[From PubMed to GPT: A narrative review of the emerging role of Large Language Models in systematic reviews].

IF 1 4区医学 Q4 OBSTETRICS & GYNECOLOGY

Gynecologie Obstetrique Fertilite & Senologie Pub Date : 2026-05-25 DOI:10.1016/j.gofs.2026.05.004

Benjamin Birene, Benjamin Popoff, Olivier Graesslin, Vincent Vuiblet, Olivier Morel

{"title":"[From PubMed to GPT: A narrative review of the emerging role of Large Language Models in systematic reviews].","authors":"Benjamin Birene, Benjamin Popoff, Olivier Graesslin, Vincent Vuiblet, Olivier Morel","doi":"10.1016/j.gofs.2026.05.004","DOIUrl":null,"url":null,"abstract":"Objective: Systematic reviews are a cornerstone of evidence-based medicine but remain time-consuming and prone to human error. Large Language Models (LLMs), such as ChatGPT or Claude, offer new opportunities for partial automation of these tasks. This articles aims to provide a critical synthesis of the current uses of LLMs across the key stages of systematic reviews and meta-analyses in healthcare.Methods: We conducted a narrative review based on a literature search in PubMed and Scopus (2019-2025), including empirical studies, scoping reviews, preprints, and technical reports discussing the use of LLMs in any stage of the systematic review process.Results: LLMs can support research question formulation, search strategy development, reference screening, data extraction, meta-analysis scripting, and result synthesis. Reported performances are often high, especially for screening and quantitative data extraction, with sensitivities of 95-98%. However, significant limitations persist: hallucinations, bias, misinterpretations, and variability across models. Independent validation remains scarce.Conclusion: LLMs show promising potential to accelerate several stages of systematic reviews, provided their use is methodologically controlled. A semi-automated approach, combining AI capabilities with human expertise, currently appears the safest. Prompt structuring, result validation, and transparent reporting of AI involvement are essential to ensure the quality and reliability of the synthesized evidence.","PeriodicalId":56056,"journal":{"name":"Gynecologie Obstetrique Fertilite & Senologie","volume":" ","pages":""},"PeriodicalIF":1.0000,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Gynecologie Obstetrique Fertilite & Senologie","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1016/j.gofs.2026.05.004","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"OBSTETRICS & GYNECOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

Objective: Systematic reviews are a cornerstone of evidence-based medicine but remain time-consuming and prone to human error. Large Language Models (LLMs), such as ChatGPT or Claude, offer new opportunities for partial automation of these tasks. This articles aims to provide a critical synthesis of the current uses of LLMs across the key stages of systematic reviews and meta-analyses in healthcare.

Methods: We conducted a narrative review based on a literature search in PubMed and Scopus (2019-2025), including empirical studies, scoping reviews, preprints, and technical reports discussing the use of LLMs in any stage of the systematic review process.

Results: LLMs can support research question formulation, search strategy development, reference screening, data extraction, meta-analysis scripting, and result synthesis. Reported performances are often high, especially for screening and quantitative data extraction, with sensitivities of 95-98%. However, significant limitations persist: hallucinations, bias, misinterpretations, and variability across models. Independent validation remains scarce.

Conclusion: LLMs show promising potential to accelerate several stages of systematic reviews, provided their use is methodologically controlled. A semi-automated approach, combining AI capabilities with human expertise, currently appears the safest. Prompt structuring, result validation, and transparent reporting of AI involvement are essential to ensure the quality and reliability of the synthesized evidence.

查看原文本刊更多论文

[从PubMed到GPT：大型语言模型在系统评价中的新兴角色的叙述性回顾]。

目的：系统评价是循证医学的基石，但仍然耗时且容易出现人为错误。大型语言模型（llm），如ChatGPT或Claude，为这些任务的部分自动化提供了新的机会。这篇文章的目的是提供一个关键的综合当前使用的法学硕士跨越系统评价和荟萃分析的关键阶段在医疗保健。方法：我们基于PubMed和Scopus（2019-2025）的文献检索进行了叙述性综述，包括实证研究、范围综述、预印本和技术报告，讨论了法学硕士在系统综述过程的任何阶段的使用。结果：法学硕士可以支持研究问题制定、搜索策略制定、参考筛选、数据提取、元分析脚本编写和结果综合。报道的性能通常很高，特别是在筛选和定量数据提取方面，灵敏度为95-98%。然而，显著的局限性仍然存在：幻觉、偏见、误解和模型之间的可变性。独立验证仍然很少。结论：法学硕士显示出有希望的潜力，加速几个阶段的系统评价，只要它们的使用是方法学控制。目前看来，将人工智能能力与人类专业知识相结合的半自动化方法是最安全的。快速构建、结果验证和人工智能参与的透明报告对于确保合成证据的质量和可靠性至关重要。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Gynecologie Obstetrique Fertilite & Senologie Medicine-Obstetrics and Gynecology

CiteScore

1.70

自引率

0.00%

发文量

170

期刊介绍： Gynécologie Obstétrique Fertilité & Sénologie est un mensuel scientifique d''information et de formation destiné aux gynécologues, aux obstétriciens, aux sénologues et aux biologistes de la reproduction. La revue, dans ses éditoriaux, articles originaux, mises au point, lettres à la rédaction et autres rubriques, donne une information actualisée ayant trait à l''obstétrique et à la gynécologie et aux différentes spécialités développées à partir de ces deux pôles : médecine de la reproduction, médecine maternelle et fœtale, périnatalité, endocrinologie, chirurgie gynécologique, cancérologie pelvienne, sénologie, sexualité, psychosomatique…