Scholarlink

<p>Conducting systematic or narrative literature reviews in Otorhinolaryngology (ENT) is a time-intensive process requiring extensive manual screening, summarization, and synthesis. The emergence of large language models (LLMs), such as ChatGPT, presents opportunities to streamline these tasks. To propose a structured methodological framework for integrating ChatGPT-based tools into ENT literature reviews and to evaluate their performance, efficiency, and limitations in a pilot application. Asix-step framework was developed aligning with PRISMA principles. The model was tested on the topic ‘Role of Nasal Microbiome in Chronic Rhinosinusitis’ using ChatGPT-4 and ChatGPT-5. Outputs were validated manually by two independent reviewers. ChatGPT demonstrated 92% accuracy in generating PubMed/Scopus-compatible search strings. AI-assisted screening reduced time by 46%, and data extraction time decreased by 38%. Summarization achieved 87% factual accuracy, though an 8% citation mismatch was noted. Reviewer satisfaction scored 4.5/5. ChatGPT-based frameworks significantly improve the efficiency of literature reviews in ENT research. When applied with structured oversight, these tools enhance productivity without compromising scientific integrity.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integration of ChatGPT-based Literature Review Tools in ENT Research: A Methodological Framework

  • 期刊论文

  • 发表日期：

    2025年11月17日

查看全文
  • Rohit Kumar,
  • Manish Kumar,
  • Ashish Kumar

摘要

Conducting systematic or narrative literature reviews in Otorhinolaryngology (ENT) is a time-intensive process requiring extensive manual screening, summarization, and synthesis. The emergence of large language models (LLMs), such as ChatGPT, presents opportunities to streamline these tasks. To propose a structured methodological framework for integrating ChatGPT-based tools into ENT literature reviews and to evaluate their performance, efficiency, and limitations in a pilot application. Asix-step framework was developed aligning with PRISMA principles. The model was tested on the topic ‘Role of Nasal Microbiome in Chronic Rhinosinusitis’ using ChatGPT-4 and ChatGPT-5. Outputs were validated manually by two independent reviewers. ChatGPT demonstrated 92% accuracy in generating PubMed/Scopus-compatible search strings. AI-assisted screening reduced time by 46%, and data extraction time decreased by 38%. Summarization achieved 87% factual accuracy, though an 8% citation mismatch was noted. Reviewer satisfaction scored 4.5/5. ChatGPT-based frameworks significantly improve the efficiency of literature reviews in ENT research. When applied with structured oversight, these tools enhance productivity without compromising scientific integrity.