[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-563827-105":3,"detail-sidebar-cat-1-en-105":77,"doc-detail-563827-en":123},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":70,"head_meta":72,"extra_data":74,"updated_unix":76},105,"en","multi-agent-self-triage-system-with-medical-flowcharts-proof-of-concept-conversational-framework","Multi-agent Self-triage System with Medical Flowcharts - proof-of-concept conversational framework","","Online health resources and large language models are increasingly used for medical decision-making, but reliability is limited by low accuracy, insufficient transparency, and susceptibility to unverified claims. A proof-of-concept conversational self-triage system guides LLMs using 100 clinically validated American Medical Association flowcharts, enabling structured and auditable patient decision support. A multi-agent pipeline retrieves the most relevant flowchart, interprets responses, and delivers personalized recommendations. Scaled evaluation on synthetic conversations shows 95.29% top-3 retrieval accuracy and 99.10% navigation accuracy.",{"@graph":14,"@context":69},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/template/","Template",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/template/general/","General",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/template/multi-agent-self-triage-system-with-medical-flowcharts-proof-of-concept-conversational-framework/563827/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/multi-agent-self-triage-system-with-medical-flowcharts-proof-of-concept-conversational-framework/563827.png","ImageObject",442,249,{"name":42,"@type":43},"Guten tag","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-11-02","2026-10-12",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",6,{"@type":57,"mainEntity":58},"FAQPage",[59,65],{"name":60,"@type":61,"acceptedAnswer":62},"How was the system evaluated and how accurate were the results?","Question",{"text":63,"@type":64},"Performance was measured at scale using synthetic simulated conversations, achieving 95.29% top-3 accuracy for flowchart retrieval (N=2,000) and 99.10% accuracy for flowchart navigation across varied conversational conditions (N=37,200).","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"Why is this approach considered more trustworthy than free-text LLM advice?",{"text":68,"@type":64},"By combining free-text conversation flexibility with the rigor of standardized clinical protocols, it provides a structured, auditable framework that reduces ambiguity and improves transparency for patient decision-making.","https://schema.org",{"og:url":32,"og:type":71,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":73,"canonical":32},"index,follow",{"doc_id":75,"site_id":7},563827,1792475288,{"code":4,"msg":78,"data":79},"success",[80,85,90,95,100,105,110,115,120],{"id":81,"doc_module":22,"doc_module_name":25,"category_name":82,"show_sort_weight":83,"slug":84},11,"Presentations",90,"presentations",{"id":86,"doc_module":22,"doc_module_name":25,"category_name":87,"show_sort_weight":88,"slug":89},12,"Resumes",80,"resumes",{"id":91,"doc_module":22,"doc_module_name":25,"category_name":92,"show_sort_weight":93,"slug":94},14,"Invoices",70,"invoices",{"id":96,"doc_module":22,"doc_module_name":25,"category_name":97,"show_sort_weight":98,"slug":99},15,"Posters",60,"posters",{"id":101,"doc_module":22,"doc_module_name":25,"category_name":102,"show_sort_weight":103,"slug":104},16,"Social Media",50,"social-media",{"id":106,"doc_module":22,"doc_module_name":25,"category_name":107,"show_sort_weight":108,"slug":109},17,"Forms",40,"forms",{"id":111,"doc_module":22,"doc_module_name":25,"category_name":112,"show_sort_weight":113,"slug":114},18,"Letters",30,"letters",{"id":116,"doc_module":22,"doc_module_name":25,"category_name":117,"show_sort_weight":118,"slug":119},21,"Paper Templates",5,"papers-templates",{"id":121,"doc_module":22,"doc_module_name":25,"category_name":29,"show_sort_weight":4,"slug":122},158,"general-158",{"code":4,"msg":78,"data":124},{"doc_id":75,"user_id":125,"nickname":42,"user_avatar":126,"doc_module":22,"category_id":121,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":127,"file_id":128,"file_url":129,"file_type":130,"file_size":131,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":132,"language":133,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":134,"faqs":135,"seo_title":136,"seo_description":12,"update_tm":137,"read_time":138},687212321768,"https://ap-avatar.wpscdn.com/avatar/a0010bdbe886d2fe77?x-image-process=image/resize,m_fixed,w_180,h_180&k=1789897067658708522","Multi-agent self-triage system with medical flowcharts  \nYujia Liu1 ,2 , Sophia Yu3 , Hongyue Jin1 , Jessica Wen4 , Alexander Qian5 , Terrence Lee6 , Mattheus Ramsis7 , Gi Won Choi8 , Lianhui Qin3 , Xin Liu9 , 10 , Edward J. Wang1 ,2  \n1 Department of Electrical and Computer Engineering, University of California San Diego, La Jolla, CA, USA.  \n2 The Design Lab, University of California San Diego, La Jolla, CA, USA.  \n3 Department of Computer Science and Engineering, University of California San Diego, La Jolla, CA, USA.  \n4 Department of Medicine, Kaiser Permanente, San Diego, CA, USA.  \n5 Department of Radiation Oncology, University of California San Francisco, San Francisco, CA, USA  \n6 Department of Emergency Medicine, University of California San Francisco, San Francisco, CA, USA  \n7 Division of Cardiology, Department of Medicine, University of California San Diego, La Jolla, CA, USA  \n8 Department of Orthopaedic Surgery, Korea University Ansan Hospital, Ansan, South Korea  \n9 Google Research, Seattle, WA, USA  \n10 Paul G. Allen School of Computer Science & Engineering, University of Washington, Seattle, WA, USA.  \nAbstract  \nOnline health resources and large language models ( LLMs) are increasingly used as a first point of contact for medical decision-making, yet their reliability in healthcare remains limited by low accuracy, lack of transparency, and susceptibility to unverified information. We introduce a proof-of-concept conversational self-triage system that guides LLMs with 100 clinically validated flowcharts from the American Medical Association, providing a structured and auditable framework for patient decision support. The system leverages a multi-agent framework consisting of a retrieval agent, a decision agent, and a chat agent to identify the most relevant flowchart, interpret patient responses, and deliver personalized, patient-friendly recommendations, respectively. Performance was evaluated at scale using synthetic datasets of simulated conversations. The system achieved 95.29% top-3 accuracy in flowchart retrieval ( N=2 ,000) and 99.10% accuracy in flowchart navigation across varied conversational styles and conditions ( N=37 ,200) . By combining the flexibility of free-text interaction with the rigor of standardized clinical protocols, this approach demonstrates the feasibility of transparent, accurate, and generalizable AI-assisted self-triage, with potential to support informed patient decision-making while improving healthcare resource utilization.  \nMain  \nSelf-triage, the process by which people assess the urgency of their symptoms and decide when and where to seek medical care, is becoming increasingly prevalent with the use of online resources. According to the National Center for Health Statistics, 58.5% of adults searched for health or medical information online during a six-month period from July to December 20221 . Popular platforms such as Web MD and Mayo Clinic offer a wide range of health content, including symptom checkers, drug information, and in-depth articles on diseases and conditions. More recently, a growing number of people have begun relying on large-language models ( LLMs) for health-related inquiries, complex probabilistic algorithms that generate human-like textual responses, following the introduction of ChatGPT2 in 2022. A U.S. online survey of 2 ,002 participants found that 95.6% of the respondents used search engines for health queries, while 32.6% reported using LLMs such as ChatGPT to address medical or health questions in 20243. The trend in self-triage towards LLMs has shifted the initial point of contact in healthcare from clinicians to online platforms, which offer greater accessibility and cost-efficiency, particularly for individuals in low-and middle-income regions with limited healthcare infrastructure.  \nAt the other end of the spectrum, emergency rooms today are frequently overwhelmed with nonurgent visits. Nearly 40% of all emergency department ( ED) visits in ","cbCaip5Zs9XcXuJc","https://ap.wps.com/l/cbCaip5Zs9XcXuJc","pdf",1515819,36,"English","# Abstract\n## System Overview\n## Evaluation Results\n# Background and Motivation\n## Shift Toward LLM-based Self-triage\n## Emergency Department Overload\n# Challenges with Existing Tools\n## Accuracy, Literacy, and Actionability\n## Hallucination and Black-box Limitations\n## Need for Auditable Shared Representations","[{\"question\":\"How was the system evaluated and how accurate were the results?\",\"answer\":\"Performance was measured at scale using synthetic simulated conversations, achieving 95.29% top-3 accuracy for flowchart retrieval (N=2,000) and 99.10% accuracy for flowchart navigation across varied conversational conditions (N=37,200).\"},{\"question\":\"Why is this approach considered more trustworthy than free-text LLM advice?\",\"answer\":\"By combining free-text conversation flexibility with the rigor of standardized clinical protocols, it provides a structured, auditable framework that reduces ambiguity and improves transparency for patient decision-making.\"}]","Multi-agent Self-triage System with Medical Flowcharts - proof-of-concept conversational framework | PDF",1791838763,13]