus. 2024;16(6):e248163. PMID: 38854295; PMCID: PMC11158416.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAhmed AAZ, Ahmed SNNA, Ezzelarab AMD, Mustafa M, Albasheer MGA, Ahmed REA, Elsayed MBGE. Clinical Impact of Artificial Intelligence-Based Triage Systems in Emergency Departments: A Systematic Review. Cureus. 2025;17(2):e85667. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.7759/cureus.85667\u003c/span\u003e\u003cspan address=\"10.7759/cureus.85667\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEl Arab RA, Al Moosa OA. The role of AI in emergency department triage: An integrative systematic review. Intensive Crit Care Nurs. 2025;79:104058. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.iccn.2025.104058\u003c/span\u003e\u003cspan address=\"10.1016/j.iccn.2025.104058\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTahernejad A, Sahebi A, Abadi ASS, et al. Application of artificial intelligence in triage in emergencies and disasters: a systematic review. BMC Public Health. 2024;24:3203. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12889-024-20447-3\u003c/span\u003e\u003cspan address=\"10.1186/s12889-024-20447-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePorto BM. Improving triage performance in emergency departments using machine learning and natural language processing: a systematic review. BMC Emerg Med. 2024;24:219. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12873-024-01135-2\u003c/span\u003e\u003cspan address=\"10.1186/s12873-024-01135-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYi N, Baik D, Baek G. The effects of applying artificial intelligence to triage in the emergency department: A systematic review of prospective studies. J Nurs Scholarsh. 2024;56(4):765\u0026ndash;77. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/jnu.13024\u003c/span\u003e\u003cspan address=\"10.1111/jnu.13024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFriedman AB, Delgado MK, Weissman GE. Artificial Intelligence for Emergency Care Triage\u0026mdash;Much Promise, but Still Much to Learn. JAMA Netw Open. 2024;7(5):e248857. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2024.8857\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2024.8857\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIlick J. Challenges in evaluating the accuracy of AI-containing digital triage systems: A systematic review. PLoS ONE. 2023;18(1):e0279636. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0279636\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0279636\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBožić V. Using Artificial Intelligence in Triage Process: Benefits, Challenges, and Considerations. Res Square Preprint. 2023. URL.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSimbo.ai. The Impact of Artificial Intelligence on Triage Systems in Emergency Departments: Enhancing Efficiency and Patient Care. Simbo Blog. 2023. Mednition. Emergency Triage: Revolutionizing Patient Care with AI for Better Outcomes. Mednition Blog. 2023. URL: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://mednition.com/blog/emergency-triage-revolutionizing-patient-care-with-ai-for-better-outcomes/\u003c/span\u003e\u003cspan address=\"https://mednition.com/blog/emergency-triage-revolutionizing-patient-care-with-ai-for-better-outcomes/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePeters MDJ, Godfrey CM, McInerney P, Baldini Soares C, Khalil H, Parker D. Methodology for JBI scoping reviews. In: Aromataris E, Munn Z, editors. Joanna Briggs Institute Reviewer\u0026rsquo;s Manual. The Joanna Briggs Institute; 2017. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://reviewersmanual.joannabriggs.org/\u003c/span\u003e\u003cspan address=\"https://reviewersmanual.joannabriggs.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMunn Z, Peters MDJ, Stern C, Tufanaru C, McArthur A, Aromataris E. Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Med Res Methodol. 2018;18:143. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12874-018-0611-x\u003c/span\u003e\u003cspan address=\"10.1186/s12874-018-0611-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTricco AC, Lillie E, Zarin W, O\u0026rsquo;Brien KK, Colquhoun H, Levac D, et al. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Ann Intern Med. 2018;169(7):467\u0026ndash;73. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.7326/M18-0850\u003c/span\u003e\u003cspan address=\"10.7326/M18-0850\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePage MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/bmj.n71\u003c/span\u003e\u003cspan address=\"10.1136/bmj.n71\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCritical Appraisal Skills Programme (CASP). CASP Checklists. Oxford: CASP. 2023. Available from Joanna Briggs Institute (JBI). JBI Critical Appraisal Tools. Adelaide: JBI; 2023. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://jbi.global/critical-appraisal-tools\u003c/span\u003e\u003cspan address=\"https://jbi.global/critical-appraisal-tools\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBramer WM, Rethlefsen ML, Kleijnen J, Franco OH. Optimal database combinations for literature searches in systematic reviews: a prospective exploratory study. Syst Rev. 2017;6:245. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13643-017-0644-y\u003c/span\u003e\u003cspan address=\"10.1186/s13643-017-0644-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHiggins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editors. Cochrane Handbook for Systematic Reviews of Interventions. Version 6.3 (updated February 2022). Cochrane; 2022. Available from. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://training.cochrane.org/handbook\u003c/span\u003e\u003cspan address=\"https://training.cochrane.org/handbook\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePorto BM. Improving triage performance in emergency departments using machine learning and natural language processing: a systematic review. BMC Emerg Med. 2024;24:219. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12873-024-01135-2\u003c/span\u003e\u003cspan address=\"10.1186/s12873-024-01135-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYi N, Baik D, Baek G, et al. The effects of applying artificial intelligence to triage in the emergency department: A systematic review of prospective studies. J Nurs Scholarsh. 2024;56(5):123\u0026ndash;34. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/jnu.13024\u003c/span\u003e\u003cspan address=\"10.1111/jnu.13024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFriedman AB, Delgado MK, Weissman GE. Artificial Intelligence for Emergency Care Triage\u0026mdash;Much Promise, but Still Much to Learn. JAMA Netw Open. 2024;7(5):e248857. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2024.8857\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2024.8857\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAhmed AAZ, Ahmed SNNA, Ezzelarab AMD, Mustafa M, Albasheer MGA, Ahmed REA, Elsayed MBGE. Clinical Impact of Artificial Intelligence-Based Triage Systems in Emergency Departments: A Systematic Review. Cureus. 2025;17(3):e85667. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.7759/cureus.85667\u003c/span\u003e\u003cspan address=\"10.7759/cureus.85667\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTyler S, Olis M, Aust N, Patel L, Simon L, Triantafyllidis C et al. Use of Artificial Intelligence in Triage in Hospital Emergency Departments: A Scoping Review. Cureus. 2025;17(4):e11158416. PMID:38854295; PMCID:PMC11158416. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cureus.com/articles/248163\u003c/span\u003e\u003cspan address=\"https://www.cureus.com/articles/248163\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEl Arab RA, Al Moosa OA. The role of AI in emergency department triage: An integrative systematic review. Intensive Crit Care Nurs. 2025;79:104058. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.iccn.2025.104058\u003c/span\u003e\u003cspan address=\"10.1016/j.iccn.2025.104058\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTahernejad A, Sahebi A, Abadi ASS, et al. Application of artificial intelligence in triage in emergencies and disasters: a systematic review. BMC Public Health. 2024;24:3203. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12889-024-20447-3\u003c/span\u003e\u003cspan address=\"10.1186/s12889-024-20447-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBožić V. Using Artificial Intelligence in Triage Process: Benefits, Challenges, and Considerations. 2023. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.researchgate.net/publication/371911400_Using_Artifical_Intelligence_in_Triage_Process_Benefits_Challenges_and_Considerations\u003c/span\u003e\u003cspan address=\"https://www.researchgate.net/publication/371911400_Using_Artifical_Intelligence_in_Triage_Process_Benefits_Challenges_and_Considerations\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIlick J. Challenges in evaluating the accuracy of AI-containing digital triage systems: A systematic review. PLoS ONE. 2022;17(12):e0279636. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0279636\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0279636\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArab RAE, Omayma Abdulaziz Al Moosa. The role of AI in emergency department triage: an integrative systematic review. Intensive Crit Care Nurs. 2025;79:104058. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.iccn.2025.104058\u003c/span\u003e\u003cspan address=\"10.1016/j.iccn.2025.104058\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMednition. Emergency Triage: Revolutionizing Patient Care with AI for Better Outcomes. 2024. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://mednition.com/blog/emergency-triage-revolutionizing-patient-care-with-ai-for-better-outcomes/\u003c/span\u003e\u003cspan address=\"https://mednition.com/blog/emergency-triage-revolutionizing-patient-care-with-ai-for-better-outcomes/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSimbo.AI. The Impact of Artificial Intelligence on Triage Systems in Emergency Departments: Enhancing Efficiency and Patient Care. 2024. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.simbo.ai/blog/the-impact-of-artificial-intelligence-on-triage-systems-in-emergency-departments-enhancing-efficiency-and-patient-care-2687317/\u003c/span\u003e\u003cspan address=\"https://www.simbo.ai/blog/the-impact-of-artificial-intelligence-on-triage-systems-in-emergency-departments-enhancing-efficiency-and-patient-care-2687317/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSimbo.AI. The Impact of AI-Driven Triage Systems on Reducing Patient Wait Times during Critical Healthcare Scenarios. 2024. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.simbo.ai/blog/the-impact-of-ai-driven-triage-systems-on-reducing-patient-wait-times-during-critical-healthcare-scenarios-2096665/\u003c/span\u003e\u003cspan address=\"https://www.simbo.ai/blog/the-impact-of-ai-driven-triage-systems-on-reducing-patient-wait-times-during-critical-healthcare-scenarios-2096665/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHsia R, Mbembati NA, Macfarlane S, Kruk ME. Access to emergency and surgical care in sub-Saharan Africa: the infrastructure gap. Health Policy Plan. 2012;27(3):234\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/heapol/czr023\u003c/span\u003e\u003cspan address=\"10.1093/heapol/czr023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlami H, Rivard L, Lehoux P, Hoffman SJ, Cadeddu SB, Savoldelli M, et al. Artificial intelligence in health care: laying the foundation for responsible, sustainable, and inclusive innovation in low- and middle-income countries. Global Health. 2020;16:52. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12992-020-00584-1\u003c/span\u003e\u003cspan address=\"10.1186/s12992-020-00584-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeslie D, Mazumder A, Peppin A, Wolters MK, Hagerty A. Does AI stand for augmenting inequality in the era of COVID-19 healthcare? BMJ. 2021;372:n242. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/bmj.n242\u003c/span\u003e\u003cspan address=\"10.1136/bmj.n242\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWHO. Ethics and Governance of Artificial Intelligence for Health: WHO Guidance. World Health Organization. 2021. Available from: European Commission. Ethics Guidelines for Trustworthy AI. High-Level Expert Group on AI. 2019. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai\u003c/span\u003e\u003cspan address=\"https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eO\u0026rsquo;Sullivan ED, Schofield SJ. Cognitive bias in clinical medicine. J R Coll Physicians Edinb. 2018;48(3):225\u0026ndash;32. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4997/JRCPE.2018.306\u003c/span\u003e\u003cspan address=\"10.4997/JRCPE.2018.306\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJobin A, Ienca M, Vayena E. The global landscape of AI ethics guidelines. Nat Mach Intell. 2019;1:389\u0026ndash;99. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s42256-019-0088-2\u003c/span\u003e\u003cspan address=\"10.1038/s42256-019-0088-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Artificial intelligence, emergency department, triage, low- and middle-income countries, healthcare efficiency, resource-limited settings","lastPublishedDoi":"10.21203/rs.3.rs-8097974/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8097974/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground:\u003c/h2\u003e\u003cp\u003eEmergency departments (EDs) in resource-limited settings face persistent challenges including overcrowding, delayed triage, and workforce shortages. Conventional triage systems often struggle under these conditions, contributing to inefficiencies and preventable morbidity. Artificial intelligence (AI)-driven triage systems have emerged as innovative tools to enhance patient prioritization, clinical decision-making, and resource allocation. However, their real-world impact on emergency care efficiency, particularly within low- and middle-income countries (LMICs), remains insufficiently defined.\u003c/p\u003e\u003ch2\u003eMethods:\u003c/h2\u003e\u003cp\u003eThis scoping review followed the Joanna Briggs Institute (JBI) methodology and adhered to PRISMA-ScR reporting standards. This review was registered with PROSPERO (1208548). Comprehensive searches were performed across PubMed, Scopus, Web of Science, IEEE Xplore, Cochrane Library, and Embase, supplemented by grey literature sources such as Google Scholar, WHO, and World Bank reports. Eligible studies examined AI-driven triage applications within emergency care settings, with relevance to LMICs or comparable resource-constrained environments. Data were charted and analyzed thematically to synthesize trends in design, implementation, and outcomes.\u003c/p\u003e\u003ch2\u003eResults:\u003c/h2\u003e\u003cp\u003eForty-three studies met inclusion criteria, comprising 21 quantitative, 12 qualitative or mixed-methods, and 10 review articles. Most research originated from high-income countries, though studies from LMICs are increasing. AI triage systems consistently improved patient flow, reduced waiting times, and enhanced alignment between triage levels and clinical urgency. In resource-limited contexts, AI supported overburdened clinicians, optimized staffing, and improved patient safety. Key challenges included algorithmic bias, limited data infrastructure, insufficient external validation, and ethical concerns surrounding transparency and accountability.\u003c/p\u003e\u003ch2\u003eConclusions:\u003c/h2\u003e\u003cp\u003eAI-driven triage systems demonstrate strong potential to enhance emergency care efficiency in resource-limited settings. However, context-specific evidence from LMICs remains limited. Future research should emphasize prospective validation, cost-effectiveness analysis, and ethical governance to ensure equitable and sustainable AI integration in emergency care.\u003c/p\u003e","manuscriptTitle":"Evaluating the Impact of Artificial Intelligence-Driven Triage Systems on Emergency Care Efficiency in Resource-Limited Settings: A Scoping Review","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-28 07:07:19","doi":"10.21203/rs.3.rs-8097974/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0ef9bf92-dcb7-49ea-83c7-f98b7c1fd33c","owner":[],"postedDate":"November 28th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-27T06:41:19+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-28 07:07:19","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8097974","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8097974","identity":"rs-8097974","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}