PDF(Pubmed)
Abstract：
BACKGROUND: Understanding stakeholders\' perception of cure in prostate cancer (PC) is essential to preparing for effective communication about emerging treatments with curative intent. This study used artificial intelligence (AI) for landscape review and linguistic analysis of definition, context and value of cure among stakeholders in PC.
METHODS: Subject-matter experts (SMEs) selected cure-related key words using Elicit, a semantic literature search engine, and extracted hits containing the key words from Medline, Sermo and Overton, representing academic researchers, health care providers (HCPs) and policymakers, respectively. NetBase Quid, a social media analytics and natural language processing tool, was used to carry out key word searches in social media (representing the general public). NetBase Quid analysed linguistics of key word-specific hit sets for key word count, geolocation and sentiments. SMEs qualitatively summarised key word-specific insights. Contextual terms frequently occurring with key words were identified and quantified.
RESULTS: SMEs identified seven key words applicable to PC (number of acquired hits) across four platforms: Cure (12429), Survivor (6063), Remission (1904), Survivorship (1179), Curative intent (432), No evidence of disease (381) and Complete remission (83). Most commonly used key words were Cure by the general public and HCPs (11815 and 224 hits), Survivorship by academic researchers and Survivor by policymakers (378 hits each). All stakeholders discussed Cure and cure-related key words primarily in early-stage PC and associated them with positive sentiments. All stakeholders defined cure differently but communicated about it in relation to disease measurements (e.g. prostate-specific antigen) or surgery. Stakeholders preferred different terms when discussing cure in PC: Cure (academic researchers), Cure rates (HCPs), Potential cure and Survivor/Survivorship (policymakers) and Cure and Survivor (general public).
CONCLUSIONS: This human-led, AI-assisted large-scale qualitative language-based research revealed that cure was commonly discussed by academic researchers, HCPs, policymakers and the general public, especially in early-stage PC. Stakeholders defined and contextualised cure in their communications differently and associated it with positive value.
METHODS: Subject-matter experts (SMEs) selected cure-related key words using Elicit, a semantic literature search engine, and extracted hits containing the key words from Medline, Sermo and Overton, representing academic researchers, health care providers (HCPs) and policymakers, respectively. NetBase Quid, a social media analytics and natural language processing tool, was used to carry out key word searches in social media (representing the general public). NetBase Quid analysed linguistics of key word-specific hit sets for key word count, geolocation and sentiments. SMEs qualitatively summarised key word-specific insights. Contextual terms frequently occurring with key words were identified and quantified.
RESULTS: SMEs identified seven key words applicable to PC (number of acquired hits) across four platforms: Cure (12429), Survivor (6063), Remission (1904), Survivorship (1179), Curative intent (432), No evidence of disease (381) and Complete remission (83). Most commonly used key words were Cure by the general public and HCPs (11815 and 224 hits), Survivorship by academic researchers and Survivor by policymakers (378 hits each). All stakeholders discussed Cure and cure-related key words primarily in early-stage PC and associated them with positive sentiments. All stakeholders defined cure differently but communicated about it in relation to disease measurements (e.g. prostate-specific antigen) or surgery. Stakeholders preferred different terms when discussing cure in PC: Cure (academic researchers), Cure rates (HCPs), Potential cure and Survivor/Survivorship (policymakers) and Cure and Survivor (general public).
CONCLUSIONS: This human-led, AI-assisted large-scale qualitative language-based research revealed that cure was commonly discussed by academic researchers, HCPs, policymakers and the general public, especially in early-stage PC. Stakeholders defined and contextualised cure in their communications differently and associated it with positive value.
摘要：
背景：了解利益相关者对前列腺癌（PC）治愈的看法对于准备有关具有治愈意图的新兴治疗方法的有效沟通至关重要。本研究使用人工智能(AI)进行景观评价和定义的语言分析，PC利益相关者之间治愈的背景和价值。
方法：主题专家（SME）使用Elicit，语义文献搜索引擎,从Medline中提取包含关键词的点击，Sermo和Overton,代表学术研究人员，医疗保健提供者(HCP)和政策制定者，分别。NetBaseQuid,社交媒体分析和自然语言处理工具，用于在社交媒体（代表公众）中进行关键字搜索。NetBaseQuid分析了针对关键字计数的特定关键字命中集的语言学，地理位置和情感。中小企业定性总结了针对关键词的见解。识别并量化了与关键词频繁出现的上下文术语。
结果：中小企业在四个平台上确定了适用于PC的七个关键字（获得的点击量）：Cure（12429），幸存者(6063),Remission(1904),幸存者（1179），治疗意图（432），无疾病证据（381）和完全缓解（83）。最常用的关键词是公众和HCP的Cure（11815和224次点击），学术研究人员的幸存者和政策制定者的幸存者（378命中）。所有利益相关者主要在早期PC中讨论了与治愈和治愈相关的关键词，并将其与积极情绪联系起来。所有利益相关者对治愈的定义都不同，但与疾病测量（例如前列腺特异性抗原）或手术有关。利益相关者在PC中讨论治愈时更喜欢不同的术语：治愈(学术研究人员)，治愈率（HCPs），潜在治愈和幸存者/幸存者（政策制定者）和治愈和幸存者（一般公众）。
结论：这个人类主导的，人工智能辅助的大规模定性基于语言的研究表明，治疗方法通常是由学术研究人员讨论的，HCP,决策者和公众，尤其是在早期PC。利益相关者在他们的沟通中不同地定义和情境化治疗，并将其与积极价值相关联。
方法：主题专家（SME）使用Elicit，语义文献搜索引擎,从Medline中提取包含关键词的点击，Sermo和Overton,代表学术研究人员，医疗保健提供者(HCP)和政策制定者，分别。NetBaseQuid,社交媒体分析和自然语言处理工具，用于在社交媒体（代表公众）中进行关键字搜索。NetBaseQuid分析了针对关键字计数的特定关键字命中集的语言学，地理位置和情感。中小企业定性总结了针对关键词的见解。识别并量化了与关键词频繁出现的上下文术语。
结果：中小企业在四个平台上确定了适用于PC的七个关键字（获得的点击量）：Cure（12429），幸存者(6063),Remission(1904),幸存者（1179），治疗意图（432），无疾病证据（381）和完全缓解（83）。最常用的关键词是公众和HCP的Cure（11815和224次点击），学术研究人员的幸存者和政策制定者的幸存者（378命中）。所有利益相关者主要在早期PC中讨论了与治愈和治愈相关的关键词，并将其与积极情绪联系起来。所有利益相关者对治愈的定义都不同，但与疾病测量（例如前列腺特异性抗原）或手术有关。利益相关者在PC中讨论治愈时更喜欢不同的术语：治愈(学术研究人员)，治愈率（HCPs），潜在治愈和幸存者/幸存者（政策制定者）和治愈和幸存者（一般公众）。
结论：这个人类主导的，人工智能辅助的大规模定性基于语言的研究表明，治疗方法通常是由学术研究人员讨论的，HCP,决策者和公众，尤其是在早期PC。利益相关者在他们的沟通中不同地定义和情境化治疗，并将其与积极价值相关联。
