[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-363192-105":59,"doc-detail-363192-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","expert-guided-prompting-and-retrieval-augmented-generation-for-emergency-medical-service-question-answering","Expert-Guided Prompting and Retrieval-Augmented Generation for Emergency Medical Service Question Answering","","Large language models (LLMs) are promising for medical question answering but often miss the structured expertise professionals rely on, such as clinical subject areas and certification levels. The work introduces EMSQA, a 24.3K-question multiple-choice dataset covering 10 subject areas and 4 certification levels, with curated subject-aligned knowledge bases. Two methods are proposed: Expert-CoT and ExpertRAG. Experiments show accuracy gains and strong performance on EMS certification simulations.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/expert-guided-prompting-and-retrieval-augmented-generation-for-emergency-medical-service-question-answering/363192/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/expert-guided-prompting-and-retrieval-augmented-generation-for-emergency-medical-service-question-answering/363192.png","ImageObject",300,407,{"name":92,"@type":93},"Angel","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-29","2026-09-23",true,{"@type":102,"interactionType":103,"userInteractionCount":19},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What problem does the paper address in LLM-based emergency medical QA?","Question",{"text":112,"@type":113},"It targets the gap where LLMs often ignore structured domain expertise, including clinical subject areas and certification levels, which professionals use in high-stakes EMS decision-making.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"What is EMSQA and what resources does it provide?",{"text":117,"@type":113},"EMSQA is a 24.3K-question multiple-choice dataset covering 10 clinical subject areas and 4 certification levels. It is paired with curated, subject-aligned knowledge bases and real-world patient care reports.",{"name":119,"@type":110,"acceptedAnswer":120},"How do Expert-CoT and ExpertRAG improve performance?",{"text":121,"@type":113},"Expert-CoT conditions chain-of-thought on subject area and certification level, while ExpertRAG grounds generation in aligned documents and patient data. Combined, they yield up to a 4.59% accuracy gain over standard RAG baselines.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},363192,1792160197,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":19,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145,"preview_page_count":19,"preview_urls":146},687207412472,"https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d","Expert-Guided Prompting and Retrieval-Augmented Generation for Emergency Medical Service Question Answering  \nXueren Ge 1 , Sahil Murtaza 1 , Anthony Cortez2 , Homa Alemzadeh 1  \n1 School of Engineering and Applied Sciences, University of Virginia  \n2 School of Medicine, University of Virginia  \n[zar8jw@virginia.edu](zar8jw@virginia.edu), [vpn9ej@virginia.edu](vpn9ej@virginia.edu), [aec3gp@virginia.edu](aec3gp@virginia.edu), [ha4d@virginia.edu](ha4d@virginia.edu)  \narXiv :2511 . 10900v2 [ cs .CL] 18 Nov 2025  \nAbstract  \nLarge language models (LLMs) have shown promise in medical question answering, yet they often overlook the domainspecific expertise that professionals depend on-such as the clinical subject areas (e.g., trauma, airway) and the certification level (e.g., EMT, Paramedic) . Existing approaches typically apply general-purpose prompting or retrieval strategies without leveraging this structured context, limiting performance in high-stakes settings. We address this gap with EMSQA, an 24.3K-question multiple-choice dataset spanning 10 clinical subject areas and 4 certification levels, accompanied by curated, subject area-aligned knowledge bases (40K documents and 2M tokens) . Building on EMSQA, we introduce (i) Expert-CoT, a prompting strategy that conditions chain-ofthought (CoT) reasoning on specific clinical subject area and certification level, and (ii) ExpertRAG, a retrieval-augmented generation pipeline that grounds responses in subject areaaligned documents and real-world patient data. Experiments on 4 LLMs show that Expert-CoT improves up to 2.05% over vanilla CoT prompting. Additionally, combining ExpertCoT with ExpertRAG yields up to a 4.59% accuracy gain over standard RAG baselines. Notably, the 32B expertiseaugmented LLMs pass all the computer-adaptive EMS certification simulation exams.  \nCode & Data—[https://uva-dsa.github.io/EMSQA](https://uva-dsa.github.io/EMSQA)  \nIntroduction  \nThe rapid advancement of large language models (LLMs) has brought new possibilities to high-stakes domains such as emergency medical services (EMS) (Weerasinghe et al. 2024), where accurate and reliable decision-making is critical. There is growing interest in leveraging LLMs for medical education (Abd-Alrazaq et al. 2023), decision support (Ge et al. 2024; Luo et al. 2025), and certification preparation (Kung et al. 2023), particularly in the context of opendomain multiple-choice question answering (MCQA) . However, while LLMs have shown promising performance on general medical QA benchmarks (Cai et al. 2024), important gaps remain between their current capabilities and the reasoning processes used by trained medical professionals.  \nCopyright © 2026, Association for the Advancement of Artificial Intelligence ([www.aaai.org](www.aaai.org)). All rights reserved.  \nRecent approaches in medical MCQA, such as chain-ofthought prompting (CoT) (Wei et al. 2022) and retrievalaugmented generation (RAG) (Lewis et al. 2020), have improved LLM performance by enhancing reasoning capabilities and incorporating external domain knowledge during inference. But these approaches often treat both reasoning and retrieval as undifferentiated processes: the model sees a question and retrieves documents or reasons directly, without considering what kind of knowledge is relevant or how a human expert would approach the task. In contrast, realworld medical professionals typically begin by identifying the subject area of the question—e.g., whether it pertains to trauma, airway management, or pharmacology—and then reason from that domain-specific perspective, using knowledge and protocols appropriate to their level of certification.  \nExisting benchmarks such as MedQA (Jin et al. 2021), MMLU-Med (Hendrycks et al. 2020) and MedMCQA (Pal, Umapathi, and Sankarasubbu 2022) lack both this structured representation of question expertise (e.g., subject area, certification level) and the associated domain-specific knowledge. This makes it difficult to align re","cbCaitHTH44ciY2h","https://ap.wps.com/l/cbCaitHTH44ciY2h","pdf",1951064,18,"English","# Abstract\n# Introduction\n# Related Works\n## Medical Quest","[{\"question\":\"What problem does the paper address in LLM-based emergency medical QA?\",\"answer\":\"It targets the gap where LLMs often ignore structured domain expertise, including clinical subject areas and certification levels, which professionals use in high-stakes EMS decision-making.\"},{\"question\":\"What is EMSQA and what resources does it provide?\",\"answer\":\"EMSQA is a 24.3K-question multiple-choice dataset covering 10 clinical subject areas and 4 certification levels. It is paired with curated, subject-aligned knowledge bases and real-world patient care reports.\"},{\"question\":\"How do Expert-CoT and ExpertRAG improve performance?\",\"answer\":\"Expert-CoT conditions chain-of-thought on subject area and certification level, while ExpertRAG grounds generation in aligned documents and patient data. Combined, they yield up to a 4.59% accuracy gain over standard RAG baselines.\"}]","Expert-Guided Prompting and Retrieval-Augmented Generation for Emergency Medical Service Question Answering | PDF",1790150310,45,[147,151,155,159],{"page_no":8,"url":148,"width":149,"height":150},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/04a534da1e39829c6d99802893d4a645.webp",983,1272,{"page_no":14,"url":152,"width":153,"height":154},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/3e11af4008934852cd8d0b76f4620b72.webp",982,1270,{"page_no":81,"url":156,"width":157,"height":158},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/bb2998aefbf158ef4c8b014aa67f27d9.webp",981,1269,{"page_no":19,"url":160,"width":161,"height":162},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-16/4578a99d824aba6702aacad15bafbb43.webp",980,1268]