The Use of Artificial Intelligence in Patient Triage in Emergency Departments: an Integrative Review

Objective: This study aims to explore how AI has been applied in patient triage in emergency services, investigating its contributions to the efficiency of care, the challenges faced in implementation, and the opportunities for response time optimization and resource allocation. Method: An integrative literature review was carried out with studies published between 2020 and 2024. The research was guided by the PICo (Population, Interest and Context) model, focusing on patients treated in emergency services (P), the application of AI for triage (I) and the hospital and emergency context (Co). The search was carried out in databases such as PubMed, LILACS, SciELO and Scopus, applying the Boolean code "Artificial Intelligence" AND "Triage" AND "Emergency Services". After screening 214 initial articles, 12 studies were selected for final analysis, based on the PRISMA guidelines. Results and Discussion: The review pointed out that AI significantly improves response time and reduces the margin of error in patient classification, especially in high-demand situations. The use of AI stood out in identifying critical outcomes, such as the need for immediate care. During the COVID-19 pandemic, AI has proven to be essential in the remote triage of high-risk patients, ensuring efficient use of resources in overcrowded settings. However, implementation faces challenges, such as resistance from health professionals and the need for integration with existing health systems. Conclusion: In summary, the use of AI in emergency services has brought significant benefits, such as increased triage efficiency, improved diagnostic accuracy, and improved resource management. However, overcoming cultural and operational barriers and setting clear ethical guidelines are essential. The careful integration of AI, with continuous training and periodic assessments, is critical to transforming care and ensuring more effective and safer patient care.